Mild Non-Gaussianities under Perturbative Control from Rapid-Turn Inflation Models

M.C. David Marsh; Ricardo Z. Ferreira; Theodor Bjorkmo

REVIEW 2 major objections 4 minor 3 cited by

Mild Non-Gaussianities under Perturbative Control from Rapid-Turn Inflation Models

T0 review · 2 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Rapid-turn inflation survives: exponentially large non-Gaussianities cancel exactly once nested commutators are accounted for.

desk verdict A solid, genuinely useful paper: it supplies the first analytic growth formula for rapid-turn inflation and shows the leading exponential non-Gaussianities cancel at tree level, but the loop-level extension is a real, self-admitted gap. read the letter →

arxiv 1908.11316 v1 pith:CNX2W5HB submitted 2019-08-29 hep-th astro-ph.CO

Theodor Bjorkmo , Ricardo Z. Ferreira , M.C. David Marsh This is my paper · ORCID

classification hep-thastro-ph.CO

keywords rapid-turninflationhyperinflationnon-Gaussianityin-informalismnestedcommutatorsimaginaryspeedofsoundtrispectrumperturbativecontrol

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper defends a class of inflationary models---rapid-turn inflation in negatively curved field spaces---against a recent argument that they produce 'hyper-large' non-Gaussianities and lose perturbative control. The authors claim that the apparently dominant, exponentially large contributions to the trispectrum and higher-point correlators cancel exactly once the nested commutators of the in-in formalism are handled correctly. After the cancellation, the connected $n$-point functions satisfy $\langle\zeta^n\rangle_c/\langle\zeta^2\rangle^{n-1}\sim 1$, so $g_{NL}\sim 1$ rather than $e^{4x}$ with the growth factor $x$, and perturbative control is maintained. The paper also supplies the first analytic WKB solution for the growth of perturbations in two-field rapid-turn models, which matches numerical results and identifies the exponent $x=(2-\sqrt{3+\xi})\pi\omega/2$. If correct, rapid-turn inflation remains an observationally viable candidate theory of the early universe.

What carries the argument

The carrying mechanism is the nested-commutator structure of the in-in (closed-time-path) formalism for correlators, together with a two-component mode-function ansatz. In the in-in expression for an $n$-point function, any nonzero term must have at least one operator on each commutator's left side Wick-contracted with an operator on its right side; terms without such cross-contractions vanish. A cross-contraction brings in the imaginary part of a product of mode functions, and because the mode functions are written as $\zeta_i(\tau)=f_i(\tau)e^x+i g_i(\tau)e^{-x}$ with real $f_i,g_i$, each imaginary part carries one power of the decaying component, $\sim e^{-x}$, instead of the growing $e^x$. The leading $e^{(4n-6)x}$ pieces in the two orderings of each commutator are therefore equal real quantities and cancel exactly, leaving the suppressed imaginary pieces and yielding $\langle\zeta^n\rangle_c/\langle\zeta^2\rangle^{n-1}\sim 1$. A secondary piece of machinery is the WKB evaluation of the two-field mode-function integral, which yields the closed-form growth exponent used to fix the normalisation and bound the turn rate.

What would settle it

Compute the tree-level four-point function in the full two-field theory beyond the leading WKB approximation, using the exact numerical mode functions, and check whether $g_{NL}$ scales as $e^{4x}$; if a time-dependent relative phase appears in the mode functions, the leading $e^{10x}$ term survives and the scaling would be visible. Equivalently, a direct numerical in-in evaluation of the trispectrum for a representative rapid-turn hyperinflation model would settle whether the cancellation is exact.

Watch

Extended reading notes

Core claim

The central discovery is that in rapid-turn inflation with $\xi<1$---the regime where the entropic mass lies below its critical value---the curvature perturbation $\zeta$ undergoes transient exponential growth near horizon crossing, characterised by a large parameter $x$, yet the higher-order correlators are not exponentially enhanced. Writing the mode functions as $\zeta_i(\tau)=f_i(\tau)e^x+i g_i(\tau)e^{-x}$ with real $f_i,g_i$, the naively leading $e^{10x}$ term in the four-point function from two cubic-interaction insertions is the real part of identical products of mode functions in the two orderings of each commutator, and it cancels exactly. Each nested commutator forces at least one Wick contraction between its left and right sides, producing a factor of the imaginary part of a product of mode functions, which scales as $e^{-x}$; with $n-2$ insertions this turns the naive $\alpha^{2n-2}e^{(4n-6)x}$ into $\alpha^{2n-2}e^{(2n-2)x}$, matching the denominator $\langle\zeta^2\rangle^{n-1}$ so that the ratio is of order one. The paper states this explicitly for the four-point function ($g_{NL}\sim 1$) and for the general $n$-point correlator. In addition, a WKB computation of the two-field linear system gives the analytic growth $\ln\gamma^2\approx(2-\sqrt{3+\xi})\pi\omega$, in good agreement with numerical results and consistent with the imaginary-speed-of-sound effective field theory.

Load-bearing premise

The load-bearing premise is that each mode function is exactly a growing real piece plus a decaying imaginary piece with no time-dependent relative phase between them; if that relative phase rotates in time, the leading exponential terms in the two commutator orderings would no longer coincide and the exact cancellation would fail.

Editorial extensions

If this is right

The connected four-point function is not exponentially amplified: $g_{NL}\sim 1$, far below current constraints $g_{NL}\lesssim 10^4\text{--}10^6$.
For every $n$, the ratio $\langle\zeta^n\rangle_c/\langle\zeta^2\rangle^{n-1}$ is of order one, so the perturbative expansion in $\zeta$ is under control despite the exponential amplification of the power-spectrum normalisation factor.
Each insertion of the cubic interaction contributes a factor of order $\alpha e^x\simeq \sqrt{P_\zeta}\ll 1$, so loop corrections are not expected to reintroduce exponential enhancement.
The analytic WKB solution gives the first closed-form expression for the perturbation growth in the two-field rapid-turn class and identifies the parameter $x=(2-\sqrt{3+\xi})\pi\omega/2$.
Combining the growth formula with power-spectrum normalisation and reheating requirements bounds the turn rate, e.g. $\omega\lesssim 96$ for hyperinflation with $\xi=-1$, leaving a large viable parameter space.

Reading between the lines

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

The cancellation mechanism is generic: any inflationary model whose mode functions are a growing-plus-decaying sum with fixed relative phase should show the same suppression, so estimates of non-Gaussianity that ignore commutator nesting will systematically overestimate the signal in transient-instability models.
A natural next step is a numerical in-in computation of the tree-level trispectrum in the full two-field theory, without the effective single-field description, to verify that $g_{NL}$ does not grow as $e^{4x}$; this would test the mode-function ansatz beyond leading WKB order.
If the cancellation persists at loop level, as the paper's power counting suggests, then the practical constraint on rapid-turn models shifts from non-Gaussianity to the power-spectrum amplitude and reheating, which already gives $\omega\lesssim 96$ in hyperinflation.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Let me get straight to the point. This paper gives the first analytic handle on the growth of ζ in rapid-turn inflation with ξ<1 and, more importantly, argues that the exponentially large non-Gaussianities found in [12] cancel at leading order because of the nested commutator structure of the in-in correlators. The WKB computation in Section 4 is real work: the growth integral (4.8)-(4.9) is evaluated in closed form, the resulting amplitude x(ω,ξ)=(2−√(3+ξ))πω/2 is new, and it matches the numerics of Mizuno & Mukohyama to better than a percent (lnγ∝0.920ω vs 0.924ω). That alone is worth a cite. The four-point cancellation in Section 3.1 is explicit and clean: the e^{10x} pieces are real symmetric products and the two orderings in the commutator cancel; only e^{6x} survives. The analogy to the occupation-number cancellation in gauge-field inflation is apt and makes the physics believable.

Soft spots, in proportion. The general-n argument in Section 3.2 is iterative and schematic, not a full derivation; the claim that every surviving term contains at least one cross-contraction per commutator is plausible but is asserted more than proven. The bigger gap is loop corrections. The authors restrict the mode function to ζ = f(τ)e^x + i g(τ)e^{-x} with no time-dependent relative phase, and footnote 2 concedes this. The stress-test note is right that this is harmless at tree level — one can rotate the growing-basis coefficient real and absorb the constant ratio into f(τ) — but the same is not shown for loop corrections to ζ. The loop discussion at the end of Section 3.2 is a power-counting heuristic; it does not control a phase that slips in through a one-loop correction. That is a specific, named gap, not a demonstrated failure.

I would send this to a serious referee. The central tree-level claim probably survives, and the WKB result is solid. Ask the referee to focus on the general-n commutator proof and the loop phase issue, and to verify the matched normalisation between α, e^{2x} and P_obs used in Eq. (2.17). The citation pattern is fine: the self-citations to [3,4] are the source of the attractor background, which is appropriate.

Who reads this: anyone working on non-geodesic multi-field inflation, effective field theories with imaginary sound speed, or the validity of perturbative inflation. It deserves a referee rather than a desk rejection.

Referee Report

2 major / 4 minor

Summary. The paper studies rapid-turn inflation models in negatively curved field spaces, where the curvature perturbation ζ undergoes a transient exponential growth before horizon crossing. The authors first provide an analytic WKB solution for the mode functions in the full two-field theory, obtaining the growth exponent x = (2−√(3+ξ))πω/2 and showing that it matches existing numerical results and the single-field EFT with imaginary speed of sound. They then address the previously claimed exponential enhancement of higher-order correlators. Using the in-in formalism, they argue that the nested commutator structure causes the naively leading exponentially large contributions to cancel: for the four-point function the e^{10x} and e^{8x} terms drop out (Eqs. 3.1–3.5), giving gNL ∼ 1, and for general n the connected ratio (3.14) is shown to be ∼1. The paper concludes that these models do not lose perturbative control and remain observationally viable.

Significance. If the conclusions hold, the paper resolves an apparent contradiction between rapid-turn inflation and observational bounds on non-Gaussianities, and it provides a useful analytic handle (the WKB growth formula) for a class of models that is otherwise studied numerically. The tree-level four-point cancellation is demonstrated explicitly with concrete mode-function scaling, and the WKB result is benchmarked against independent numerics in Figure 2, with quantitative agreement. The general-n argument is less complete, and the extension to loop corrections is only heuristic; this is the main gap in support of the paper's central perturbative-control claim.

major comments (2)

[Sec. 3.2, Eq. (3.2), footnote 2] The cancellation proof relies on the mode-function decomposition ζ(τ) = f(τ) e^x + i g(τ) e^{-x} with f, g real and no time-dependent relative phase. Footnote 2 explicitly concedes that time-dependent phases in one of the two terms are not considered and 'could show up in loop corrections to ζ'. The subsequent loop discussion at the end of Sec. 3.2 is only a power-counting heuristic and does not control this possibility. If a one-loop correction to ζ introduces a time-dependent relative phase θ(τ), the leading e^{10x} and e^{8x} cancellations in Eq. (3.1) are no longer exact; the uncancelled four-point amplitude would scale as α^6 e^{10x}, yielding gNL ∼ λ e^{4x}, which for x ∼ ω ∼ 90 is astronomically large. Since the abstract and Sec. 5 assert that there is 'no problem with perturbative control', the central claim is currently conditional on an unproven property of loop corrections. The manuscript should either prove the absence of such phases at loop level or restrict the no-loss-of-control claim to tree level.
[Sec. 3.2, after Eq. (3.13)] The general-n cancellation is established by an iterative argument rather than a complete derivation. In particular, the claim that every non-zero term in an n-nested commutator must contain contractions across each commutator, and that each such commutator contributes a factor e^{-2x} to the scaling, is stated without a fully rigorous combinatorial treatment when the interaction Hamiltonian contains derivatives (as in Eq. (3.3)) and when multiple operators within the same H_int may be contracted with each other. The four-point example is explicit, but the extension to arbitrary n rests on a schematic argument (Eqs. 3.9–3.14) that would benefit from a complete proof or a clearly stated conjecture with supporting evidence.

minor comments (4)

[Sec. 2.2, Eq. (2.15)] The mass term is written as '−H^2ω^2(ξ−1)σ2'; this should presumably be '−H^2ω^2(ξ−1)σ^2' for consistency with the rest of the equation.
[Sec. 3.1, before Eq. (3.2)] There is a typo: 'for simplicitly' should be 'for simplicity'.
[Sec. 2.3, Eq. (2.19)] The notation f^{(n−2)}_{NL} is introduced without definition; clarifying the placement of the superscript relative to the NL subscript (e.g., f_{NL}^{(n−2)}) would help the reader track the standard hierarchy of non-Gaussian shapes.
[Sec. 4, after Eq. (4.13)] The bound (4.13) uses a minimal reheating temperature T_min from BBN but does not explain how H_min is computed from T_min. A sentence with the standard relation H_min ∝ T_min^2/M_Pl and the relevant prefactor would make the estimate reproducible.

Circularity Check

0 steps flagged · score 2.0 of 10

No load-bearing circularity: the commutator-cancellation argument and the WKB growth computation are self-contained, with only minor non-load-bearing self-citations.

full rationale

The central claim of the paper, that the naively exponentially large non-Gaussianities cancel in the in-in formalism, is derived from the explicit algebraic structure of nested commutators in Eqs. (3.1)-(3.14), using the mode-function decomposition ζ_i(τ)=f_i(τ)e^x+i g_i(τ)e^{-x} introduced in Eq. (3.2). This decomposition is taken from the imaginary-speed-of-sound EFT mode function (2.16), attributable to references [10,12], which are not the present authors' work, and the cancellation does not rely on a fitted amplitude. The growth parameter x is not an input fitted to the correlators: it is computed from the quadratic action through the WKB integral (4.7), evaluated in Eqs. (4.10)-(4.12), and benchmarked against the independent numerical results of reference [2]. The paper does use self-citations for background material: [3,4] supply the rapid-turn attractor dynamics, and [36], which shares an author, provides an analogy from axially coupled gauge fields. However, none of these citations is load-bearing: the displayed calculations in Sections 3 and 4 stand on their own, and no equation reduces to its own input by construction. Footnote 2 does concede that time-dependent relative phases could appear in loop corrections to ζ and were not considered; this is an explicit limitation of the argument's domain, not a circular step, and it does not affect the tree-level cancellation argument. Overall, the paper's predictions are not statistically forced by fitted data, and the derivation chain is self-contained at the level claimed.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The central calculation depends on standard in-in/Wick technology, on the validity of the imaginary-speed EFT, and on two specific modeling choices: the mode-function phase structure and equal initial power for the two branches. None of these is fitted to the target result; the WKB growth is benchmarked against independent numerics from [2].

assumptions (6)

standard math The in-in formalism and Wick contractions correctly compute the connected correlators.
Used throughout Section 3 as the standard framework for inflationary correlation functions.
domain assumption The single-field EFT with imaginary speed of sound describes the rapid-turn two-field system in the relevant regime.
Used in Sections 2.3 and 3 to estimate correlators; Section 4 provides supporting evidence from the two-field action but does not prove the EFT at the level of interactions.
domain assumption Mode functions admit the decomposition zeta = f(tau) e^x + i g(tau) e^{-x} with f, g real and no time-dependent relative phase.
Footnotes and Eq. (3.2) in Section 3.1: this is the load-bearing premise for the exact cancellation of the leading exponential terms and for the e^{-2x} suppression per commutator.
domain assumption Hubble friction can be neglected during the transient growth phase.
Used to derive the WKB mode functions in Section 4; justified by large omega and exponential growth, and checked against numerics in Figure 2.
domain assumption The zeta_+ and zeta_- branches have roughly equal power at the start of the unstable phase.
Used after Eq. (4.11) to normalize the total power growth ln gamma^2; an order-one offset in the initial amplitudes would shift the growth by a constant.
domain assumption Derivative interactions do not alter the exponential scaling or the commutator cancellation.
Stated in Section 3.1; derivatives are dropped when estimating the correlator scaling, with the claim that they do not affect the exponential factors.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Mild Non-Gaussianities under Perturbative Control from Rapid-Turn Inflation Models." pith.science (2026). https://pith.science/paper/CNX2W5HB

@misc{pith2026190811316,
  author       = {Pith},
  title        = {Pith review of: Mild Non-Gaussianities under Perturbative Control from Rapid-Turn Inflation Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CNX2W5HB}},
  note         = {Machine review of arXiv:1908.11316}
}

abstract

Inflation can be supported in very steep potentials if it is generated by rapidly turning fields, which can be natural in negatively curved field spaces. The curvature perturbation, $\zeta$, of these models undergoes an exponential, transient amplification around the time of horizon crossing, but can still be compatible with observations at the level of the power spectrum. However, a recent analysis (based on a proposed single-field effective theory with an imaginary speed of sound) found that the trispectrum and other higher-order, non-Gaussian correlators also undergo similar exponential enhancements. This arguably leads to `hyper-large' non-Gaussianities in stark conflict with observations, and even to the loss of perturbative control of the calculations. In this paper, we provide the first analytic solution of the growth of the perturbations in two-field rapid-turn models, and find it in good agreement with previous numerical and single-field EFT estimates. We also show that the nested structure of commutators of the in-in formalism has subtle and crucial consequences: accounting for these commutators, we show analytically that the naively leading-order piece (which indeed is exponentially large) cancels exactly in all relevant correlators. The remaining non-Gaussianities of these models are modest, and there is no problem with perturbative control from the exponential enhancement of $\zeta$. Thus, rapid-turn inflation with negatively curved field spaces remains a viable and interesting class of candidate theories of the early universe.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Pushing the Primordial Frontier: Exact Linear Solutions in Multifield Inflation
astro-ph.CO 2026-06 unverdicted novelty 8.0 of 10

Exact analytic solutions for coupled linear perturbations in two-field inflation provide a closed-form primordial power spectrum that interpolates weak, strong, light, and heavy field regimes.
Searching for Folded Primordial Non-Gaussianity with Galaxy Surveys
astro-ph.CO 2026-07 conditional novelty 7.0 of 10

The galaxy bispectrum, not the power spectrum, is the main probe of folded primordial non-Gaussianity; narrow folded signals are washed out by Fourier-space binning.
Pushing the Primordial Frontier: Cosmological Collider Signatures at Strong Mixing
astro-ph.CO 2026-07 conditional novelty 7.0 of 10

Exact analytic squeezed-limit bispectra for strongly mixed two-field inflation, nonperturbative in the curvature-isocurvature mixing λ.

Reference graph

Works this paper leans on

42 extracted references · 2 canonical work pages · cited by 3 Pith papers

[12]

Fumagalli, S

J. Fumagalli, S. Garcia-Saenz, L. Pinol, S. Renaux-Petel and J. Ronayne,Hyper non-Gaussianities in inﬂation with strongly non-geodesic motion, 1902.03221

arXiv 1902
[1]

A. R. Brown,Hyperbolic Inﬂation, Phys. Rev. Lett.121 (2018) 251601, [1705.03023]

arXiv 2018
[2]

Mizuno and S

S. Mizuno and S. Mukohyama,Primordial perturbations from inﬂation with a hyperbolic ﬁeld-space, Phys. Rev. D96 (2017) 103533, [1707.05125]

arXiv 2017
[3]

Bjorkmo and M

T. Bjorkmo and M. C. D. Marsh,Hyperinﬂation generalised: from its attractor mechanism to its tension with the ‘swampland conditions’, JHEP 04 (2019) 172, [1901.08603]

arXiv 2019
[4]

Bjorkmo,Rapid-Turn Inﬂationary Attractors, Phys

T. Bjorkmo,Rapid-Turn Inﬂationary Attractors, Phys. Rev. Lett.122 (2019) 251301, [1902.10529]

arXiv 2019
[5]

Christodoulidis, D

P. Christodoulidis, D. Roest and E. I. Sfakianakis,Angular inﬂation in multi-ﬁeld α-attractors, 1803.09841

arXiv
[6]

Cremonini, Z

S. Cremonini, Z. Lalak and K. Turzynski,Strongly Coupled Perturbations in Two-Field Inﬂationary Models, JCAP 1103 (2011) 016, [1010.3021]

arXiv 2011
[7]

Renaux-Petel and K

S. Renaux-Petel and K. Turzyński,Geometrical Destabilization of Inﬂation, Phys. Rev. Lett. 117 (2016) 141301, [1510.01281]

arXiv 2016

Show all 42 references

[8]

Renaux-Petel, K

S. Renaux-Petel, K. Turzyński and V. Vennin,Geometrical destabilization, premature end of inﬂation and Bayesian model selection, JCAP 1711 (2017) 006, [1706.01835]

2017 arXiv
[9]

Garcia-Saenz, S

S. Garcia-Saenz, S. Renaux-Petel and J. Ronayne,Primordial ﬂuctuations and non-Gaussianities in sidetracked inﬂation, JCAP 1807 (2018) 057, [1804.11279]

2018 arXiv
[10]

Garcia-Saenz and S

S. Garcia-Saenz and S. Renaux-Petel,Flattened non-Gaussianities from the eﬀective ﬁeld theory of inﬂation with imaginary speed of sound, JCAP 1811 (2018) 005, [1805.12563]

2018 arXiv
[11]

Grocholski, M

O. Grocholski, M. Kalinowski, M. Kolanowski, S. Renaux-Petel, K. Turzyński and V. Vennin,On backreaction eﬀects in geometrical destabilisation of inﬂation, 1901.10468

1901 arXiv
[13]

Christodoulidis, D

P. Christodoulidis, D. Roest and E. I. Sfakianakis,Scaling attractors in multi-ﬁeld inﬂation, 1903.06116

1903 arXiv
[14]

Christodoulidis, D

P. Christodoulidis, D. Roest and E. I. Sfakianakis,Attractors, Bifurcations and Curvature in Multi-ﬁeld Inﬂation, 1903.03513

1903 arXiv
[15]

Bravo, G

R. Bravo, G. A. Palma and S. Riquelme,A Tip for Landscape Riders: Multi-Field Inﬂation Can Fulﬁll the Swampland Distance Conjecture, 1906.05772

1906 arXiv
[16]

D. A. Easson, R. Gregory, D. F. Mota, G. Tasinato and I. Zavala,Spinﬂation, JCAP 0802 (2008) 010, [0709.2666]

2008 arXiv
[17]

Achucarro, J.-O

A. Achucarro, J.-O. Gong, S. Hardeman, G. A. Palma and S. P. Patil,Mass hierarchies and non-decoupling in multi-scalar ﬁeld dynamics, Phys. Rev. D84 (2011) 043502, [1005.3848]

2011 arXiv
[18]

Achucarro, J.-O

A. Achucarro, J.-O. Gong, S. Hardeman, G. A. Palma and S. P. Patil,Features of heavy physics in the CMB power spectrum, JCAP 1101 (2011) 030, [1010.3693]

2011 arXiv
[19]

Achucarro, J.-O

A. Achucarro, J.-O. Gong, S. Hardeman, G. A. Palma and S. P. Patil,Eﬀective theories of single ﬁeld inﬂation when heavy ﬁelds matter, JHEP 05 (2012) 066, [1201.6342]

2012 arXiv
[20]

Achucarro, V

A. Achucarro, V. Atal, S. Cespedes, J.-O. Gong, G. A. Palma and S. P. Patil,Heavy ﬁelds, reduced speeds of sound and decoupling during inﬂation, Phys. Rev. D86 (2012) 121301, [1205.0710]. – 17 –

2012 arXiv
[21]

Cespedes, V

S. Cespedes, V. Atal and G. A. Palma,On the importance of heavy ﬁelds during inﬂation, JCAP 1205 (2012) 008, [1201.4848]

2012 arXiv
[22]

Hetz and G

A. Hetz and G. A. Palma,Sound Speed of Primordial Fluctuations in Supergravity Inﬂation, Phys. Rev. Lett.117 (2016) 101301, [1601.05457]

2016 arXiv
[23]

X. Chen, G. A. Palma, W. Riquelme, B. Scheihing Hitschfeld and S. Sypsas,Landscape tomography through primordial non-Gaussianity, Phys. Rev. D98 (2018) 083528, [1804.07315]

2018 arXiv
[24]

X. Chen, G. A. Palma, B. Scheihing Hitschfeld and S. Sypsas,Reconstructing the Inﬂationary Landscape with Cosmological Data, Phys. Rev. Lett.121 (2018) 161302, [1806.05202]

2018 arXiv
[25]

Aragam, S

V. Aragam, S. Paban and R. Rosati,Multi-ﬁeld Inﬂation in High-Slope Potentials, 1905.07495

1905 arXiv
[26]

Garcia-Saenz, L

S. Garcia-Saenz, L. Pinol and S. Renaux-Petel,Revisiting non-Gaussianity in multiﬁeld inﬂation with curved ﬁeld space, 1907.10403

1907 arXiv
[27]

Achucarro, G

A. Achucarro, G. Palma, D.-G. Wang and Y. Welling,Origin of ultra-light ﬁelds during inﬂation and their suppressed non-Gaussianity, 1908.06956

1908 arXiv
[28]

Achucarro, E

A. Achucarro, E. J. Copeland, O. Iarygina, G. A. Palma, D.-G. Wang and Y. Welling, Shift-Symmetric Orbital Inﬂation: single ﬁeld or multi-ﬁeld?, 1901.03657

1901 arXiv
[29]

Chakraborty, R

D. Chakraborty, R. Chiovoloni, O. Loaiza-Brito, G. Niz and I. Zavala,Fat Inﬂatons, Large Turns, and theη-problem, 1908.09797

1908 arXiv
[30]

Achucarro and G

A. Achucarro and G. A. Palma,The string swampland constraints require multi-ﬁeld inﬂation, 1807.04390

arXiv
[31]

D. H. Lyth and A. Riotto,Particle physics models of inﬂation and the cosmological density perturbation, Phys. Rept. 314 (1999) 1–146, [hep-ph/9807278]

1999 arXiv
[32]

Baumann and L

D. Baumann and L. McAllister,Inﬂation and String Theory. Cambridge Monographs on Mathematical Physics. Cambridge University Press, 2015, 10.1017/CBO9781316105733

2015 doi
[33]

Cicoli, V

M. Cicoli, V. Guidetti, F. G. Pedro and G. P. Vacca,A geometrical instability for ultra-light ﬁelds during inﬂation?, JCAP 1812 (2018) 037, [1807.03818]

2018 arXiv
[34]

Ooguri and C

H. Ooguri and C. Vafa,On the Geometry of the String Landscape and the Swampland, Nucl. Phys. B766 (2007) 21–33, [hep-th/0605264]

2007 arXiv
[35]

Cicoli, V

M. Cicoli, V. Guidetti and F. G. Pedro,Geometrical Destabilisation of Ultra-Light Axions in String Inﬂation, JCAP 1905 (2019) 046, [1903.01497]

2019 arXiv
[36]

R. Z. Ferreira, J. Ganc, J. Noreña and M. S. Sloth,On the validity of the perturbative description of axions during inﬂation, JCAP 1604 (2016) 039, [1512.06116]

2016 arXiv
[37]

Akrami et al.,Planck 2018 results

Planck collaboration, Y. Akrami et al.,Planck 2018 results. IX. Constraints on primordial non-Gaussianity, 1905.05697

2018 arXiv
[38]

P. D. Meerburg et al.,Primordial Non-Gaussianity, 1903.04409

1903 arXiv
[39]

Groot Nibbelink and B

S. Groot Nibbelink and B. J. W. van Tent,Scalar perturbations during multiple ﬁeld slow-roll inﬂation, Class. Quant. Grav.19 (2002) 613–640, [hep-ph/0107272]

2002 arXiv
[40]

Gordon, D

C. Gordon, D. Wands, B. A. Bassett and R. Maartens,Adiabatic and entropy perturbations from inﬂation, Phys. Rev. D63 (2001) 023506, [astro-ph/0009131]. – 18 –

2001 arXiv
[41]

Sasaki and E

M. Sasaki and E. D. Stewart,A General analytic formula for the spectral index of the density perturbations produced during inﬂation, Prog. Theor. Phys.95 (1996) 71–78, [astro-ph/9507001]

1996 arXiv
[42]

Langlois and S

D. Langlois and S. Renaux-Petel,Perturbations in generalized multi-ﬁeld inﬂation, JCAP 0804 (2008) 017, [0801.1085]. – 19 –

2008 arXiv

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [12]

Fumagalli, S

J. Fumagalli, S. Garcia-Saenz, L. Pinol, S. Renaux-Petel and J. Ronayne,Hyper non-Gaussianities in inﬂation with strongly non-geodesic motion, 1902.03221

arXiv 1902

[2] [1]

A. R. Brown,Hyperbolic Inﬂation, Phys. Rev. Lett.121 (2018) 251601, [1705.03023]

arXiv 2018

[3] [2]

Mizuno and S

S. Mizuno and S. Mukohyama,Primordial perturbations from inﬂation with a hyperbolic ﬁeld-space, Phys. Rev. D96 (2017) 103533, [1707.05125]

arXiv 2017

[4] [3]

Bjorkmo and M

T. Bjorkmo and M. C. D. Marsh,Hyperinﬂation generalised: from its attractor mechanism to its tension with the ‘swampland conditions’, JHEP 04 (2019) 172, [1901.08603]

arXiv 2019

[5] [4]

Bjorkmo,Rapid-Turn Inﬂationary Attractors, Phys

T. Bjorkmo,Rapid-Turn Inﬂationary Attractors, Phys. Rev. Lett.122 (2019) 251301, [1902.10529]

arXiv 2019

[6] [5]

Christodoulidis, D

P. Christodoulidis, D. Roest and E. I. Sfakianakis,Angular inﬂation in multi-ﬁeld α-attractors, 1803.09841

arXiv

[7] [6]

Cremonini, Z

S. Cremonini, Z. Lalak and K. Turzynski,Strongly Coupled Perturbations in Two-Field Inﬂationary Models, JCAP 1103 (2011) 016, [1010.3021]

arXiv 2011

[8] [7]

Renaux-Petel and K

S. Renaux-Petel and K. Turzyński,Geometrical Destabilization of Inﬂation, Phys. Rev. Lett. 117 (2016) 141301, [1510.01281]

arXiv 2016