{"id":"a8d86b74-f1b2-4d46-b015-752e810da32f","arxiv_id":"2412.04310","paper_version":3,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Eigenstate phase-space projections in chaotic systems are correlated along short-time classical trajectories, producing quantum trails that cause a wavepacket to remember its initial trajectory and weakly break ergodicity.","lead":"A single-author paper shows that in chaotic quantum systems, eigenstates have elongated correlations, called quantum trails, along classical trajectories that are not periodic, provided the wavepacket spreads slowly. This implies a long-time memory effect: a localized wavepacket remains more likely to be found along its initial short-time trajectory, weakly breaking ergodicity.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central mechanism hinges on e^{-iHt}|z0> ≈ |z_t> over the short-time window, but this fidelity is never directly measured; if it decays faster than assumed, the trails and memory effect lack their stated explanation.","rationale":"I agree with the reader that the untested wavepacket-fidelity assumption is the weakest link. The algebraic core of the paper (if |z_t>≈e^{-iHt}|z0>, then Q_E(z_t)≈Q_E(z0)) is sound and the numerics are suggestive, but the entire physical interpretation rests on the antecedent being true for the same finite window used to measure correlations. The paper's own definition of the trail length is the time for which this antecedent holds, so using the observed trails as evidence for it is circular. The concern is concrete and falsifiable: a direct fidelity calculation is well within the reach of the existing numerical method and would settle whether the central mechanism operates in the stadium billiard. I therefore do not propose changing the reader's conditional verdict; the paper should be accepted only if this check is supplied.","tokens_in":13722,"tokens_out":10533,"duration_ms":119966,"concrete_test":"Using the same boundary-integral solver and the same k, σ, and λ as in Fig. 2, compute F(t)=|⟨z_t|e^{-iHt}|z0⟩|^2 for an ensemble of initial coherent states z0 and times t∈[0,3/λ], and report the time-resolved mean and median F(t). Then compare (i) the time at which mean F(t) drops below 0.5 with the trail length visible in Fig. 2(d); and (ii) the prediction \\bar Q(z_t|z0)/\\bar Q(z_g|z0)=1+F(t) from Eq. (3) with the directly computed contrast in Fig. 3 for a fixed t in the window. If mean F(t) stays above 0.5 throughout [0,3/λ] and the predicted contrast matches, the concern is resolved; if not, the observed trails need an alternative account.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's core implication is conditional: if the coherent pointer state at the classical phase-space point z_t has high overlap with the true propagated wavepacket, α=|⟨z_t|e^{-iHt}|z0⟩|^2≈1, then Q_E(z_t)≈Q_E(z0) for every eigenstate and the memory effect follows. In the stadium calculation this condition is asserted rather than verified. The only quantitative evidence offered is the two-point correlation in Fig. 2(f), but that correlation is the very effect the condition is meant to explain, so the argument is circular at this step. The issue is not merely technical: for k≈95, σ=1/√k≈0.10, and λ≈0.86k≈82, the window t∈[0,3/λ] corresponds to several boundary collisions and to a Lyapunov stretching factor e^3≈20. A phase-space Gaussian with that stretching has an overlap with an isotropic coherent state of order 2/(√γ+1/√γ)≈0.43 at the end of the window, so α may already be far from 1. If the actual fidelity is low, the eigenstate correlations seen in Fig. 2(d,e) and the enhanced \\bar Q in Fig. 3 cannot be attributed to the proposed weakly-dispersing wavepacket mechanism; they could instead be artifacts of the finite wavepacket width, caustics, or conventional scarring. Checking F(t) directly is therefore required before the central claim is accepted beyond the conditional level.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces the notion of 'quantum trails': in a chaotic system, if a localized wave packet follows a classical trajectory with little dispersion, then energy eigenstates have phase-space projections Q_E(z) that are correlated along that short-time nonperiodic trajectory. This correlation is claimed to produce a 'memory effect' in the time-averaged phase-space projection, leading to weak ergodicity breaking. The author derives the expansion of the time-averaged projection in the energy basis (Eq. (2)), gives a heuristic derivation of the contrast formula (Eq. (3), detailed in Appendix A), and presents numerical evidence from the stadium billiard: phase-space unwarping visualizations of eigenstates, Pearson correlation curves in Fig. 2(f), and time-averaged distributions in Fig. 3. The Supplemental Material compares quantum with classical time-averaged distributions and defines a timescale t* for the onset of ergodicity breaking.","tokens_in":14054,"tokens_out":5420,"duration_ms":60248,"significance":"If the central premise is verified, the work is conceptually significant: it extends the notion of quantum scarring beyond unstable periodic orbits to generic nonperiodic trajectories and identifies a mechanism for weak ergodicity breaking that does not require eigenstate localization. The algebraic identity in Eq. (2) is exact, and the numerical comparison against Haar-random and random-wave states is a sensible control. The unwarping visualization is effective and the inclusion of a direct classical-versus-quantum comparison in the Supplemental Material is a strength. However, the paper's central claim is conditional on the wave-packet fidelity |<z_t|e^{-iHt}|z0>|^2 being close to 1 over the trail length, and this quantity is never directly computed or bounded. The significance is therefore real but currently rests on an unverified premise.","major_comments":[{"comment":"The load-bearing condition of the paper, alpha = |<zt|e^{-iHt}|z0>|^2 ≈ 1 over the 'short-time trajectory', is asserted but never directly measured. The trail length is defined by t < 3/lambda, and for the parameters of Fig. 2 (k ≈ 94.68, sigma ≈ 0.103, lambda ≈ 0.86k) this gives t up to about 0.037 in the billiard units, corresponding to a path length of about 3.5 and hence several collisions with the boundary, with a Lyapunov stretching factor e^3 ≈ 20. A direct computation of F(t) = |<zt|e^{-iHt}|z0>|^2, including intervals across collisions, is required to support the claim that weakly dispersing wave packets are the cause of the trails. Without it, the correlations seen in Fig. 2(f) demonstrate the existence of trail-like correlations but do not establish that they originate from the proposed wave-packet mechanism rather than from other sources such as boundary-induced caustics or conventional scarring.","section":"General intuition / Fig. 2(f) / SM Sec. I"},{"comment":"The contrast formula in Eq. (11) is essentially a restatement of the trail condition: the predicted enhancement is 1 + alpha, where alpha is exactly the overlap that defines the trail. Since alpha is not independently measured, the numerical enhancement in Fig. 3 is a necessary consequence of the eigenstate correlations through Eq. (2) but does not serve as an independent test of the mechanism. The paper should either report F(t) explicitly or clearly present Eq. (11) as a conditional relation rather than as a validated prediction.","section":"Eqs. (3) and (11) / Appendix A"},{"comment":"The derivation of Eq. (11) relies on two uncontrolled assumptions: the Porter-Thomas-type distribution of normalized projections (used to replace the variance by 1) and the statistical independence of <E|z0> and <E|rt> in Eq. (8). The first assumption is standard for chaotic eigenstates but is known to fail for visibly scarred states, several of which appear in Fig. 2(e) (e.g., n = 5005 and 5009). The second is plausible but not verified. Given that the central numerical contrast is not directly compared with Eq. (11), the author should at least check the covariance structure against the numerically available eigenstates to show that the assumptions are consistent with the stadium data.","section":"Appendix A"}],"minor_comments":[{"comment":"The phrase 'In the bottom of Fig. (3)' should be 'In the bottom panels of Fig. 3'.","section":"Page 4"},{"comment":"The phrase 'does not loose contrast' should be 'does not lose contrast'.","section":"Page 5"},{"comment":"The text says 'As discussed in the main test' but should say 'main text'.","section":"SM Sec. III"},{"comment":"The caption says 'consecutive eigenstates' but the row shows n = 5005–5011 while panel (d) shows n = 5016; please clarify the indexing or adjust the wording.","section":"Fig. 2(e)"},{"comment":"The cutoff t < 3/lambda is introduced in the quantitative analysis without a sensitivity study; since the length of the trails and the correlation curves depend on this cutoff, a sentence on how the results change with the chosen cutoff would strengthen the presentation.","section":"Fig. 2(f)"},{"comment":"The note added acknowledges the closely related work of Ref. [56] on 'birthmarks' but does not discuss the relation or difference between that work and the quantum trails presented here; a brief comparative sentence would help situate the contribution.","section":"Note added"}],"recommendation":"major_revision","confidential_remarks":"The paper is potentially publishable and the core picture is appealing, but the missing direct verification of the fidelity |<zt|e^{-iHt}|z0>|^2 is a load-bearing gap. A single additional figure showing F(t) for the same parameters as Figs. 2 and 3, together with a discussion of its decay over the trail window, would change the recommendation to minor revision or accept. The overlap with Ref. [56] is acknowledged, so novelty is not the main concern; the main issue is evidence for the proposed mechanism rather than the existence of the correlations."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a genuinely new observation, not just a reframing of scars. Pizzi shows that in the stadium billiard, eigenstates carry elongated correlations along non-periodic classical trajectories, and that these correlations produce a long-time memory effect in the time-averaged phase-space density. That is a real conceptual step beyond periodic-orbit scarring, and the unwarping visualization in Fig. 2 is a nice trick for seeing three-dimensional phase space on a page. The main equation (2) is just the energy-basis expansion of the time average, and the argument from a non-dispersing wavepacket to a trail is simple and correct.\n\nThe numerical evidence for the trails themselves is convincing: the phase-space plots of eigenstates show radially stretched features that are not present in Haar-random or random-wave superpositions, and the Pearson correlation in Fig. 2(f) quantifies the effect. The memory effect in Fig. 3 and the classical-quantum comparison in the supplement make the case that this is a genuine quantum interference effect, not classical ergodicity.\n\nWhere I'm more cautious is the mechanism. The paper's central premise is that a localized wavepacket follows its classical trajectory with high overlap, |<z_t|e^{-iHt}|z0>|^2 ~ 1, over the window [0, 3/λ]. That overlap is never directly computed or plotted. Instead, the paper infers the premise from the very eigenstate correlations it is meant to explain. That is not circular in a strict logical sense, but it leaves the explanation underdetermined. If the fidelity is actually low over that window—and a rough phase-space estimate suggests it could drop to ~0.4 by t=3/λ due to Lyapunov stretching alone—then the observed correlations might arise from other effects, like finite wavepacket width or scar contributions near periodic orbits. This is testable: just compute F(t) for a few wavepackets in the same billiard. A referee should ask for that.\n\nThe contrast formula in Appendix A is a secondary soft spot: it assumes Porter-Thomas fluctuations and statistical independence of the residual state |r_t>, neither of which is numerically checked. That's a minor issue, since the main qualitative claim doesn't depend on the exact contrast.\n\nThere's no code or data deposit, which is a shame for a paper whose main evidence is numerical.\n\nOverall: worth a serious referee. The phenomenon is real and the paper is readable. I'd send it out with a request for a direct fidelity check and a reproducibility statement.","headline":"A clean and genuinely new observation about eigenstate correlations in chaotic billiards, but the central wavepacket-fidelity premise is asserted rather than measured.","tokens_in":14574,"tokens_out":3362,"would_cite":true,"duration_ms":33344,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["05.45.Mt"],"model":"deepseek-v4-flash","headline":"Chaotic eigenstates carry elongated 'quantum trails' along any weakly dispersing classical path, periodic or not.","keywords":["quantum chaos","quantum trails","quantum scars","ergodicity breaking","phase-space correlations","stadium billiard","wave packet dynamics","memory effects"],"falsifier":"Compute the fidelity $|\\langle z_t|e^{-i\\hat H t}|z_0\\rangle|^2$ directly for a Gaussian wave packet in the stadium billiard at $k\\approx 94.7$ launched from the center at $\\theta_0=20^\\circ$, over times up to $3/\\lambda$ with $\\lambda\\approx 0.86 k$; if it drops far below one before the radial extent of the trails seen in Fig. 2, the stated identity cannot explain the correlations. A complementary test: repeat the correlation analysis in a billiard where wave packets disperse rapidly, such as a mushroom billiard, and check that the phase-space correlation length shrinks back to $\\sim h$.","tokens_in":13467,"feed_emoji":"🌀","tokens_out":11675,"duration_ms":99836,"temperature":0.7,"pith_summary":"This paper claims that energy eigenstates of a chaotic system are not the featureless speckles usually assumed: if a wave packet launched along a classical trajectory keeps its shape for a while, the eigenstates acquire elongated correlations—'quantum trails'—along that trajectory, even when the trajectory is not periodic. The result is demonstrated in the stadium billiard, where the trails make the long-time, time-averaged phase-space distribution of a wave packet cling to its own short-time classical path, an effect that breaks ergodicity. This matters because it generalizes quantum scars—enhancement along unstable periodic orbits—to essentially all trajectories, and it shows that wave-function localization is not necessary for memory: correlations of the eigenstates along trajectories suffice.","feed_headline":"Quantum trails etch memory into chaotic eigenstates","feed_subtitle":"Eigenstates stay correlated along short-time trajectories, so late-time distributions cling to their early path.","key_machinery":"The load-bearing object is the quantum trail: an elongated speckle, of width $\\sim h$ and length larger than $h$, in the phase-space projection $Q_E(z)$ of an eigenstate, aligned along a classical trajectory $z_t$. The trail exists precisely where the pointer-state fidelity $|\\langle z_t|e^{-i\\hat H t}|z_0\\rangle|^2$ remains close to one; that condition transfers the eigenstate projection at $z_0$ to $z_t$ via the phase $e^{-iEt}$, creating correlations on scales far exceeding $h$. The numerical analysis relies on an unwarping of the three-dimensional phase space onto the page: a fan of trajectories launched from one point is parameterized by polar coordinates $(t,\\theta)$, so trails become radial streaks and their length can be read off directly. The contrast of the memory effect is then estimated by binning energies and assuming the normalized phase-space projections follow a Porter-Thomas distribution, yielding the factor $1+|\\langle z_t|e^{-i\\hat H t}|z_0\\rangle|^2$.","core_discovery":"The paper's central claim is captured by a single identity: if a pointer state $|z_t\\rangle$ attached to a classical trajectory $z_t$ tracks the real dynamics, $e^{-i\\hat H t}|z_0\\rangle \\approx |z_t\\rangle$, then $\\langle E|z_t\\rangle \\approx e^{-iEt}\\langle E|z_0\\rangle$ and therefore $Q_E(z_t) \\approx Q_E(z_0)$ even when the phase-space distance $|z_t-z_0|$ is much larger than the localization length $h$. Numerically, eigenstates of the stadium billiard show such extended correlations along short-time trajectories, persisting across collisions with the boundary, whereas random wave superpositions show trails that break at each collision. Because these trails correlate $Q_E(z_0)$ with $Q_E(z_t)$, the time-averaged projection $\\bar Q(z|z_0)$ is enhanced along the trajectory by a factor approximately $1+|\\langle z_t|e^{-i\\hat H t}|z_0\\rangle|^2$, producing visible caustics in real space and a persistent memory of the initial condition. Since the classical stadium billiard is ergodic, the enhancement is a genuine quantum ergodicity breaking: it narrows as $k\\to\\infty$ but does not lose contrast.","pith_inferences":["The contrast formula $\\bar Q(z_t|z_0)/\\bar Q(z_g|z_0) \\approx 1 + |\\langle z_t|e^{-i\\hat H t}|z_0\\rangle|^2$ suggests that a measurement of the late-time enhancement along a known trajectory could serve as a direct experimental probe of the wave packet's fidelity, effectively a quantum-classical correspondence meter.","A natural test of the mechanism's scope is to compare trail length with the Lyapunov time across billiards with different horizon or curvature; the paper does not report such a sweep.","Because the mechanism needs only correlations and not localization, it may be relevant to transport and thermalization in disordered or quasiperiodic systems, where trail-like phase-space correlations could slow relaxation without many-body localization."],"forward_implications":["Quantum scars become a special case: trails appear along every weakly dispersing trajectory, periodic or not, so enhancement of eigenstates is generic rather than confined to unstable periodic orbits.","A system initialized in a localized wave packet keeps an enhanced probability of being found along that same short-time trajectory at arbitrarily long times, a weak ergodicity breaking that also shows up as caustics in real space.","Going deeper into the semiclassical limit narrows the trails without reducing their contrast, so ergodicity is restored only asymptotically as $k\\to\\infty$.","The same mechanism transfers to many-body systems if $z$ parametrizes a variational family and $z_t$ follows the time-dependent variational principle, suggesting trails and memory effects should be looked for beyond single-particle billiards."],"supporting_citations":[{"why":"introduces quantum scars on unstable periodic orbits, the phenomenon the trails generalize.","marker":"[14]"},{"why":"supplies the theory of eigenfunction scars and their link to ergodicity breaking that the memory effect extends.","marker":"[15]"},{"why":"provides the random-wave model of irregular eigenstates used as the baseline for the speckle comparison.","marker":"[11]"},{"why":"gives the pointer-state phase-space representation and the return-probability framework the argument builds on.","marker":"[33]"},{"why":"supplies the boundary integral method used to compute the eigenstates and wave-packet dynamics numerically.","marker":"[40]"},{"why":"proves ergodicity of the stadium billiard, the classical limit that the quantum time-averaged distribution is compared against.","marker":"[38]"},{"why":"establishes ergodicity for nowhere dispersing billiards, supporting the claim that the classical ensemble fills phase space uniformly.","marker":"[39]"},{"why":"gives the Porter-Thomas distribution assumed for the phase-space projections when estimating the contrast of the memory effect.","marker":"[27]"}],"fun_headline_variants":["Quantum trails in chaos break ergodicity with memory","Chaotic eigenstates remember their launching path","Short-time paths lasso long-time quantum probabilities","Quantum trails: memory that survives chaotic mixing","Eigenstates cling to past paths, defying chaotic mixing"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that a Gaussian wave packet in the stadium billiard stays close to its classical trajectory—$|\\langle z_t|e^{-i\\hat H t}|z_0\\rangle|^2 \\approx 1$—for a finite time, including across collisions; the paper infers this fidelity from the observed trails and their length rather than measuring it directly, so if packets disperse faster than assumed the trails and the memory effect would have a different origin.","fun_headline_variants_meta":{"raw":{"variants":["Quantum trails in chaos break ergodicity with memory","Chaotic eigenstates remember their launching path","Short-time paths lasso long-time quantum probabilities","Quantum trails: memory that survives chaotic mixing","Eigenstates cling to past paths, defying chaotic mixing"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00038,"raw_usage":{"total_tokens":2042,"prompt_tokens":991,"completion_tokens":1051,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":607,"completion_tokens_details":{"reasoning_tokens":978}},"tokens_in":607,"tokens_out":1051,"duration_ms":11327,"temperature":1.0,"reasoning_tokens":978,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T21:32:20.081320+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the fidelity $|\\langle z_t|e^{-i\\hat H t}|z_0\\rangle|^2$ directly for a Gaussian wave packet in the stadium billiard at $k\\approx 94.7$ launched from the center at $\\theta_0=20^\\circ$, over times up to $3/\\lambda$ with $\\lambda\\approx 0.86 k$; if it drops far below one before the radial extent of the trails seen in Fig. 2, the stated identity cannot explain the correlations. A complementary test: repeat the correlation analysis in a billiard where wave packets disperse rapidly, such as a mushroom billiard, and check that the phase-space correlation length shrinks back to $\\sim h$.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"introduces quantum scars on unstable periodic orbits, the phenomenon the trails generalize."},{"cited_title":"Kaplan and E","cited_arxiv_id":null,"evidence_quote":"supplies the theory of eigenfunction scars and their link to ergodicity breaking that the memory effect extends."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the random-wave model of irregular eigenstates used as the baseline for the speckle comparison."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"gives the pointer-state phase-space representation and the return-probability framework the argument builds on."},{"cited_title":"B ¨acker, Numerical aspects of eigenvalue and eigenfunction computations for chaotic quantum systems, in The mathemati- cal aspects of quantum maps (Springer, 2003) pp","cited_arxiv_id":null,"evidence_quote":"supplies the boundary integral method used to compute the eigenstates and wave-packet dynamics numerically."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"proves ergodicity of the stadium billiard, the classical limit that the quantum time-averaged distribution is compared against."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"establishes ergodicity for nowhere dispersing billiards, supporting the claim that the classical ensemble fills phase space uniformly."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"gives the Porter-Thomas distribution assumed for the phase-space projections when estimating the contrast of the memory effect."}],"review_version":1}