Pith. Sign in

REVIEW

Evolution of fNL to the adiabatic limit

PaperPDF from arXiv

data files

ReviewThe objections and verdict, as JSON Paper breakdownThe authors' claims, conditions and measurements, as JSON Signed receiptProof this review has not been altered
X Bluesky LinkedIn Reddit Hacker News
Joseph Elliston, David J. Mulryne, David Seery, Reza Tavakol

Cosmology and Nongalactic Astrophysics · Submitted 10 Jun 2011 · 92 references

Pith's verdict: A careful, honest numerical study showing that asymptotic fNL in multi-field inflation depends on when and how the adiabatic limit is reached, not just on the potential.

T0 · AI review

How this review was made: An AI referee (selfhosted/deepseek-v4.1-flash) read the paper on November 10, 2026 and answered a fixed list of questions about it. T0 is the starting level every paper gets. The levels describe how deeply Pith has checked the paper, never the authors.

  1. T0 AI review An AI read the paper and answered a fixed list of questions.
  2. T1 Audit Automatic checks that the references and files the paper relies on exist and match.
  3. T2 Compute The paper's code or calculations re-run with the same results.
  4. T3 Formal statement The claims written out in a formal math language a computer can read.
  5. T4 Proof Those claims proved, and the proof checked by computer.

How Pith checks papers →

· 2 major objections · 7 minor · Ask about this paper · The compiled record · Read the full review →
Challenge this review Comment (0) Read on arXiv ↗
Cite ▾
arXiv ID1106.2153 (version 2)
Pith ID2RYOKLWE, Pith's permanent record of this paper and its review
CitationDavid J. Mulryne, David Seery, Joseph Elliston, Reza Tavakol. "Evolution of fNL to the adiabatic limit." arXiv:1106.2153 (2011). https://pith.science/paper/2RYOKLWE
Cite this reviewBibTeX for Pith's review of this paper
Claim paper

Why Pith says this

Highlight any text to comment on it.

Referee report

How Pith checks papers → · Full audit of this paper →

The paper

The paper claims

The final amplitude of local non-Gaussianity in multi-field inflation depends on how inflation ends, not only on the shape of the potential, because the curvature perturbation keeps evolving until every isocurvature mode has decayed.

The reading

Inflation driven by more than one scalar field does not settle the statistics of the primordial density ripples at horizon exit. The curvature perturbation keeps being fed by the slower, transverse directions until those isocurvature modes die away — the paper calls that endpoint the adiabatic limit — and only then is the prediction frozen in. The authors ask whether the final amplitude of local non-Gaussianity, $f_{NL}$, depends on the route taken to that limit rather than on the potential alone, and they answer with numerical simulations that it does: changing when slow roll breaks down, or how fast the universe reheats, changes the number. That matters because $f_{NL}$ is a quantity observers actually measure, so if it can be shifted by physics at the end of inflation then a measured value constrains the whole history rather than just the inflationary plateau.

What carries the argument

The argument is carried by the $delta N$ formalism read as a statement about bundles of phase-space trajectories. An ensemble of smoothed spatial patches traces a collimated bundle; the curvature perturbation measures how the bundle's width maps onto expansion histories, and the adiabatic limit is the caustic where the bundle degenerates to a line, i.e. where $partial phi^c_k / partial phi^*_j to 0$. The decay rate of those derivatives, estimated as $e^{-int eta_s , dN}$ with $eta_s$ the lightest isocurvature mass in Hubble units, decides whether the final slice remembers anything. For sum-separable potentials the $delta N$ coefficients are known in closed form, and dropping the exponentially damped $c$-term leaves an $f_{NL}$ fixed by horizon-crossing data. For the transient cases the machinery is a matched asymptotic expansion: trajectories near a ridge disperse as $chi_c = chi_* D$ with $D = e^{beta(H_*^2 - H_c^2)}$ until the turn at $|chi_{text{turn}}| sim |g_0|/m_chi^2$, which yields the peak estimates $f_{NL}|_{text{peak}} approx eta_{chi *} delta_*$ for both ridge and valley with $delta_{text{valley}} = 1/delta_{text{ridge}}$.

Extended reading notes

Core claim

The paper's central claim is that the asymptotic value of local $f_{NL}$ is not a function of the inflationary potential alone. Whenever the isocurvature modes are not exhausted before slow roll fails, the value inherited by late-time observers depends on model-specific data: the time at which slow roll breaks down, and the timescale of reheating. The authors demonstrate this with explicit numerical runs — a hybrid model where a waterfall transition freezes a transiently large $f_{NL}$, a two-field axion-plus-quadratic model where the limit is reached only after the light field has begun oscillating and the final value sits below the slow-roll analytic estimate, and a model with no reconvergence in field space where the final $f_{NL}$ tracks the decay rate $Gamma$ used to model reheating. Against this, they identify a class where the answer *is* clean: in a sum-separable potential whose isocurvature modes decay before the end of slow roll, the $c$-terms in the $delta N$ coefficients are exponentially suppressed and $f_{NL}$ reduces to an 'enhanced' $eta$-parameter of the single dominant field, $f_{NL} approx -tfrac{5}{6} M_P^2 V''_phi/V_phi$ evaluated at horizon crossing, where the enhancement comes from the field's small share of the total energy density. Product-separable potentials cannot do this; there $f_{NL}$ stays of order the slow-roll parameters. Finally, the authors show that the two transient mechanisms — descent from a ridge and convergence into a valley — produce bispectra of opposite sign, inherited from the local $eta$ of the transverse field, and give approximate expressions for the peak value that scale as $1/chi_*$.

Load-bearing premise

The claim that the final $f_{NL}$ is sensitive to reheating is a statement about one specific family of histories: all fields decaying perturbatively at a common rate $Gamma$ into a single thermalized radiation fluid, and the hybrid transition modelled by a homogeneous waterfall value of order $H^3/M$.

Editorial extensions

If this is right

  • A measured $f_{NL}$ cannot be inverted to a unique potential: two models that agree on the inflationary plateau but differ in their reheating history can land on different values, so an observational bound becomes a joint bound on inflation and on reheating.
  • Large $f_{NL}$ at the adiabatic limit is achievable during slow roll in sum-separable potentials with one dominant light direction, so a large value does not by itself signal an unexhausted isocurvature mode.
  • Product-separable potentials are excluded as sources of observable $f_{NL}$ at the natural adiabatic limit, since the value remains of order the slow-roll parameters there.
  • The sign of the bispectrum distinguishes ridge dynamics (negative) from valley dynamics (positive) even when the magnitudes coincide, giving a discriminator that does not require an accurate peak height.
  • A fast waterfall transition can preserve a transiently large $f_{NL}$, which makes hybrid endings a workable route to observable non-Gaussianity provided the transition completes in less than an e-fold.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the spread in $f_{NL}$ induced by reheating is comparable to the spread between competing inflationary models, then forecasts treating $f_{NL}$ as a clean model discriminator are over-optimistic unless the post-inflationary history is specified alongside the potential.
  • The ridge/valley sign rule suggests a cheap numerical diagnostic: run a given potential with the bundle started on either side of a separatrix and check whether the sign of the bispectrum tracks the local $eta$ of the transverse field, as the peak formulas predict.
  • The sensitivity is largest precisely when focusing is only just incomplete at the end of slow roll, so models tuned near that boundary are where a full numerical treatment buys the most over the analytic horizon-crossing estimate.
  • The authors' remark that many fields may relax the fine-tuning needed for large $f_{NL}$ implies the observable is a sharper discriminator in the two-field regime than in the large-$N$ regime, where generic initial conditions already give a large value.
Original abstract

We study inflationary perturbations in multiple-field models, for which zeta typically evolves until all isocurvature modes decay--the "adiabatic limit". We use numerical methods to explore the sensitivity of the nonlinear parameter fNL to the process by which this limit is achieved, finding an appreciable dependence on model-specific data such as the time at which slow-roll breaks down or the timescale of reheating. In models with a sum-separable potential where the isocurvature modes decay before the end of the slow-roll phase we give an analytic criterion for the asymptotic value of fNL to be large. Other examples can be constructed using a waterfall field to terminate inflation while fNL is transiently large, caused by descent from a ridge or convergence into a valley. We show that these two types of evolution are distinguished by the sign of the bispectrum, and give approximate expressions for the peak fNL.

The compiled record

The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.

Compiled by classifier-1.13-20260917 via OpenRouter, question bank abstract-p1-20260923, 2026-09-23: 1 request, 4,386 input tokens, $0.0002 reported.

What the abstract itself states, as probabilities: claims first 0.03 · states limitation 0.41 · human participants 0.01 · controlled comparison 0.03 · randomized assignment 0.01 · states quantitative result 0.15 · compares with prior results 0.10 · mentions code or data availability 0.02.

The abstract, sentence by sentence

arXiv listing abstract v2. Results and the main claim are underlined; .

We study inflationary perturbations in multiple-field models, for which zeta typically evolves until all isocurvature modes decay--the "adiabatic limit". We use numerical methods to explore the sensitivity of the nonlinear parameter fNL to the process by which this limit is achieved, finding an appreciable dependence on model-specific data such as the time at which slow-roll breaks down or the timescale of reheating. In models with a sum-separable potential where the isocurvature modes decay before the end of the slow-roll phase we give an analytic criterion for the asymptotic value of fNL to be large. Other examples can be constructed using a waterfall field to terminate inflation while fNL is transiently large, caused by descent from a ridge or convergence into a valley. We show that these two types of evolution are distinguished by the sign of the bispectrum, and give approximate expressions for the peak fNL.

background contribution method result interpretation main claim

What the classifier reports for each sentence. The number is the probability it gave the label.
#sentencerolethe authors' own workstates the main conclusionstrength
s0 We study inflationary perturbations in multiple-field models, for which zeta typically evolves until all isocurvature modes decay--the "adiabatic limit". background 0.81 0.18 0.06 asserted 0.77
s1 We use numerical methods to explore the sensitivity of the nonlinear parameter fNL to the process by which this limit is achieved, finding an appreciable dependence on model-specific data such as the time at which slow-roll breaks down or the timescale of reheating. contribution 0.43 0.96 0.52 asserted 1.00
s2 In models with a sum-separable potential where the isocurvature modes decay before the end of the slow-roll phase we give an analytic criterion for the asymptotic value of fNL to be large. result 0.81 0.91 0.19 asserted 1.00
s3 Other examples can be constructed using a waterfall field to terminate inflation while fNL is transiently large, caused by descent from a ridge or convergence into a valley. result 0.54 0.60 0.08 asserted 0.63
s4 We show that these two types of evolution are distinguished by the sign of the bispectrum, and give approximate expressions for the peak fNL. result 0.92 0.94 0.19 asserted 1.00

What kind of paper the classifier reports this is: theoretical result 0.80 · evidence from simulation or computation 1.00.

What was examined, and what was not

sources read
  • arXiv listing abstract v2
not examined
  • full text
  • figures and tables
  • references and citing sentences
  • relations to other papers
levels run
  • P abstract
not run
  • S sections
  • T body sentences
  • C citations
  • R cross-paper relations

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. Every number in the measurements table is the authors', checked by code. Record as JSON

Figures
Figure 1. Evolution of fNL for the model (5.1) calculated numerically (solid red line) with the hybrid transition included for the parameter values in the text, the (red) dot-dashed line is added to illustrate the nearly constant final level of fNL. The (blue) dashed line, represents the analytical evolution, with no hybrid transition included. be calculated at the adiabatic limit using (4.1) after dropping the c-term provided inflation ends when the final field gracefully exits from slow-roll. Whether this occurs depends on the number of fields and the choice of fi . Taking fi = f < MP for…
Figure 2. Evolution of fNL for the model of Eq. (5.2) (first set of parameter choices). The solid red line is a numerical calculation. The blue dashed line is an analytic prediction. The horizontal green dashed line represents the analytically calculated adiabatic limiting value. should produce a positive spike, perhaps followed by a plateau. Finally, as the isocurvature modes are exhausted, the system should evolve to the adiabatic limit (4.8). These features are clearly visible in
Figure 3. Evolution of fNL for Eq. (5.2) (second set of parameter choices). The solid red line is a numerical calculation. The blue dashed line is an analytic prediction. Only the final few e-folds are shown. The horizontal green dashed line represents the analytically calculated adiabatic limiting value. The solid vertical line indicates when inflation ends, computed using the exact equations of motion. As the axion rolls, inflation momentarily restarts and the slow-roll expressions cease to be a good approximation. – 20 –
Figure 4. Evolution of fNL for the model (5.3). The solid red line is a numerical calculation for the parameter values quoted in the text and Γ = 0. – 22 –
Figure 5. Upper panel: numerical evolution of fNL (solid red line) for the parameter values quoted and Γ = (V0/10)1/2/MP. The blue dashed line represents the corresponding plot with Γ = (V0/100)1/2/MP, and the green dot-dashed line represents the analytical evolution. The an￾alytic evolution terminates when the χ field reaches zero, because the slow-roll expressions can’t evolve past this point. Lower panel: magnified in the vicinity of the end of inflation. The asymptotic value of fNL depends on Γ, and therefore on microphysical details of the reheating phase. Where the parameters Λi and fi…
Figure 6. Evolution of fNL for the Nf-axion model, calculated analytically (solid blue line) under the slow-roll approximation. The horizontal dot-dashed red line is the asymptotic value computed using the horizon-crossing approximation. (This is unreliable in the present case.) The vertical black line corresponds to ǫ = 1. Since fNL has not reached the adiabatic limit at this point, this model is an example in which the adiabatic limit is reached after slow-roll ends. Therefore, numerical analysis is required to obtain a reliable value for fNL. There is a specific case where this model can…

Figures reproduced from arXiv:1106.2153 by the authors.

References
How later papers used it

59 later papers cite it

Evidence: substantial. 38 of the citing papers share no author with it, and 6 of those carry a confident label from Pith's citation classifier.

Pith read 15 of the 160 citing sentences, choosing only those whose citation markers point at this paper alone, and wrote the paraphrases from them. The summary, the counts, the evidence level, the links and the quotes come from code; the summary counts the papers listed below. A later paper is listed under a label when Pith's citation classifier gives one of its sentences that label at probability 0.8 or more (0.95 for support), Pith's reading finds that the citing paper itself did it, and a second check confirms the paraphrase against the sentence. This section lists uses only. It does not list criticism or disagreement, so their absence here says nothing about how the paper has been received. Papers that share an author with this one are listed apart, under "By the same authors". A quote is the citing authors' own sentence, shown when their paper's licence is CC BY, CC BY-SA or CC0. JSON

Uses its method or tool 2 of 5 papers with this label · 2 by the same authors

By the same authors

How the literature uses it

59 arXiv papers cite this in 160 sentences

Each citing sentence is taken from the citing paper's LaTeX source and typed by Pith's classifier. A label appears at probability 0.8 or more (0.95 for support), a contest at 0.7 or more; the quoted words are the citing authors'. How each label was checked · JSON

Papers citing it · 120 sentences without a confident label

Builds on it 2 papers
  • [ref] Here we give a summary of those results and further elucidate them by considering an inflection point model, which is of the first type, and which can produce a large positive or negative value of $$ at the adiabatic limit.

    Evolution of non-Gaussianity in multi-scalar field models · Introduction · 0.85

  • More generally however, it would be interesting to apply similar techniques to [ref] in the case of non-standard kinetic terms, to further understand what we can infer about the dynamics of inflation from observations of non-Gaussianity.

    Local non-Gaussianity from rapidly varying sound speeds · Conclusions · 0.82

Uses its method or tool 5 papers
Compares with it 4 papers
  • This time evolution of $$ closely matches the behaviour described by Elliston et al., associated with rolling between two approximate critical points of the potential [ref] .

    Non-Gaussianity in D3-brane inflation · Large non-Gaussianity from rapid change of angular minima · 0.94

  • This definition of the reheating hypersurface is more refined than that of [ref] , where reheating was initiated at a surface of constant density.

    Reheating, Multifield Inflation and the Fate of the Primordial Observables · Numerical Code · 0.89

  • This is in contrast to other two field models which produce a large non-Gaussianity due to inflationary dynamics [ref] , which require both carefully choosen model parameters, and offer no explanation for the origin of the finely tuned horizon crossing conditions.

    Non-Gaussianity from the hybrid potential · Conclusions · 0.99

Cites it as support 3 papers
Cites it as background 22 papers

What it cites · 124 sentences without a confident label

Builds on 1 paper
Uses the method or tool of 1 paper
Uses data or values from 1 paper
Compares with 2 papers
  • We have checked a small number of these cases numerically and find agreement with the formulae given in Ref. [ref] , although we have only considered examples of positive curvature rather than ridges.]

    Large non-Gaussianity from two-component hybrid inflation · Transitory models with interruption · 0.96

  • (This model has some similarity to the scenario of Boubekeur & Lyth [ref] .) Therefore, most of the axions can be replaced by a single effective field with a quadratic potential, retaining the full cosine only for the axion closest to the hilltop, [equation] where $Λ$ and $f$ are constants.

    Detecting a small perturbation through its non-Gaussianity · Large non-Gaussianity at the natural adiabatic limit · 0.81

Cites as support 1 paper
Cites as background 16 papers

Discussion

Disagree with the review, or know something it missed? Leave a comment, highlight any passage above to comment on it, or challenge this review formally.