Pith. sign in

REVIEW 3 major objections 5 minor 30 references

Recent two-loop divergences in on-shell EFT RG functions are spurious: they come from omitted non-minimal source terms, and restoring them restores finiteness.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-03 16:12 UTC pith:BS4DALZW

load-bearing objection Thomsen gives a convincing fix for the spurious two-loop on-shell EFT divergences by adding non-minimal source terms; the all-orders claim still leans on an unproven completeness assumption. the 3 major comments →

arxiv 2512.14260 v2 pith:BS4DALZW submitted 2025-12-16 hep-ph hep-th

On the Renormalization Group in EFTs: On-Shell Bases, Ambiguities, and Divergences

Anders Eller Thomsen This is my paper
classification hep-ph hep-th
keywords renormalization groupeffective field theoryon-shell basisnon-minimal source termsRG-finitenessflavor rotationsfield redefinitionsvacuum functional
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

Recent two-loop calculations of renormalization-group (RG) functions in effective field theories using 'on-shell' operator bases have turned up unphysical divergences, even in theories with no flavor structure. This paper argues those divergences are an artifact of a truncated formulation: converting the theory to an on-shell basis by field redefinitions silently drops non-minimal source terms from the vacuum functional, so the on-shell Lagrangian no longer renormalizes all Green's functions. Restoring those terms—treating the redundant EOM operators as couplings of new source terms—fully renormalizes the vacuum functional and makes the on-shell RG functions finite, with any leftover divergence amounting to an unphysical flavor rotation. The paper also shows that on-shell beta functions are intrinsically ambiguous along flavor directions and proposes a physical coupling space, the quotient by the flavor group, where the RG flow would be unique. The stakes are practical: comparisons between different beta-function calculations must account for these flavor-rotation ambiguities.

Core claim

The central claim is that the puzzling two-loop poles in on-shell EFT RG functions are not a failure of the renormalization group but a consequence of incomplete renormalization. When a theory is mapped from an off-shell (Green's) basis to an on-shell basis by field redefinitions, the source term J·η becomes J·(η + r Q(η)); the new r-dependent terms are non-minimal source terms (NMSTs). They can be dropped for S-matrix elements, but not for renormalizing the vacuum functional. Keeping them defines the full on-shell formulation, which is equivalent to the off-shell formulation and separates on-shell couplings from redundant ones. The recursive formulas for the beta functions and field anomalo

What carries the argument

The central object is the full on-shell vacuum functional, W'on[g,r,J] = -i log ∫ Dη exp(i(S_on[η,g] + ∫ J·(η + r Q(η)))), in which the redundant operators eliminated by the on-shell projection are not discarded but kept as non-minimal source terms with couplings r. The r's enlarge the on-shell space to Von × Vred, which is isomorphic to the off-shell space, so the formulation is invertible and fully renormalizable. The load-bearing formulas are the recursive pole relations for β_on and γ_on derived from RG invariance of the bare Lagrangian, now including β_red ∂_r terms; together with the flavor Ward identity these yield the RG-finiteness conditions. A second piece of machinery is the flavo

Load-bearing premise

The argument rests on the expectation, stated in footnote 22, that every counterterm outside the on-shell basis—including operators with multiple source insertions—can be shifted into δr counterterms or eliminated by a field redefinition η→η+f(η,J); if any such multi-source counterterm resisted elimination, the on-shell vacuum functional would not be fully renormalized and the claimed RG-finiteness would fail.

What would settle it

A three-loop full on-shell calculation exposing a pole in a beta function or anomalous dimension that is not proportional to a flavor rotation acting on the couplings, or an explicit counterterm with two source insertions that cannot be absorbed by any field redefinition η→η+f(η,J), would falsify the claim that including NMSTs restores RG-finiteness.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • The two-loop pole anomalies reported in on-shell EFT RG functions are predicted to cancel in a full on-shell calculation; no new physics or modification of the RG is required.
  • The finite parts of standard on-shell beta functions remain valid as obtained from truncated calculations; the missing NMST contributions affect the pole parts and the flow of Green's functions, not the running of S-matrix observables.
  • Any residual divergence in a correct full on-shell RG function must lie in the flavor-rotation direction; a divergence outside that direction signals an error or an incomplete counterterm sector.
  • Different on-shell beta-function calculations can legitimately differ by flavor rotations, through counterterm choices, projection choices, or embedding choices; the unambiguous flow lives on the quotient Von/GF.
  • The 't Hooft consistency relations for on-shell beta functions hold only modulo flavor rotations in the truncated formulation; restoring NMSTs restores the standard recursive pole structure in the full formulation.

Where Pith is reading between the lines

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

  • A testable extension: at three loops, a consistent full on-shell calculation should produce no non-flavor pole; flavor-type poles could appear at two loops in theories where the redundant-coupling contribution to the one-loop wave-function renormalization is non-Hermitian, which the paper identifies as the reason two-loop divergences appear earlier than in renormalizable theories.
  • An automated-tool implication: beta functions from different EFT codes should agree after projection onto flavor invariants or after fixing a flavor gauge; raw matrix-valued beta functions need not match. This gives a practical cross-check rule.
  • The quotient construction suggests a possible definition of a scheme-independent physical beta function; if a practical parametrization of Von/GF could be found (e.g., via basis invariants), RG running could be formulated without any flavor ambiguity.
  • The argument may extend to matching calculations: any field redefinition that removes composite operators from sources should keep the corresponding redundant couplings if one requires a fully renormalized vacuum functional; otherwise similar spurious poles could appear in matching coefficients.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. This paper addresses the recent observation that two-loop RG functions in EFTs computed in on-shell operator bases can carry spurious 1/ε poles. The author's proposal is that these divergences are artifacts of truncating the vacuum functional: passing from an off-shell basis to an on-shell basis by field redefinitions introduces non-minimal source terms (NMSTs), and counterterms for those NMSTs are required to renormalize all Green's functions. The paper introduces a "full on-shell formulation" with redundant couplings r_α, derives the corresponding Callan–Symanzik equation, and argues that the full on-shell RG functions are RG-finite, with any residual divergence generating only a flavor rotation. Two examples are worked explicitly: the dimension-six scalar O(n) model and the dimension-five νSMEFT. In the scalar case, the previously found two-loop divergence in γ_ϕ is cancelled by a β_red ∂_r Z contribution. In the νSMEFT case, the Hermitian divergence in γ_ℓ is cancelled and the remaining anti-Hermitian divergence is shown to satisfy the RG-finiteness condition with the Yukawa β-function. The paper also develops a geometric picture of the flavor-group ambiguities and proposes a physical coupling space V_on/G_F.

Significance. If the central claim is correct, the paper resolves a genuine puzzle in the EFT literature: it explains why on-shell two-loop RG functions can diverge even when the underlying theory is healthy, and it shows that the finite part of the on-shell β-functions is unaffected by the missing NMSTs. The two explicit examples are convincing as demonstrations of the mechanism, and the decomposition into Hermitian and anti-Hermitian pieces in Eqs. (5.19)–(5.21) is particularly clear. The geometric discussion in Section 6 is a useful conceptual framework, even if its practical utility remains open. The paper is also honest in flagging the main unproven step as an expectation. However, the all-orders statement of RG-finiteness is stronger than what the manuscript proves, and the gap is load-bearing for the advertised conclusion.

major comments (3)
  1. [Sec. 4.2, footnote 22] The central renormalizability claim rests on the assertion that every counterterm operator not contained in the on-shell basis, including multi-source terms, can be shifted into δr counterterms or removed by a source-dependent field redefinition η → η + f(η, J). This is presented only as "We expect... " and no proof is supplied. This is load-bearing: if even one multi-source counterterm resisted elimination, the full on-shell vacuum functional would not be renormalized by the counterterms in Eqs. (4.11)–(4.15), the Callan–Symanzik equation (4.16) would not describe a finite flow, and the residual divergences would not need to be pure flavor rotations. The two examples in Section 5 involve only single-source NMSTs, so they do not test the completeness assumption. The all-orders claim should therefore be either proven, or explicitly restricted to the cases where the counterterm structure i
  2. [Sec. 4.2, after Eq. (4.16)] The conclusion "Based on this equivalence, we expect both sets of RG functions to generate finite RG flow" is not a proof. The equivalence between the off-shell and full on-shell formulations is plausible at the level of renormalized Green's functions, but the derivation of the RG functions requires that the bare coupling/source mapping is compatible with renormalization to all orders. The paper does not show that the source-dependent field redefinitions of footnote 22 preserve the loop counting or the pole structure needed for Eqs. (4.13)–(4.15). The explicit examples verify the mechanism only at one- and two-loop order for a single NMST coupling. Thus the paper's central claim, as stated in the abstract and conclusion, overreaches the evidence provided. I would recommend either supplying the missing proof or weakening the conclusions to "the full on-shell formulation is RG-finite at le
  3. [Sec. 5, Eqs. (5.16)–(5.21)] The νSMEFT example is the most substantive test, but it also illustrates the incompleteness concern. The cancellation of the Hermitian part of γ_ℓ relies on the specific form of Z_ℓ^{(1)} in Eq. (5.19), which is computed for a single redundant coupling r_ℓ. If additional multi-source counterterms are present, they could contribute to γ_ℓ and to β^Y_ν through the β_red ∂_β terms in Eq. (4.13), and the simple relation (5.21) might no longer hold. The paper should state explicitly which classes of NMST counterterms are assumed absent or argued to be removable, and how the examples are representative of the general case.
minor comments (5)
  1. [Sec. 2.3, Eq. (2.16)] The phrase "The invalid assertion that all poles vanish" is a bit abrupt; consider "The generally unjustified assumption that all poles vanish".
  2. [Sec. 3.2, Eq. (3.14)] The notation Δγ(λ(g))·g is used before the fundamental vector field notation is introduced in Section 6. A one-sentence definition at first use would improve readability.
  3. [Sec. 5.1, Eq. (5.7)] The notation for the missing NMST contribution, written as a stop sign or strikethrough, is nonstandard. It would be clearer to define a symbol such as Δγ_NMST.
  4. [Sec. 6.3, Note 5] The discussion of the SM stabilizer H∘ is useful, but the relation of the anomalous U(1)_{B+L} to the arguments about the local RG may deserve a brief comment or citation, since the paper elsewhere relies on anomaly-free flavor symmetries.
  5. [Sec. 7, Conclusion] The conclusion states that "the β-functions of the on-shell couplings are shown to be independent of the redundant couplings" but the proof in Section 4.2 uses the same completeness assumption flagged above; the wording should be qualified.

Circularity Check

0 steps flagged

No significant circularity: the NMST cancellation is derived from off-shell counterterms and field-redefinition maps, not fitted to the target divergences; the flagged completeness assumption is a correctness risk, not a circular step.

full rationale

The central mechanism is not circular. In the O(n) example, the NMST contribution to the two-loop anomalous dimension is derived from the redundant-coupling counterterm δ(1)_1 r (Eq. 5.8) and the r-dependent wave-function renormalization (Eq. 5.10), both obtained from the off-shell counterterms of Ref. [11] via the explicit map (5.6). Equation (5.11) then cancels the published truncated-on-shell divergence (5.7) algebraically; no parameter is tuned to the target divergence. The νSMEFT example is likewise derived: δ(1)_1 r_ℓ (5.18) follows from the off-shell counterterms of Ref. [12] and the map (5.15), and Z(1)_{1,ℓ} (5.19) is a direct diagrammatic result; Eqs. (5.20)–(5.21) follow rather than being imposed. The reliance on Ref. [4] for the RG-finiteness criterion is a self-citation, but it is not circular: the criterion is independently supported by the flavor Ward identity (3.10) and is not an input to the NMST construction. The Matchete self-citation [22] is to a public tool used for diagram evaluation. The only substantive gap is the completeness expectation in footnote 22 — 'We expect that such terms can be eliminated...' — but this is an explicitly flagged unproven assumption about multi-source counterterms, not a reduction of the paper's conclusion to its inputs. It is a correctness risk, not a circularity. The derivation is self-contained against the off-shell formulation and the observed cancellation is genuinely nontrivial.

Axiom & Free-Parameter Ledger

0 free parameters · 5 axioms · 1 invented entities

The central claim is analytical, not numerical. There are no data fits or hand-tuned constants; the redundant couplings r_α are fixed by field-redefinition maps, not fitted to the target divergences. The main load-bearing assumptions are invariance of the path-integral measure under perturbative field redefinitions, equivalence between off-shell and full on-shell formulations, asserted completeness of the NMST counterterm sector (explicitly flagged in footnote 22), correctness of the external counterterms used in the examples, and adoption of the RG-finiteness criterion from Ref [4]. The NMSTs are formal bookkeeping terms with no independent physical handle.

axioms (5)
  • domain assumption Perturbative field redefinitions leave the path-integral measure invariant in dimensional regularization, so the vacuum functional is unchanged under η→Ξ(η,λ) (Note 1, Eq. 4.2).
    Used to equate the off-shell and full on-shell vacuum functionals; if the Jacobian contributed, the NMST framework would need modification.
  • domain assumption Physically equivalent on-shell points are exactly those related by a flavor rotation: [g']=[g] ⇔ g'∈G_F·g (Eq. 3.5), and V_off/∼ ≅ V_on/G_F.
    Underlies the entire ambiguity analysis and the definition of the physical coupling space; well-motivated by field-redefinition theorems but asserted rather than proven in the paper.
  • ad hoc to paper Any counterterm operator not in the on-shell basis can be shifted to a counterterm for the NMSTs; multi-source counterterms are removable by field redefinitions η→η+f(η,J) (Sec 4.2, footnote 22).
    Load-bearing for full on-shell renormalizability; explicitly stated as an expectation, not demonstrated.
  • domain assumption The one- and two-loop counterterms of Refs [11,12] used in the examples are correct.
    The example cancellations rely on external counterterm results and the Matchete package; a transcription error in either would propagate.
  • domain assumption The RG-finiteness criterion and recursive pole relations of Ref [4] apply to the full on-shell formulation.
    Central criterion used to interpret residual divergences as flavor rotations; adopted from prior work by the same author.
invented entities (1)
  • Non-minimal source terms (NMSTs) with redundant couplings r_α no independent evidence
    purpose: Restore full renormalizability of the vacuum functional in the on-shell basis and cancel spurious two-loop RG divergences, e.g., r1(ϕ·ϕ)ϕ in the O(n) model and r_ℓ H̃ N_R^c in the νSMEFT example.
    These are formal source terms, not new particles or forces; they have no observable handle and are invisible in S-matrix elements. They are introduced as bookkeeping for the renormalized Green's functions, so they are an invented formal device with no independent physical evidence.

pith-pipeline@v1.3.0-alltime-deepseek · 30814 in / 16205 out tokens · 129328 ms · 2026-08-03T16:12:26.894807+00:00 · methodology

0 comments
read the original abstract

Recent results for two-loop renormalization group (RG) functions in effective field theories exhibit unphysical divergences when calculated in an on-shell operator basis. We demonstrate that this can be understood to be a result of omitting non-minimal source terms in the renormalized vacuum functional, which are essential to maintaining renormalizability of and describing the RG flow of Green's functions in an on-shell framework. With the inclusion of the missing source terms, any remaining divergences are ambiguous, generating only unphysical RG flow directed along flavor rotations, and the RG functions are RG-finite. We carefully examine the role of flavor rotations in generating ambiguities in both on- and off-shell RG functions and explore the geometry of a physical coupling space.

Figures

Figures reproduced from arXiv: 2512.14260 by Anders Eller Thomsen.

Figure 1
Figure 1. Figure 1: Sketch of the RG flow in the space of off-shell couplings. The horizontal axis indicates the [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Illustration of the RG flow in full on-shell formulation. The horizontal axis indicates the [PITH_FULL_IMAGE:figures/full_fig_p021_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: The divergence of the NMST of the scalar O( [PITH_FULL_IMAGE:figures/full_fig_p025_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: The divergence of the NMST of the ℓ field in νSMEFT on the left is renormalized with a counterterm of the ordinary source term (right). The field source J is denoted with a cross on a gray background; the dots are renormalizable couplings (mass insertion and Yukawa interaction); and the external fermion line connects to the rest of the Green’s function, indicated by a hatched blob. in the off-shell Lagrang… view at source ↗
Figure 5
Figure 5. Figure 5: Sketch of the projection of the off-shell [PITH_FULL_IMAGE:figures/full_fig_p033_5.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

30 extracted references · 1 canonical work pages

  1. [1]

    ’t Hooft,Dimensional regularization and the renormalization group,Nucl

    G. ’t Hooft,Dimensional regularization and the renormalization group,Nucl. Phys. B61 (1973) 455–468

    1973

  2. [2]

    A. V. Bednyakov, A. F. Pikelner and V. N. Velizhanin,Three-loop SM beta-functions for matrix Yukawa couplings,Phys. Lett. B737(2014) 129–134, [1406.7171]

    Pith/arXiv arXiv 2014

  3. [3]

    Herren, L

    F. Herren, L. Mihaila and M. Steinhauser,Gauge and Yukawa coupling beta functions of two-Higgs-doublet models to three-loop order,Phys. Rev. D97(2018) 015016, [1712.06614]

    Pith/arXiv arXiv 2018

  4. [4]

    Herren and A

    F. Herren and A. E. Thomsen,On ambiguities and divergences in perturbative renormalization group functions,JHEP06(2021) 116, [2104.07037]

    Pith/arXiv arXiv 2021

  5. [5]

    C. G. Callan, Jr.,Broken scale invariance in scalar field theory,Phys. Rev. D2(1970) 1541–1547

    1970

  6. [6]

    Symanzik,Small distance behavior in field theory and power counting,Commun

    K. Symanzik,Small distance behavior in field theory and power counting,Commun. Math. Phys.18(1970) 227–246

    1970

  7. [7]

    Jack and H

    I. Jack and H. Osborn,Analogs for the c Theorem for Four-dimensional Renormalizable Field Theories,Nucl. Phys. B343(1990) 647–688

    1990

  8. [8]

    Fortin, B

    J.-F. Fortin, B. Grinstein and A. Stergiou,Limit Cycles and Conformal Invariance, JHEP01(2013) 184, [1208.3674]

    Pith/arXiv arXiv 2013

  9. [9]

    Jack and H

    I. Jack and H. Osborn,Constraints on RG Flow for Four Dimensional Quantum Field Theories,Nucl. Phys. B883(2014) 425–500, [1312.0428]. –39–

    Pith/arXiv arXiv 2014

  10. [10]

    E. E. Jenkins, A. V. Manohar, L. Naterop and J. Pag` es,Two loop renormalization of scalar theories using a geometric approach,JHEP02(2024) 131, [2310.19883]

    Pith/arXiv arXiv 2024

  11. [11]

    A. V. Manohar, J. Pag` es and J. Roosmale Nepveu,Field redefinitions and infinite field anomalous dimensions,JHEP05(2024) 018, [2402.08715]

    Pith/arXiv arXiv 2024

  12. [12]

    Zhang,Two-loop renormalization group equations in theνSMEFT,JHEP06(2025) 106, [2504.00792]

    D. Zhang,Two-loop renormalization group equations in theνSMEFT,JHEP06(2025) 106, [2504.00792]

    Pith/arXiv arXiv 2025

  13. [13]

    Naterop and P

    L. Naterop and P. Stoffer,Renormalization-group equations of the LEFT at two loops: dimension-six operators,2507.08926

    arXiv

  14. [14]

    Georgi,On-shell effective field theory,Nucl

    H. Georgi,On-shell effective field theory,Nucl. Phys. B361(1991) 339–350

    1991

  15. [15]

    Arzt,Reduced effective Lagrangians,Phys

    C. Arzt,Reduced effective Lagrangians,Phys. Lett. B342(1995) 189–195, [hep-ph/9304230]

    Pith/arXiv arXiv 1995

  16. [16]

    J. C. Criado and M. P´ erez-Victoria,Field redefinitions in effective theories at higher orders,JHEP03(2019) 038, [1811.09413]

    Pith/arXiv arXiv 2019

  17. [17]

    ’t Hooft and M

    G. ’t Hooft and M. J. G. Veltman,DIAGRAMMAR,NATO Sci. Ser. B4(1974) 177–322

    1974

  18. [18]

    M. B. Einhorn and J. Wudka,The Bases of Effective Field Theories,Nucl. Phys. B876 (2013) 556–574, [1307.0478]

    Pith/arXiv arXiv 2013

  19. [19]

    Buchmuller and D

    W. Buchmuller and D. Wyler,Effective Lagrangian Analysis of New Interactions and Flavor Conservation,Nucl. Phys. B268(1986) 621–653

    1986

  20. [20]

    M. B. Einhorn and J. Wudka,Effective beta functions for effective field theory,JHEP08 (2001) 025, [hep-ph/0105035]

    Pith/arXiv arXiv 2001

  21. [21]

    K. G. Chetyrkin,Combinatorics ofR-,R −1-, andR ∗-operations and asymptotic expansions of feynman integrals in the limit of large momenta and masses, 1701.08627

    Pith/arXiv arXiv

  22. [22]

    Fuentes-Mart ´ ın, M

    J. Fuentes-Mart ´ ın, M. K¨ onig, J. Pag` es, A. E. Thomsen and F. Wilsch,A proof of concept for matchete: an automated tool for matching effective theories,Eur. Phys. J. C83 (2023) 662, [2212.04510]

    Pith/arXiv arXiv 2023

  23. [23]

    G. F. Giudice, C. Grojean, A. Pomarol and R. Rattazzi,The Strongly-Interacting Light Higgs,JHEP06(2007) 045, [hep-ph/0703164]

    Pith/arXiv arXiv 2007

  24. [24]

    Keren-Zur,The local RG equation and chiral anomalies,JHEP09(2014) 011, [1406.0869]

    B. Keren-Zur,The local RG equation and chiral anomalies,JHEP09(2014) 011, [1406.0869]

    Pith/arXiv arXiv 2014

  25. [25]

    M. P. Bento, J. P. Silva and A. Trautner,The basis invariant flavor puzzle,JHEP01 (2024) 024, [2308.00019]

    Pith/arXiv arXiv 2024

  26. [26]

    Baume, B

    F. Baume, B. Keren-Zur, R. Rattazzi and L. Vitale,The local Callan-Symanzik equation: structure and applications,JHEP08(2014) 152, [1401.5983]

    Pith/arXiv arXiv 2014

  27. [27]

    Anselmi,A General Field-Covariant Formulation Of Quantum Field Theory,Eur

    D. Anselmi,A General Field-Covariant Formulation Of Quantum Field Theory,Eur. Phys. J. C73(2013) 2338, [1205.3279]

    Pith/arXiv arXiv 2013

  28. [28]

    Naterop and P

    L. Naterop and P. Stoffer,Renormalization-group equations of the LEFT at two loops: dimension-five effects,JHEP06(2025) 007, [2412.13251]

    Pith/arXiv arXiv 2025

  29. [29]

    M. J. Pflaum,Analytic and Geometric Study of Stratified Spaces. Springer Berlin, Heidelberg, 2001, 10.1007/3-540-45436-5

    work page doi:10.1007/3-540-45436-5 2001

  30. [30]

    Bredon,Introduction to compact transformation groups

    G. Bredon,Introduction to compact transformation groups. Academic Press, 1972. –40–

    1972