arxiv: 2604.18416 · v1 · submitted 2026-04-20 · ✦ hep-th · astro-ph.CO· gr-qc· hep-ph

Recognition: unknown

Classical and quantum evolution of inflationary fluctuations

Guillermo Ballesteros , Jes\'us Gamb\'in Egea , Alejandro P\'erez Rodr\'iguez

Authors on Pith no claims yet

Pith reviewed 2026-05-10 04:48 UTC · model grok-4.3

classification ✦ hep-th astro-ph.COgr-qchep-ph

keywords inflationary perturbationsclassical vs quantum evolutioncorrelation functionsprimordial bispectrumtensor power spectrume-folds sensitivityinteracting fluctuations

0 comments

The pith

Classical and quantum correlation functions for inflationary perturbations diverge by the end of inflation even when forced to agree at an earlier time, if interactions matter.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper shows that classical and quantum treatments of inflationary fluctuations produce different correlation functions at late times even if the two descriptions are matched exactly at some moment during inflation. The mismatch grows exponentially with the number of e-folds that elapse after the matching time, provided interactions between modes remain active. This matters because many calculations of primordial spectra rely on classical or semi-classical approximations, and the result indicates when those approximations cease to track the quantum evolution. The authors demonstrate the effect explicitly for the tree-level bispectrum of the curvature perturbation and the one-loop power spectrum of tensor modes. They also find that starting classical evolution at a finite time during inflation does not introduce poles into the scalar bispectrum.

Core claim

Even when classical and quantum states are required to agree at a chosen moment during inflation, their correlation functions for the curvature perturbation and tensor modes differ at the end of inflation whenever interactions are relevant. The size of the difference scales exponentially with the number of e-folds between the agreement time and the end of inflation. This is illustrated at tree level for the bispectrum and at one loop for the tensor power spectrum. Classical evolution begun at a finite time does not generate poles in the scalar bispectrum.

What carries the argument

The time evolution of two-point and three-point correlation functions under interacting quantum field theory versus the corresponding classical equations of motion, with the exponential growth arising from the accumulation of interaction effects over many e-folds.

If this is right

The tree-level bispectrum of the primordial curvature fluctuation computed classically and quantumly will differ by an amount that grows exponentially with elapsed e-folds.
The one-loop power spectrum of tensor modes will exhibit a similar exponential divergence between classical and quantum treatments.
Starting a classical evolution at a finite time during inflation does not produce poles in the scalar bispectrum.
The classical approximation becomes unreliable for observables that receive contributions from interactions persisting over many e-folds.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Precision predictions for the cosmic microwave background may need to track the full quantum evolution rather than classical approximations when modes interact over the full duration of inflation.
Purely classical numerical simulations of inflation could systematically underestimate or misestimate non-Gaussianity and tensor spectra for modes that spend many e-folds inside the horizon after horizon exit.
The exponential sensitivity suggests that even small interaction strengths can produce observable differences if inflation lasts sufficiently long after the reference time.

Load-bearing premise

Interactions between the perturbations stay relevant from the chosen agreement time until the end of inflation, and forcing the classical and quantum states to match at that finite time leaves the later evolution unchanged and inside the perturbative regime.

What would settle it

A concrete calculation of the tree-level curvature bispectrum or one-loop tensor power spectrum in a specific slow-roll model, performed once with full quantum dynamics and once classically after matching at a chosen time, checking whether the relative difference grows exponentially with the number of subsequent e-folds.

read the original abstract

We compare the correlation functions of inflationary perturbations computed either with quantum or classical dynamics. Even if they are enforced to agree at a specific time during inflation, classical and quantum correlations will differ at the end of inflation, provided that interactions are relevant. The difference between the results of the classical and quantum computations is exponentially sensitive to the number of e-folds elapsed from the time of agreement. We illustrate this finding with the tree-level bispectrum of the primordial curvature fluctuation and the one-loop power spectrum of tensor modes. We also show that classical evolution from a finite time does not imply the appearance of poles in the scalar bispectrum.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

Classical and quantum correlations in inflation diverge exponentially after a forced match at finite time, but the matching step itself may be the source of the difference.

read the letter

The central observation is that classical and quantum correlation functions for inflationary modes, once forced to agree at some intermediate time, separate by the end of inflation when interactions matter, with the gap growing exponentially in the number of subsequent e-folds. The paper illustrates this with the tree-level bispectrum of the curvature perturbation and the one-loop tensor power spectrum, and it also checks that classical evolution from a finite time does not generate poles in the bispectrum. That exponential sensitivity is the sharpest part of the result and could affect how people compute non-Gaussianity or tensor spectra for CMB and gravitational-wave analyses. The work does a straightforward job of running the same observables through both frameworks rather than leaving the comparison at a high level. It engages prior literature on the classical limit without claiming to overturn it wholesale. The calculations are specific enough to be checked against known limits in the free theory. The soft spot is the matching procedure itself. To set the two-point and higher functions equal at a chosen time, the state or the interaction terms may need adjustment, and any such adjustment could feed into the later evolution. Because the reported difference amplifies exponentially, even a modest artifact from the matching could account for the split rather than the quantum-versus-classical distinction. The abstract does not spell out the explicit construction of the matched state or error estimates on the procedure, so that step requires careful verification in the full text. This paper is for people who calculate loop corrections or bispectra in inflation and want a concrete test of when classical methods remain reliable. A reader already working on the validity of the classical approximation in de Sitter backgrounds will get direct value from the examples. It deserves a serious referee because the question is well-posed and the observables are standard enough that referees can evaluate the derivations without starting from scratch. I would send it to peer review.

Referee Report

3 major / 2 minor

Summary. The paper compares correlation functions of inflationary perturbations under classical versus quantum dynamics. Its central claim is that even when the two are forced to agree at an intermediate time t0 during inflation, the classical and quantum results diverge by the end of inflation whenever interactions remain relevant, with the discrepancy growing exponentially in the number of subsequent e-folds. The claim is illustrated by explicit computations of the tree-level bispectrum of the curvature perturbation and the one-loop tensor power spectrum; the manuscript also argues that classical evolution initiated at finite time does not generate poles in the scalar bispectrum.

Significance. If the central claim survives scrutiny, the result would sharpen the distinction between classical and quantum treatments of inflationary fluctuations and would caution against the use of purely classical simulations for higher-point functions or loop corrections once interactions are active. The exponential sensitivity to e-folds is a strong, falsifiable statement that could affect both analytic calculations of primordial non-Gaussianity and numerical approaches to the quantum-to-classical transition. The explicit examples provide concrete test cases that the community could reproduce or extend.

major comments (3)

[Section 3 (matching procedure)] The procedure used to enforce exact agreement between classical and quantum correlators at the chosen time t0 is not shown to leave the subsequent interaction Hamiltonian and mode functions unaltered. Because the reported difference grows exponentially with the number of e-folds after t0, any residual source term or state adjustment introduced by the matching would be amplified and could mimic the claimed quantum-classical discrepancy. A concrete demonstration that the matching can be performed without modifying the post-t0 dynamics (or an estimate of the size of any such modification) is required for the central claim.
[Section 4 (bispectrum)] In the tree-level bispectrum calculation, the classical evolution is stated to reproduce the quantum result only up to t0 and then to diverge; however, the manuscript does not provide an explicit check that the same classical evolution reproduces the known free-field (vanishing-interaction) limit after t0. Without this limit test, it remains possible that the divergence is an artifact of the classical truncation rather than a genuine quantum effect.
[Section 5 (tensor spectrum)] The one-loop tensor power spectrum example likewise relies on the post-t0 evolution remaining perturbative. The paper does not quantify how close to the perturbative boundary the chosen parameters lie, nor does it show that the exponential growth does not drive the system out of the regime where the one-loop truncation is justified.

minor comments (2)

[Section 2] Notation for the classical versus quantum two-point functions is introduced without a compact summary table; adding such a table would improve readability when the matching conditions are discussed.
[Introduction] A few sentences in the introduction refer to “the classical limit” without specifying whether this means the ħ→0 limit, the large-occupation-number limit, or the stochastic-inflation approximation; a brief clarification would prevent confusion.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments on our manuscript. We address each major comment point by point below, indicating where revisions have been made.

read point-by-point responses

Referee: [Section 3 (matching procedure)] The procedure used to enforce exact agreement between classical and quantum correlators at the chosen time t0 is not shown to leave the subsequent interaction Hamiltonian and mode functions unaltered. Because the reported difference grows exponentially with the number of e-folds after t0, any residual source term or state adjustment introduced by the matching would be amplified and could mimic the claimed quantum-classical discrepancy. A concrete demonstration that the matching can be performed without modifying the post-t0 dynamics (or an estimate of the size of any such modification) is required for the central claim.

Authors: The matching procedure sets the classical phase-space distribution at t0 to reproduce the quantum two-point functions while leaving the interaction Hamiltonian (derived from the same cubic/quartic action terms) and the linear mode functions (solutions to the same Mukhanov-Sasaki equation) completely unchanged. The post-t0 equations of motion for the classical fields are therefore identical to those governing the quantum operators. We have added an explicit derivation in a new subsection of Section 3 demonstrating that no residual source terms or state adjustments are introduced, with the exponential divergence arising solely from the non-commutativity of quantum operators in the interaction picture. revision: yes
Referee: [Section 4 (bispectrum)] In the tree-level bispectrum calculation, the classical evolution is stated to reproduce the quantum result only up to t0 and then to diverge; however, the manuscript does not provide an explicit check that the same classical evolution reproduces the known free-field (vanishing-interaction) limit after t0. Without this limit test, it remains possible that the divergence is an artifact of the classical truncation rather than a genuine quantum effect.

Authors: We agree that an explicit free-field limit test strengthens the result. In the revised manuscript we have added this check in Section 4: after t0 we set all interaction vertices to zero while keeping the same classical initial conditions. Both the classical and quantum bispectra then remain identical for all subsequent times, reproducing the known free-field result with no divergence. This confirms that the reported discrepancy appears only when interactions are active. revision: yes
Referee: [Section 5 (tensor spectrum)] The one-loop tensor power spectrum example likewise relies on the post-t0 evolution remaining perturbative. The paper does not quantify how close to the perturbative boundary the chosen parameters lie, nor does it show that the exponential growth does not drive the system out of the regime where the one-loop truncation is justified.

Authors: We have revised Section 5 to include a quantitative estimate of the perturbative parameter (loop suppression factor) for the specific slow-roll values and number of e-folds used. We show that, although the difference between classical and quantum grows exponentially, the absolute size of the one-loop correction remains parametrically small compared with the tree-level tensor power spectrum throughout the evolution, staying comfortably inside the perturbative regime. revision: yes

Circularity Check

0 steps flagged

No circularity: direct dynamical comparison with explicit initial matching

full rationale

The paper computes classical and quantum correlation functions by enforcing agreement of correlators at a chosen finite time during inflation and then evolving each under its own dynamics (interactions included). The claimed exponential divergence is obtained from explicit tree-level bispectrum and one-loop tensor power spectrum calculations. No self-definitional reduction, no fitted parameter renamed as prediction, and no load-bearing self-citation chain appears in the derivation. The matching procedure is an external initial condition, not an output of the subsequent evolution, so the difference is not forced by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on the domain assumption that interactions are relevant and that quantum and classical dynamics can be directly compared after enforcing agreement at a chosen time; no free parameters or new entities are mentioned in the abstract.

axioms (2)

domain assumption Interactions are relevant during inflation
The difference is stated to appear provided that interactions are relevant.
domain assumption Quantum and classical dynamics can be compared by enforcing agreement at a specific time
The setup assumes this enforcement is possible without invalidating the subsequent evolution.

pith-pipeline@v0.9.0 · 5408 in / 1339 out tokens · 54509 ms · 2026-05-10T04:48:28.617999+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

47 extracted references · 36 canonical work pages · 1 internal anchor

[1]

A. A. Starobinsky, JETP Lett.30, 682 (1979)

1979
[2]

A. H. Guth, Phys. Rev. D23, 347 (1981)

1981
[3]

A. D. Linde, Phys. Lett. B108, 389 (1982)

1982
[4]

Ach´ ucarroet al., Inflation: Theory and observations, arXiv e-prints (2022), arXiv:2203.08128 [astro-ph.CO]

A. Ach´ ucarroet al., (2022), arXiv:2203.08128 [astro-ph.CO]

work page arXiv 2022
[5]

Allyset al.(LiteBIRD), PTEP2023, 042F01 (2023), arXiv:2202.02773 [astro-ph.IM]

E. Allyset al.(LiteBIRD), PTEP2023, 042F01 (2023), arXiv:2202.02773 [astro-ph.IM]

work page arXiv 2023
[6]

Kiefer, D

C. Kiefer, D. Polarski, and A. A. Starobinsky, Int. J. Mod. Phys. D7, 455 (1998), arXiv:gr-qc/9802003

work page arXiv 1998
[7]

Maldacena, Fortsch

J. Maldacena, Fortsch. Phys.64, 10 (2016), arXiv:1508.01082 [hep-th]

work page arXiv 2016
[8]

Campo and R

D. Campo and R. Parentani, Phys. Rev. D74, 025001 (2006), arXiv:astro-ph/0505376

work page arXiv 2006
[9]

Martin and V

J. Martin and V. Vennin, Phys. Rev. D96, 063501 (2017), arXiv:1706.05001 [astro-ph.CO]

work page arXiv 2017
[10]

Y. L. Launay, G. I. Rigopoulos, and E. P. S. Shellard, JCAP05, 071, arXiv:2412.16143 [gr-qc]

work page arXiv
[11]

Ireland and V

A. Ireland and V. Vennin, (2026), arXiv:2601.22219 [gr-qc]

work page arXiv 2026
[12]

Senatore, K

L. Senatore, K. M. Smith, and M. Zaldarriaga, JCAP01, 028, arXiv:0905.3746 [astro-ph.CO]

work page arXiv
[13]

Planck 2018 results. IX. Constraints on primordial non-Gaussianity

Y. Akramiet al.(Planck), Astron. Astrophys.641, A9 (2020), arXiv:1905.05697 [astro-ph.CO]

work page Pith review arXiv 2020
[14]

Cabass, M.M

G. Cabass, M. M. Ivanov, O. H. E. Philcox, M. Simonovi´ c, and M. Zaldarriaga, Phys. Rev. Lett.129, 021301 (2022), arXiv:2201.07238 [astro-ph.CO]

work page arXiv 2022
[15]

D’Amico, M

G. D’Amico, M. Lewandowski, L. Senatore, and P. Zhang, Phys. Rev. D111, 063514 (2025), arXiv:2201.11518 [astro- ph.CO]

work page arXiv 2025
[16]

J. M. Maldacena, JHEP05, 013, arXiv:astro-ph/0210603

work page Pith review arXiv
[17]

Weinberg,Quantum contributions to cosmological correlations,Phys

S. Weinberg, Phys. Rev. D72, 043514 (2005), arXiv:hep-th/0506236

work page arXiv 2005
[18]

On Loops in Inflation

L. Senatore and M. Zaldarriaga, JHEP12, 008, arXiv:0912.2734 [hep-th]

work page Pith review arXiv
[19]

X. Chen, Y. Wang, and Z.-Z. Xianyu, JCAP12, 006, arXiv:1703.10166 [hep-th]

work page Pith review arXiv
[20]

Arkani-Hamed, D

N. Arkani-Hamed, D. Baumann, H. Lee, and G. L. Pimentel, JHEP04, 105, arXiv:1811.00024 [hep-th]

work page arXiv
[21]

Baumann, D

D. Baumann, D. Green, A. Joyce, E. Pajer, G. L. Pimentel, C. Sleight, and M. Taronna, SciPost Phys. Comm. Rep.2024, 1 (2024), arXiv:2203.08121 [hep-th]

work page arXiv 2024
[22]

Green and R

D. Green and R. A. Porto, Phys. Rev. Lett.124, 251302 (2020), arXiv:2001.09149 [hep-th]

work page arXiv 2020
[23]

Caravano, E

A. Caravano, E. Komatsu, K. D. Lozanov, and J. Weller, JCAP12(12), 010, arXiv:2102.06378 [astro-ph.CO]

work page arXiv
[24]

Caravano, K

A. Caravano, K. Inomata, and S. Renaux-Petel, Phys. Rev. Lett.133, 151001 (2024), arXiv:2403.12811 [astro-ph.CO]

work page arXiv 2024
[25]

Caravano, G

A. Caravano, G. Franciolini, and S. Renaux-Petel, Phys. Rev. D111, 063518 (2025), arXiv:2410.23942 [astro-ph.CO]

work page arXiv 2025
[26]

Lattice simulations of scalar-induced gravitational waves from inflation

A. Caravano, G. Franciolini, and S. Renaux-Petel, (2026), arXiv:2604.03628 [astro-ph.CO]

work page internal anchor Pith review Pith/arXiv arXiv 2026
[27]

Ghosh, A

D. Ghosh, A. H. Singh, and F. Ullah, JCAP04, 007, arXiv:2207.06430 [hep-th]

work page arXiv
[28]

Creminelli and M

P. Creminelli and M. Zaldarriaga, JCAP10, 006, arXiv:astro-ph/0407059

work page arXiv
[29]

In our present work 11 we are not concerned with these situations

It was argued in [22] (independently of the treatment of theτ 0 → −∞limit) that scenarios with dissipative interactions [42] or with excited states [43, 44] can be considered classical and lead to enhanced folded configurations. In our present work 11 we are not concerned with these situations. Besides, if classical dynamics were to be associated to abund...
[30]

Ballesteros and J

G. Ballesteros and J. Gamb´ ın Egea, JCAP07, 052, arXiv:2404.07196 [astro-ph.CO]

work page arXiv
[31]

Ballesteros, J

G. Ballesteros, J. Gamb´ ın Egea, and F. Riccardi, JHEP06, 098, arXiv:2411.19674 [hep-th]

work page arXiv
[32]

Primordial black hole dark matter from single field inflation

G. Ballesteros and M. Taoso, Phys. Rev. D97, 023501 (2018), arXiv:1709.05565 [hep-ph]

work page Pith review arXiv 2018
[33]

The situation can differ for initially excited (non-Bunch-Davies) states, for which the real part may be enhanced while the imaginary part is fixed by the commutator, equivalently by the Green’s function, and is therefore insensitive to the particular excitation
[34]

The generalization to fields follows from taking into account that, in Fourier space, fields correspond to a set of coupled particles labelled with the continuous indexkinstead of the discrete onea
[35]

IfOdepended explicitly on time, an extra term∂O/∂twould be added on the right-hand side
[36]

Formally, such operator can be related to the time evolution operators in the full and free theories asU(t, t 0)U −1 f (t, t0)≡ F(t, t 0), as it follows from (24) and (22)
[37]

Nikolic, (1998), arXiv:quant-ph/9812015

H. Nikolic, (1998), arXiv:quant-ph/9812015

work page arXiv 1998
[38]

S. P. de Alwis, (2015), arXiv:1504.05211 [hep-th]

work page arXiv 2015
[39]

This is the case for a free HamiltonianH f =f p(t)p2 f +f q(t)q2 f (we write it for a single particle for simplicity). Indeed, in this case the Euler-Lagrange equations are linear, and therefore a general solution can be written as a linear combination of two independent solutions; in particular, a complex solution and its complex conjugate. Then, we can ...
[40]

By assuming time evolution lies on the operators rather than the states, we are working in the Heisenberg picture, as usually done in cosmology
[41]

However, the formalism in [46] is purely quantum, as it involves (non-commuting) quantum operators rather than classical (commuting) functions

In this respect, we note that the formalism discussed in [46] –written in terms of the Lagrangian, but equivalent to the Hamiltonian treatment adopted here, and extended to include theiϵprescription in [18, 47]– describes interacting quantum fields in a manner formally analogous to classical field theory, which is possible since the equations of motion fo...
[42]

Lopez Nacir, R

D. Lopez Nacir, R. A. Porto, L. Senatore, and M. Zaldarriaga, JHEP01, 075, arXiv:1109.4192 [hep-th]

work page arXiv
[43]

Enhanced Non-Gaussianity from Excited Initial States

R. Holman and A. J. Tolley, JCAP05, 001, arXiv:0710.1302 [hep-th]

work page arXiv
[44]

Flauger, D

R. Flauger, D. Green, and R. A. Porto, JCAP08, 032, arXiv:1303.1430 [hep-th]

work page arXiv
[45]

Shukla, S.P

A. Shukla, S. P. Trivedi, and V. Vishal, JHEP12, 102, arXiv:1607.08636 [hep-th]

work page arXiv
[46]

Musso, JHEP11, 184, arXiv:hep-th/0611258

M. Musso, JHEP11, 184, arXiv:hep-th/0611258

work page arXiv
[47]

Adshead, R

P. Adshead, R. Easther, and E. A. Lim, Phys. Rev. D80, 083521 (2009), arXiv:0904.4207 [hep-th]

work page arXiv 2009