Comment on "Redundancy Channels in the Conformal Bootstrap" by S. R. Kousvos and A. Stergiou

Slava Rychkov

Comment on "Redundancy Channels in the Conformal Bootstrap" by S. R. Kousvos and A. Stergiou

T0 review · 0 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read This comment defends identifying operators that differ only by contact terms, such as phi^3 and d^2 phi at the Wilson-Fisher fixed point, as the same CFT operator with equal scaling dimensions.

desk verdict A short, plainly argued comment that correctly identifies the Kousvos–Stergiou dispute as terminological; the physics is sound and it deserves publication as a comment, even though it stops short of a worked two-scheme comparison. read the letter →

arxiv 2507.13070 v1 pith:PAJ453A6 submitted 2025-07-17 hep-th cond-mat.stat-mech

Slava Rychkov This is my paper · ORCID

classification hep-thcond-mat.stat-mech

keywords conformalfieldtheoryWilson-Fisherfixedpointepsilonexpansioncompositeoperatorsequationsofmotioncontacttermsrenormalizationgroupscalingdimensions

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

The comment defends a specific practice in perturbative conformal field theory: when two operators, such as $\phi^3$ and $\partial^2\phi$ at the Wilson-Fisher fixed point, are related by the equations of motion, their correlation functions agree at non-coincident points and differ only by contact terms. Since a CFT only defines correlation functions away from coincident points, the two operators should be identified and assigned the same scaling dimension. The comment argues that the criticized work's 'full way', which renormalizes the complete operator list including contact terms before dropping them, is unnecessary and that the disagreement is terminological rather than substantive. If the comment is right, the economical identification yields the correct CFT data with strictly less computation.

What carries the argument

The machinery is the equation-of-motion contact-term identification: at the Wilson-Fisher fixed point the EOM makes $\partial^2\phi$ proportional to $\phi^3$ up to delta-function terms, so away from coincident points there is only one independent operator. The argument combines this with the rule that CFT scaling dimensions are read from non-coincident correlation functions, which justifies working modulo contact terms before diagonalizing the renormalization mixing matrix. That is what allows the 'economical way' to operate with a minimal set of composite operators.

What would settle it

Compute a non-coincident correlation function involving $\phi^3$ at the Wilson-Fisher fixed point to subleading order in $\epsilon$ using the full mixing matrix that includes contact-term operators, and check whether the extracted anomalous dimension agrees with the one obtained by identifying $\phi^3$ with $\partial^2\phi$.

Watch

Extended reading notes

Core claim

The central claim is that CFT data in perturbative renormalization-group computations should be extracted from correlation functions at non-coincident points, so operators whose difference is a contact term are the same CFT operator. At the Wilson-Fisher fixed point, $\phi^3$ and $\partial^2\phi$ are related by the equation of motion and therefore have equal scaling dimensions in the CFT sense. The nonhomogeneous scale transformation of such operators is purely a contact-term effect, invisible to the CFT. The comment states that renormalizing the full mixing matrix including EOM operators, as done in the criticized work, and then discarding those operators cannot change the non-coincident correlators, so the difference between the two treatments is only a matter of terminology.

Load-bearing premise

The load-bearing premise is that CFT data lives entirely in non-coincident correlation functions, so contact terms can be discarded when computing scaling dimensions.

Editorial extensions

If this is right

At the Wilson-Fisher fixed point, $\phi^3$ and $\partial^2\phi$ carry the same scaling dimension in the CFT sense.
Renormalizing contact terms and then discarding them cannot change the CFT correlation functions at separated points.
The difference between the comment's treatment and the criticized treatment is terminological; both give the same CFT scaling dimensions.
The economical approach is sufficient for computing CFT data and avoids unnecessary operator renormalization.
The same identification applies to EOM-related operators in other RG flows, such as the conserved-current/quartic-operator pair discussed for the $O(N)$ fixed point.

Reading between the lines

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

A testable extension is to check explicitly at subleading order that contact-term renormalization does not feed back into the eigenvalues of the mixing matrix that control non-coincident correlators, since the comment leaves that demonstration implicit.
If the identification principle is right, conformal bootstrap equations that treat these operators as independent channels are carrying gauge degrees of freedom, and the physical spectrum is obtained by modding them out.
The same logic predicts that any EOM relation between operators of different classical dimensions yields equal CFT scaling dimensions at the fixed point, which is testable in multiscalar Wilson-Fisher fixed points.

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

Slava,

This is a compact comment, and it does what a good comment should: it states the disagreement crisply and takes a clear position. The substantive claim is that in a CFT, operators whose correlators differ by contact terms are the same operator, because CFT scaling dimensions are read off from non-coincident points. That is the standard understanding, and I think Rychkov is right to push back on framing it as a flaw in the original paper.

What is actually new here is the explicit contrast between the two procedures: the "full way" of renormalizing the entire mixing matrix, including EOM directions, and the "economical way" of identifying operators modulo contact terms before renormalizing. The comment argues these give identical CFT data. That is the heart of the matter, and the logic is sound: contact terms are delta-function supported, so they cannot affect the power-law scaling of non-coincident correlators. The reader's worry that contact terms could sneak into the mixing matrix and shift the eigenvalues that survive is excluded by locality. So the central argument holds.

The soft spot is that this is asserted rather than shown. An explicit one-loop check, comparing the two schemes in a simple case like the Wilson–Fisher fixed point and confirming the surviving eigenvalues match, would shut the door completely. The comment does not provide that. In a longer paper that might be a real gap, but for a comment responding to a published criticism, it is a minor omission. The references are appropriate, and there is no circularity—the earlier work is cited for context, not as evidence.

This is not a new result, and it introduces no new technique. Its value is as a clarification, and it should be useful to anyone working in epsilon expansions or perturbative CFT who has to navigate conflicting conventions. I would send it to a referee; it deserves a proper peer review, not a desk rejection. A referee may ask for the one-loop check, but even without it the comment is correct and worthwhile.

I would not cite it in my own work in the next year, but I would point a student to it when they ask about EOM operators and contact terms. Bring it to a reading group only if you are actively working in this corner of the bootstrap.

Net: send it back for review. It will be a fine, if modest, contribution.

Best,

[Your name]

Referee Report

0 major / 5 minor

Summary. This manuscript is a short comment responding to Kousvos and Stergiou's critique of Rychkov and Tan's epsilon-expansion computation of CFT data. The author defends the identification, modulo contact terms, of the composite operators φ^3 and ∂^2φ at the Wilson-Fisher fixed point, which leads to equal scaling dimensions in the CFT sense. The argument is that CFT correlation functions are defined at non-coincident points, so operators whose correlators differ only by contact terms—such as those proportional to the equations of motion—should be identified; Kousvos and Stergiou's full renormalization-matrix treatment retains these contact directions and assigns them formal "scaling dimensions" in a perturbative RG sense, which are not part of the CFT spectrum. The author concludes that the disagreement is terminological and that the more economical scheme is preferable.

Significance. If accepted, the comment clarifies a recurring point of confusion in perturbative CFT: the status of equation-of-motion operators in RG dimension calculations. It does not present new calculations, but it articulates a standard and internally consistent viewpoint, and it explicitly separates the CFT notion of scaling dimension from the formal RG eigenvalue. The argument is not circular: it relies on the definition of CFT correlators at separated points and on locality of contact terms. The main limitation is that the equivalence of the economical and full renormalization schemes is asserted rather than demonstrated with a concrete one-loop example, though the claim is standard and I do not regard the absence of such a check as a correctness gap.

minor comments (5)

[Economical-way paragraph] The guarantee that renormalizing modulo contact terms gives correct CFT scaling dimensions would be easier to evaluate with an explicit one-loop comparison for the φ^3/∂^2φ sector, or with a citation to a proof in the RG literature; without this, the reader must rely on the stated locality argument.
[Role-of-equations-of-motion paragraph] The premise "CFT only describes correlation functions away from coincident points" is an oversimplification when taken literally: contact terms carry Ward-identity and anomaly information in a CFT. Since the argument concerns only the extraction of scaling dimensions, I suggest reformulating the premise accordingly.
[Item 1 in the introductory list] There is a typo: "minimal subtaction" should read "minimal subtraction."
[Third paragraph] There is a typo: "one one independent operator" should read "one independent operator."
[References] Reference [4] cites the arXiv version; if the comment is published after Kousvos and Stergiou's paper appears in a journal, the reference should be updated.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the comment's definitional argument is self-contained and does not rely on self-citation as evidence.

full rationale

The paper is a comment that makes a terminological and definitional argument rather than a derivational prediction. Its central claim is that operators whose correlation functions differ only by contact terms should be identified in a CFT, because CFT data is defined from correlation functions at non-coincident points. This claim is justified by the stated definition of what a CFT operator is, not by fitting parameters, not by a derivation from prior results, and not by the cited prior work [1] as evidence. The self-citation to Rychkov and Tan only identifies the earlier work being defended; the defense itself rests on the definition of scaling dimensions via non-coincident correlators. There is no equation in the comment that reduces to its own input, no fitted input is renamed as a prediction, and no uniqueness theorem is imported from the authors. The reader-identified concern that contact terms might feed into renormalization mixing and affect non-coincident correlators is a possible correctness or completeness issue, but it is not an instance of circular reasoning. The comment is therefore not circular.

Assumptions & free parameters 0 free parameters · 2 assumptions · 0 invented entities

The comment relies on standard CFT and RG assumptions, not on new postulated entities. The two axioms above are domain-specific premises that are widely accepted; the dispute concerns their application, not their validity.

assumptions (2)

domain assumption CFT only describes correlation functions away from coincident points, so contact terms can be neglected when extracting CFT data.
Stated in the opening and throughout: 'CFT only describes the operators whose correlation functions are nonzero at non-coincident points.' This premise is essential for the identification of EOM-related operators.
domain assumption Operators related by equations of motion have correlation functions that differ only by contact terms, so they represent the same CFT operator at non-coincident points.
Invoked in the phi^3 and d^2 phi example: 'both phi^3 and d^2 phi are nontrivial, but they are related by the EOM, so away from coincident points there is again just one independent operator.'

how reviews work

0 comments

Cite this review

Pith. "Pith review of Comment on "Redundancy Channels in the Conformal Bootstrap" by S. R. Kousvos and A. Stergiou." pith.science (2026). https://pith.science/paper/PAJ453A6

@misc{pith2026250713070,
  author       = {Pith},
  title        = {Pith review of: Comment on "Redundancy Channels in the Conformal Bootstrap" by S. R. Kousvos and A. Stergiou},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PAJ453A6}},
  note         = {Machine review of arXiv:2507.13070}
}

read the original abstract

Recent work by Kousvos and Stergiou criticises our work with Zhong Ming Tan [arXiv:1505.00963]. The issue is CFT scaling dimension computations in perturbative Renormalization Group. We identified operators whose correlation functions differ by contact terms. This is allowed because CFT only describes correlation functions away from coincident points. They instead renormalize the contact terms, which are eventually dropped. Our way is not only correct, but preferable as it operates with the minimal set of quantities.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

4 extracted references · 2 canonical work pages

[1]

The ϵ-expansion from conformal field theory,

S. Rychkov and Z. M. Tan, “The ϵ-expansion from conformal field theory,” J. Phys. A 48 no. 29, (2015) 29FT01, arXiv:1505.00963 [hep-th]

arXiv 2015
[2]

A. N. Vasil’ev, The Field Theoretic Renormalization Group in Critical Behavior Theory and Stochastic Dynamics. Chapman and Hall/CRC, 1998

work page 1998
[3]

Kleinert and V

H. Kleinert and V. Schulte-Frohlinde, Critical Properties of ϕ4-Theories. World Scientific, 2001

work page 2001
[4]

Redundancy Channels in the Conformal Bootstrap,

S. R. Kousvos and A. Stergiou, “Redundancy Channels in the Conformal Bootstrap,” arXiv:2507.05338v1 [hep-th]. – 3 –

arXiv

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [1]

The ϵ-expansion from conformal field theory,

S. Rychkov and Z. M. Tan, “The ϵ-expansion from conformal field theory,” J. Phys. A 48 no. 29, (2015) 29FT01, arXiv:1505.00963 [hep-th]

arXiv 2015

[2] [2]

A. N. Vasil’ev, The Field Theoretic Renormalization Group in Critical Behavior Theory and Stochastic Dynamics. Chapman and Hall/CRC, 1998

work page 1998

[3] [3]

Kleinert and V

H. Kleinert and V. Schulte-Frohlinde, Critical Properties of ϕ4-Theories. World Scientific, 2001

work page 2001

[4] [4]

Redundancy Channels in the Conformal Bootstrap,

S. R. Kousvos and A. Stergiou, “Redundancy Channels in the Conformal Bootstrap,” arXiv:2507.05338v1 [hep-th]. – 3 –

arXiv