Pith. sign in

REVIEW 2 major objections 7 minor 5 references

A new look at Perturbation Theory in QFT and Resolvent Series

T0 review · 2 major / 7 minor · reviewed 2026-08-07 · deepseek-v4-flash

Pith's one-line read One resolvent identity underlies QFT perturbation theory.

desk verdict A solid pedagogical translation of perturbation theory into finite-dimensional resolvent language, with a real but fixable gap in the scattering matrix section. read the letter →

arxiv 2506.11572 v1 pith:O7GUT7W2 submitted 2025-06-13 math-ph math.MPquant-ph

classification math-phmath.MPquant-ph MSC 81Q1547A1081T18
keywords resolventseriesperturbationtheoryquantumfieldDysonscatteringmatrixFeynmandiagramsfinite-dimensionalmatriceseigenvalue
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

These notes argue that the elementary resolvent identity $$(A+B)^{-1}=\sum_{m=0}^\infty (-1)^m $A^{{-1}}$($BA^{{-1}}$)^m$$ is the central tool of perturbation theory: eigenvalue corrections, Dyson series, scattering amplitudes, and diagram sums all appear as particular cases of expanding the inverse of a sum of matrices. The paper works systematically with finite-dimensional matrices, setting aside infinite-particle systems and divergent integrals, on the principle that a formula that cannot be proved for finite matrices has little hope in infinite dimension. It attributes the ultraviolet and infrared divergences of quantum field theory to the naive application of the lemma with unbounded operators, where $\|B\|=\infty$ or $\|A^{-1}\|=\infty$, rather than to the formula itself. The payoff is a finite-matrix translation of the standard perturbation-theoretic apparatus in which each QFT object is obtained by substituting a resolvent expansion into a resolvent identity.

What carries the argument

The load-bearing object is Lemma 1.1, the resolvent series $$(A+B)^{-1}=\sum_{m=0}^\infty (-1)^m $A^{{-1}}$($BA^{{-1}}$)^m,$$ valid when $\|A^{-1/2}BA^{-1/2}\|<1$; the paper calls it the most central tool in perturbation theory. The lemma is used as an exact finite-order identity with explicit remainder, and every later object—eigenvalues, eigenvectors, scattering entries, diagram sums—is obtained by writing the relevant quantity as $(A+B-z)^{-1}$ or its exponential analogue and expanding. The companion exponential Dyson series is linked to the same identity through the Fourier representation of the resolvent, $(A+B+i\tau)^{-1}=i\int_0^\infty e^{it(A+B)-\tau t}dt$.

What would settle it

Take a small finite-dimensional system, compute exactly the Cesàro averaged correlation $\tau\int_0^\infty e^{-\tau t}\langle v_i,e^{-itA}e^{i2t(A+B)}e^{-itA}v_j\rangle dt$ as a function of $\tau$, and compare it with the resolvent series $i\tau\langle v_i,(2(A+B)-\lambda_i-\lambda_j+i\tau)^{-1}v_j\rangle$; a discrepancy in any fixed order as $\tau\to0^+$ would show that the scattering series is not the expansion of that averaged object. A second test is to exhibit an unbounded-operator pair for which $\|B\|=\infty$ yet the resolvent expansion converges, or $\|B\|<\infty$ yet the series diverges, which would break the paper's identification of divergences with infinite norms.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central claim is that a single lemma—the Neumann-type resolvent expansion for the inverse of a sum—unifies the perturbation-theoretic machinery of quantum field theory. The eigenvalue perturbation series is obtained by writing $\lambda(\epsilon)$ as a contour integral of $\operatorname{Tr}((z-A-\epsilon B)^{-1})$ and expanding the resolvent; the eigenvector corrections follow from the same expansion applied to the Schur-complement block form. The scattering matrix is rewritten as the resolvent entry $$S_{ij}=\lim_{\tau\to0^+} i\tau\langle v_i,(2(A+B)-\lambda_i-\lambda_j+i\tau)^{-1}v_j\rangle,$$ and its perturbation series is then read off from the resolvent lemma. In this way the notes present the standard objects of formal QFT—Dyson series, diagrams, Born and Rutherford scattering, Feynman parameters—as translations of finite-matrix identities into infinite dimension.

Load-bearing premise

The load-bearing premise is that the formal scattering limit $\lim_{t\to\infty} e^{-itA}e^{i2t(A+B)}e^{-itA}$ can be replaced by the resolvent average $\lim_{\tau\to0^+} i\tau\langle v_i,(2(A+B)-\lambda_i-\lambda_j+i\tau)^{-1}v_j\rangle$; the paper itself concedes that the true $t\to\infty$ limit does not exist for finite-dimensional matrices because of Poincaré recurrence, so the scattering formulas rest on an unproved intermediate-time approximation.

Editorial extensions

If this is right

  • Eigenvalue perturbation theory becomes a corollary of the resolvent lemma: for an isolated eigenvalue $\lambda(0)$ of $A$, the series $\lambda(\epsilon)=\sum_{k=0}^\infty \lambda^{(k)}\epsilon^k$ holds for small $\epsilon$, with the standard first and second order terms.
  • The Dyson series and the resolvent series are two faces of the same identity, connected by a Fourier transform in time, so a perturbation expansion can be chosen according to whether the problem is time-independent or time-dependent.
  • Scattering amplitudes are computable from the resolvent series; first order gives the Born approximation, and the Rutherford formula follows after summing over final states, while second order reproduces the four-diagram sum in the motivating QFT text.
  • When a symmetry $U$ commutes with $A$ and $B$, the resolvent expansion factorizes on eigenspaces of $U$, which can drastically reduce the intermediate states that need to be summed and underlies a tree-diagram uniqueness statement.
  • The Feynman-parameter representation emerges from the same expansion by writing $t_i=x_i t$ in the exponential series, so the Feynman parameter $x_i$ is interpreted as the fraction of time the system spends in state $k_i$.

Reading between the lines

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

  • A testable consequence the paper leaves implicit: in a finite-dimensional truncation of a QFT, the scattering 'matrix' should be defined through the Cesàro/resolvent average, not through a time limit that provably does not exist; comparing the two in small systems would show whether the difference is negligible at physical times.
  • The divergence analysis suggests an order-of-limits diagnostic for renormalization: compute the resolvent expansion in a finite-dimensional truncation and ask whether $\tau\to0^+$ and the truncation dimension $N\to\infty$ commute; a failure of commutativity would locate the UV/IR problem in the infinite-dimensional passage rather than in any finite-order term.
  • The counting argument in the tree-diagram proposition points to a general characterization: loop diagrams are exactly those for which the momentum-energy conservation equations leave free variables, so the loop number is the dimension of the solution space of that linear system.
  • If the resolvent identity is truly universal, one could attempt to generate the counterterm structure of a renormalizable theory by choosing the split $H=A'+B'$ so that the divergent parts of the expansion are absorbed into $A'$; the author gestures at dressed particles but leaves this construction open.
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.

Referee Report

2 major / 7 minor

Summary. This manuscript presents a set of lecture notes aiming to unify perturbative QFT techniques through finite-dimensional resolvent expansions. The central object is Lemma 1.1, the resolvent identity (A+B)^{-1}=Σ(-1)^m A^{-1}(BA^{-1})^m, and the paper derives from it series for eigenvalues (Proposition 1.2), Dyson series (Section 2), a scattering-matrix expansion (Section 3), symmetry-restricted diagram rules (Sections 4–5), Fourier and Feynman-parameter integral formulas (Section 6), and eigenvector perturbation formulas, including an adiabatic approach (Section 8). The stated program is that many perturbative QFT objects should be seen as particular cases of the resolvent lemma, with UV/IR divergences arising from unbounded operators that violate the lemma's convergence condition. The text compares its formulas with Talagrand's book and Peskin's textbook, and includes worked examples such as the Born approximation and a three-particle diagram.

Significance. The notes are useful pedagogically: the derivations are elementary, self-contained, and several formulas reproduce standard textbook results, including the eigenvalue perturbation expansion, the Dyson series, the Born approximation, and the three-particle diagram matching Talagrand's Equation 13.84. The explicit comparisons with external references are a strength. However, the paper's central claim of a universal resolvent-lemma framework is not established, because the scattering-matrix section defines S_ij through a t→∞ limit that the author concedes does not exist in finite dimension and then replaces it by an Abel/τ-regularized resolvent limit. This is a load-bearing gap: Sections 1 and 9 present the scattering matrix as a particular case of Lemma 1.1. The candid admissions of this issue and of the formal nature of Proposition 8.5 are commendable, but they also delimit the rigor of the notes. If Section 3 is honestly recast as a study of the τ-regularized operator, the manuscript can be a valuable pedagogical contribution; in its current form, the advertised universality is stronger than what is proved.

major comments (2)
  1. [Section 3, Lemma 3.1 and Proposition 3.2] The scattering matrix S_ij is not defined in the paper's own finite-dimensional setting. The author states that the t→∞ limit of ⟨v_i,e^{-itA}e^{i2t(A+B)}e^{-itA}v_j⟩ cannot exist for finite matrices by Poincaré recurrence, and Lemma 3.1 only proves equality under the antecedent that the limit exists. Since that antecedent is generically false, Proposition 3.2 is a series for the τ-regularized Abel object lim_{τ→0+} iτ⟨v_i,(2(A+B)-λ_i-λ_j+iτ)^{-1}v_j⟩, not for a scattering amplitude. A concrete check illustrates the gap: for A=diag(0,0) and B=gσ_x, the left side is i sin(2gt), which has no limit, while the resolvent-side limit is 0. This is load-bearing because Sections 1 and 9 present the scattering matrix as a particular case of Lemma 1.1. The manuscript should either define S_ij explicitly as the Abel limit and restrict all subsequent claims to that regularized object, or prove an actual scattering limit in a setting where it exists.
  2. [Section 8.5, Proposition 8.5] The proof of Proposition 8.5 is acknowledged by the author to be only formal, because the remainder term diverges in the η→∞ regime used to identify the ϵ^k terms of the eigenvector series. This is a genuine limitation: as stated, Proposition 8.5 invites the reader to treat an unproved identity as a rigorous expansion. The resolvent-based formulas of Section 8.1 do not depend on this proposition, so that main derivation is not affected, but the correspondence table in Section 9 should be marked to indicate which entries are rigorous and which are formal.
minor comments (7)
  1. [Section 1, Lemma 1.1] Lemma 1.1 should state explicitly that A is Hermitian positive definite, or that A^{-1/2} is defined through a specified functional calculus; without this, the norm condition is not well defined for arbitrary invertible matrices.
  2. [Section 3, Proposition 3.2] The condition ||(A−λτ)^{-1/2}B(A−λτ)^{-1/2}||<1 uses a square root of a normal but non-Hermitian matrix; please define this square root explicitly, or state a sufficient iteration condition such as ||(A−λτ)^{-1}B||<1 that follows directly from the resolvent identity.
  3. [Section 6.2, Remark 6.2] The displayed formula contains a stray dot in '˙B(i,k1)'; it appears to be a typographical artifact and should be removed.
  4. [Section 6.3, Proposition 6.4] In the final display of the proof, the denominator contains 'iτ t'; after the Laplace integration the time variable should not remain, and this is likely a typo for 'iτ'.
  5. [Section A.1, Example A.1] In the last line of the λ_i^{(4)} expression, the final denominator is written as 'z − λ_j'; this should be an evaluated residue expression with denominator '(λ_i − λ_j)' or the loop integral should be kept.
  6. [Section 8.2, equation (14) and Figure 2] The denominator convention in equation (14) uses (z−λ_{k_j}) while Figure 2 uses (λ_{k_2}−z); please harmonize the sign convention so that the displayed example matches the general formula.
  7. [Section 9, summary table] The summary table presents correspondences with [TAL22] and [Pes18] without indicating which entries are rigorous and which are formal; adding such indications would help readers distinguish proved statements from heuristic analogies.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the derivations reduce to the resolvent identity and are checked against, not assumed from, textbook results.

full rationale

The paper's central derivation chain begins with Lemma 1.1, which is proved directly from the resolvent identity (A+B)^-1 = A^-1 - A^-1B(A+B)^-1, and all later objects—eigenvalue perturbations, Dyson series, scattering-matrix coefficients, diagram sums, and eigenvector corrections—are expanded by applying that lemma or equivalent Duhamel/Cauchy formulas. The textbook results of Talagrand and Peskin are used as comparison targets, not as premises: the paper derives formulas and then states correspondences such as "This gives a similar expression as [TAL22, Theorem 12.4.1]" and "we recover [TAL22, Equation 13.84]". The only potentially weak point is the scattering-matrix definition in Section 3, where the t→∞ limit is not guaranteed to exist; the author explicitly acknowledges this and uses an Abel-mean/resolvent identification under an existence hypothesis. That is a mathematical gap or assumption, not a circular reduction: the resolvent series for S_ij(τ) does not presuppose the value or existence of the t→∞ limit. There are no fitted parameters, no predictions derived from data, and no load-bearing self-citations. The self-contained algebra stands independently of any external claim, so the circularity score is 0.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

No free parameters appear; the notes are self-contained derivations of standard perturbation formulas from the resolvent identity. The main unstated input is that matrix square roots are available and the norm condition holds. Several physical assumptions are informal, notably the existence of the scattering limit and the adiabatic gap condition.

assumptions (5)
  • domain assumption A is a Hermitian (or at least positive definite) matrix so that A^{-1/2} exists and the norm condition ||A^{-1/2}BA^{-1/2}|| < 1 is meaningful.
    Lemma 1.1 and Proposition 3.2 use A^{-1/2} without stating positivity; for scattering, A is later treated as Hermitian with real eigenvalues.
  • standard math The resolvent identity and the geometric (Neumann) series converge under the stated norm bound.
    Used in Lemma 1.1, Proposition 1.2, and Proposition 3.2; this is standard functional calculus.
  • domain assumption The scattering matrix limit can be replaced by the tau->0+ resolvent boundary value, and this remains a valid approximation in finite dimension.
    Section 3, Lemma 3.1: the paper uses this despite noting the t->infty limit cannot exist due to Poincare recurrence.
  • domain assumption In the adiabatic eigenvector formula, the spectral gap and C1 bounds on eigenvectors hold uniformly, and the eta->infty limit can be matched term by term with the epsilon expansion.
    Proposition 8.4 and Proposition 8.5; the matching is stated informally and the remainder is acknowledged to diverge.
  • domain assumption The basis states |p_1,...,p_l> with A diagonal and a conserved quantum number U (total momentum) correctly encode the physical content of Feynman diagrams.
    Section 5.1 hypotheses replace Fock space and Wick theorem; the diagrams are admittedly not equivalent to usual Feynman diagrams (footnote 8).

how reviews work

0 comments
Cite this review

Pith. "Pith review of A new look at Perturbation Theory in QFT and Resolvent Series." pith.science (2026). https://pith.science/paper/O7GUT7W2

@misc{pith2026250611572,
  author       = {Pith},
  title        = {Pith review of: A new look at Perturbation Theory in QFT and Resolvent Series},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/O7GUT7W2}},
  note         = {Machine review of arXiv:2506.11572}
}
read the original abstract

We give a short introduction for elementary mathematical tools used in the context of Quantum Field Theory. These notes were motivated by a reading group in Lyon on Talagrand's book {\guillemotleft}What is Quantum Field Theory, A First Introduction for Mathematicians{\guillemotright}.

Figures

Figures reproduced from arXiv: 2506.11572 by the authors.

Figure 1
Figure 1. Example of Diagram to compute Bk1k2Bk2k3 λk2−λτ with k1 = |p1, p2, p3⟩, k2 = |p3, p4⟩, k3 = |p5, p6⟩, Definition 5.1. (Diagram value) We denote S (ℓ,T ) ij (τ ) the r-h-s term of (7) and such that S (ℓ) ij (τ ) = X T S (ℓ,T ) ij (τ ). 5.2 Tree Diagram We give here a short justification of the well known fact in QFT : if the Feynman diagram is a tree, then the computation is not too problematic. Proposition 5.2. If t… view at source ↗
Figure 2
Figure 2. Example of diagrams T1,T2,T3 with v = k1 = |p1⟩, k2 = |p2, p3⟩, k3 = |p3, p4, p5⟩ corresponding to the terms ve (1,T1) k2 (z) = Bk1k2 λk2−z , ve (2,T2) k3 (z) = P k2 Bk3k2Bk2k1 (λk3−z)(λk2−z) and ⟨b, v(1,T1 (z)⟩ = P k2 Bk1k2Bk2k1 (λk2−z) [PITH_FULL_IMAGE:figures/full_fig_p018_2.png] view at source ↗
Figure 3
Figure 3. Example of diagrams corresponding ⟨b, v(1,T1 (z)⟩ 2 and ⟨ve (2,T2) i , Cve (1,T1) j ⟩ = P Bk1k2Bk2k3 Ck3k4Bk4k5 (λk2−λbi)(λk3−λbj )(λk4−λbj ) with T1, T2 the same as in [PITH_FULL_IMAGE:figures/full_fig_p018_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

5 extracted references · 4 canonical work pages

  1. [1]

    Random operators , volume 168

    Michael Aizenman and Simone Warzel. Random operators , volume 168. American Mathematical Soc., 2015

  2. [2]

    Lectures on the local semicircle law for wigner matrices

    Florent Benaych-Georges and Antti Knowles. Lectures on the local semicircle law for wigner matrices. arXiv preprint arXiv:1601.04055 , 2016

  3. [3]

    u rg Fr \

    J \"u rg Fr \"o hlich and Thomas Spencer. Absence of diffusion in the anderson tight binding model for large disorder or low energy. Communications in Mathematical Physics , 88(2):151--184, 1983

  4. [4]

    An Introduction to quantum field theory

    Michael E Peskin. An Introduction to quantum field theory . CRC press, 2018

  5. [5]

    TALAGRAND

    MICHEL. TALAGRAND. WHAT IS A QUANTUM FIELD THEORY?: A First Introduction for Mathematicians . CAMBRIDGE University Press, 2022

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.