Pith. Sign in

REVIEW

Wilson Loops and Minimal Surfaces

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
Nadav Drukker, David J. Gross, Hirosi Ooguri

High Energy Physics - Theory · Submitted 27 Apr 1999 · 29 references

Pith's verdict: The Legendre-transform prescription is the real contribution and it is right; the loop-equation verification is a partial result that the abstract dresses slightly better than the body supports.

T0 · AI review

How this review was made: An AI referee (selfhosted/deepseek-v4.1-flash) read the paper on October 29, 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 →

· 4 major objections · 6 minor · Ask about this paper · The compiled record · Read the full review →
Challenge this review Comment (0) Read on arXiv ↗
Cite ▾
arXiv IDhep-th/9904191 (version 1)
Pith IDRVTL7OUA, Pith's permanent record of this paper and its review
CitationDavid J. Gross, Hirosi Ooguri, Nadav Drukker. "Wilson Loops and Minimal Surfaces." arXiv:hep-th/9904191 (1999). https://pith.science/paper/RVTL7OUA
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 paper argues that the correct string dual of a Wilson loop is not the minimal-surface area but its Legendre transform, which cancels the linear ultraviolet divergence for BPS loops satisfying $x$-dot squared equals $y$-dot squared.

The reading

The paper asks whether the proposed string description of a Wilson loop in four-dimensional $N=4$ super Yang-Mills is internally consistent. It isolates a class of loops, called BPS loops, whose couplings to the gauge field and the six scalars satisfy $x$-dot squared equals $y$-dot squared, and argues that their expectation values are free of the linear ultraviolet divergence. The central proposal is that the correct string action is the Legendre transform of the minimal-surface area with respect to the six Neumann string coordinates, not the area itself. The authors formulate the large-$N$ loop equation for these loops and check, for smooth loops and cusps, that the minimal surface in $AdS_5$ times $S^5$ gives the expected answer. The check is partial: at a cusp the coefficient depends on an unknown function for loops that violate the constraint, and the self-intersection case is not reproduced.

What carries the argument

The carrying object is the Legendre transform of the area with respect to the boundary momenta conjugate to the six Neumann string coordinates. The Neumann condition is written using the induced complex structure on the worldsheet, and this is what ties the string boundary data to the loop's coupling to the six scalars. Hamilton-Jacobi variation shows that the transformed action is a functional of the boundary momenta, and for a smooth loop the transformed linear divergence becomes a difference of the two length functionals, which vanishes under the BPS constraint. The loop equation is the second functional derivative with respect to the loop coordinates, and its right-hand side counts cusps and self-intersections.

Extended reading notes

Core claim

The central claim is that the Wilson loop of $N=4$ super Yang-Mills theory should be computed from a minimal surface in $AdS_5$ times $S^5$ using a Legendre-transformed action, not the Nambu-Goto area. The six $S^5$ coordinates obey Neumann boundary conditions set by the loop's scalar couplings, and when the loop satisfies $x$-dot squared equals $y$-dot squared, the Legendre transform subtracts a boundary term proportional to the integral of $y$-dot over the cutoff, cancelling the linear divergence of the area and leaving a finite BPS-loop expectation value. The paper then formulates the large-$N$ loop equation for such loops and shows that the minimal surface gives zero for a smooth loop, while a cusp produces a non-zero term of order the 't Hooft coupling. The authors present this as evidence that the minimal-surface ansatz solves the loop equation, while noting that the cusp coefficient and the intersection case are not fully under control.

Load-bearing premise

The load-bearing premise is the Neumann boundary condition with unit normalization, which the authors explicitly say they cannot derive from first principles, so if it is wrong the cancellation of the linear divergence and the finite BPS-loop claim do not follow.

Editorial extensions

If this is right

  • If the Legendre-transform prescription is right, smooth BPS loops have finite expectation values and the area-only Wilson-loop formula is incomplete.
  • The minimal surface in $AdS_5$ times $S^5$ satisfies the large-$N$ loop equation for smooth BPS loops, which supports the duality at the level of loop equations.
  • Cusps and self-intersections produce non-zero loop-equation sources of order the 't Hooft coupling, so the loop equation is not simply zero for all singular loops.
  • The constraint $x$-dot squared equals $y$-dot squared is reinterpreted geometrically and as a BPS condition: only half the local supersymmetry in loop space acts on such a loop.
  • Zig-zag symmetry is expected only when the loop does not couple to the scalars, and the minimal-surface picture reproduces this when the boundary condition is imposed at $Y=\epsilon$.

Reading between the lines

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

  • The Legendre-transform lesson likely extends to other AdS/CFT observables with Neumann worldvolume directions: area-only calculations may need boundary terms whenever the dual operator sources momentum rather than position.
  • If the unknown function controlling loops away from the constraint could be fixed by requiring the loop equation at cusps, the same excited-string picture might give a workable definition of non-BPS Wilson loops.
  • The unresolved self-intersection case may point to a missing contact term or loop-dependent prefactor in the classical minimal-surface prescription rather than a failure of the duality.
Original abstract

The AdS/CFT correspondence suggests that the Wilson loop of the large N gauge theory with N=4 supersymmetry in 4 dimensions is described by a minimal surface in AdS_5 x S^5. We examine various aspects of this proposal, comparing gauge theory expectations with computations of minimal surfaces. There is a distinguished class of loops, which we call BPS loops, whose expectation values are free from ultra-violet divergence. We formulate the loop equation for such loops. To the extent that we have checked, the minimal surface in AdS_5 x S^5 gives a solution of the equation. We also discuss the zig-zag symmetry of the loop operator. In the N=4 gauge theory, we expect the zig-zag symmetry to hold when the loop does not couple the scalar fields in the supermultiplet. We will show how this is realized for the minimal surface.

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, 6,073 input tokens, $0.0003 reported.

What the abstract itself states, as probabilities: claims first 0.03 · states limitation 0.71 · human participants 0.01 · controlled comparison 0.05 · randomized assignment 0.01 · states quantitative result 0.04 · compares with prior results 0.59 · mentions code or data availability 0.01.

The abstract, sentence by sentence

arXiv listing abstract v1. Results and the main claim are underlined; .

The AdS/CFT correspondence suggests that the Wilson loop of the large N gauge theory with N=4 supersymmetry in 4 dimensions is described by a minimal surface in AdS_5 x S^5. We examine various aspects of this proposal, comparing gauge theory expectations with computations of minimal surfaces. There is a distinguished class of loops, which we call BPS loops, whose expectation values are free from ultra-violet divergence. We formulate the loop equation for such loops. To the extent that we have checked, the minimal surface in AdS_5 x S^5 gives a solution of the equation. We also discuss the zig-zag symmetry of the loop operator. In the N=4 gauge theory, we expect the zig-zag symmetry to hold when the loop does not couple the scalar fields in the supermultiplet. We will show how this is realized for the minimal surface.

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 The AdS/CFT correspondence suggests that the Wilson loop of the large N gauge theory with N=4 supersymmetry in 4 dimensions is described by a minimal surface in AdS_5 x S^5. background 1.00 0.05 0.25 hedged possibility 0.99
s1 We examine various aspects of this proposal, comparing gauge theory expectations with computations of minimal surfaces. contribution 0.92 0.87 0.06 asserted 0.79
s2 There is a distinguished class of loops, which we call BPS loops, whose expectation values are free from ultra-violet divergence. contribution 0.52 0.24 0.07 asserted 1.00
s3 We formulate the loop equation for such loops. contribution 0.77 0.91 0.06 asserted 0.95
s4 To the extent that we have checked, the minimal surface in AdS_5 x S^5 gives a solution of the equation. result 0.94 0.88 0.38 hedged possibility 0.68
s5 We also discuss the zig-zag symmetry of the loop operator. contribution 0.97 0.58 0.04 asserted 0.77
s6 In the N=4 gauge theory, we expect the zig-zag symmetry to hold when the loop does not couple the scalar fields in the supermultiplet. background 0.66 0.11 0.06 hedged possibility 0.60
s7 We will show how this is realized for the minimal surface. contribution 0.81 0.80 0.10 asserted 0.99

What kind of paper the classifier reports this is: theoretical result 1.00 · evidence from proof or derivation 0.72.

What was examined, and what was not

sources read
  • arXiv listing abstract v1
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. (a) At one loop, there is a linear divergence from the propagator connecting coincident points. The divergence is proportional to the circumference of the loop. (b) At cusps and intersections, an additional logarithmic divergence ap￾pears when the 2 external legs approach the singular point. 10-dimensional theory, the 10-dimensional rotational invariance of the Feyn￾man rule is recovered in the coincidence limit. Therefore the contractions of the external indices by the Feynman rule produce only rotational invariant combinations of ( ˙x µ , iy˙ i ), namely a polynomial of ( ˙x 2 −
Figure 2. A minimal surface for a Wilson loop with a cusp. The regularized area is evaluated over the shaded region. The result is that the area of the surface has a logarithmic divergence as well as a linear divergence. It behaves as A = L 2πǫ − 1 2π F(Ω, Θ) log L ǫ + · · · , (3.19) where Ω and Θ are the cusp angles in R4 and S 5 respectively. When either Θ or Ω vanishes, we can express F(Ω, Θ)/2π in terms of elliptic integrals. In fig. 3. we show the numerical evaluation of the function F(Ω, 0) in the solid curve. This is to be compared with the perturbative expression (2.12) shown
Figure 3. The solid curve shows the function F(Ω, 0)/2π, which appears in the logarithmic divergence of the minimal surface with the cusp of angle Ω. This is compared with the perturbative result (2.12) at the cusp shown in the dashed curve. The dotted curve is half of the perturbative result (2.13) at an intersection. modified at the cusp as x˙ 2 = (1 + f 2 0 ) ˙y 2 , (3.20) where f0 = f(ϕ = Ω/2) is the minimal value of f(ϕ). (4) Intersection: The minimal surface for a self-intersecting loop is just the sum of 2 cusps. The only difference is that, by the exchange symmetry of the 2 com
Figure 4. The comparison of the two regularization prescrip￾tions. The boundary conditions are imposed at Y = 0 in (a) and at Y = ǫ in (b). The shaded regions represent the regularized areas. divergence in the area of the minimal surface in AdS5 is proportional to the circumference of the loop5 . The linear divergence arises from the leading behavior of the surface at small Y , i.e. near the boundary of AdS5. In this section, we have computed the regularized area by imposing the boundary condition at the boundary Y = 0 of AdS5 and integrating the area element over the part of the surface Y ≥ ǫ
Figure 5. The zig-zag loop; the loop goes in one direction along C1 and comes back along C2. The two segments C1 and C2 are parallel and their distance η is less that the gauge theory UV cutoff ǫ. In perturbation theory, one can easily prove that the zig-zag symmetry holds for the Wilson loop (2.4) when ˙y i = 0. Suppose we have a segment C1 of a loop which goes in one direction and another segment C2 which comes back parallel to C1 but in the opposite direction, as shown in fig. 5. If the distance η between C1 and C2 is much less than the UV regularization ǫ of the gauge theory, there is one
Figure 6. The area of a loop with a zig-zag (a) is roughly the same as the loop without it (b). with or without the backtracking. At first sight, the zig-zag symmetry of the minimal surface appears in contradiction with the gauge theory expectation since we know the minimal surface ending along a smooth loop on the boundary of AdS5 obeys the constraint ˙x 2 = ˙y 2 and therefore ˙y i 6= 0. In the gauge theory, we do not expect zig-zag symmetry when ˙y i is non-zero and constant. A close examination of the boundary condition, however, reveals that the situation is more subtle. It is true th
Figure 7. (a) A self-intersecting loop which corresponds to a single trace operator and (b) A pair of loops obtained by reconnecting the loop at the intersection. surface bounded by C1 and C2 stays within η from Y = ǫ, and ˙y 2 vanishes as η/ǫ → 0. If we take the latter point of view, the apparent contradiction with the gauge theory expectation disappears since the minimal surface in question is related to the Wilson loop which does not couple to the scalar fields in the segments C1 and C2. This is exactly the situation in which zig-zag symmetry arises in the gauge theory. One may argue that

Figures reproduced from arXiv:hep-th/9904191 by the authors.

References
How later papers used it

170 later papers cite it

Evidence: substantial. 155 of the citing papers share no author with it, and 39 of those carry a confident label from Pith's citation classifier.

Pith read 22 of the 314 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 28 papers with this label
How the literature uses it

166 arXiv papers cite this in 307 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 · 180 sentences without a confident label

Builds on it 8 papers
Uses its method or tool 28 papers
  • which is nothing but the divergence removed from the action by the Legendre transformation [ref] to give $S_reg.$; thus we see that the regularized action vanishes for these solutions.

    Supersymmetric Wilson Loops in Diverse Dimensions · Calibrated surfaces · 0.87

  • In order to regularize the integral we must Legendre transform to the variables that are conjugate to the transverse directions to the Wilson loop [ref] (see also [ref] for a similar setup to ours).

    Precision Holography for 5D Super Yang-Mills · The classical on-shell action · 0.96

  • The potential piece, to compare with the gauge theory, is extracted by removing the divergence from ([ref]) as in [ref] .

    Holographic Interface-Particle Potential · String and potential · 0.96

Uses its data or values 1 paper
Compares with it 10 papers
  • We have verified that the answer given in [ref] can be obtained by the regularisation procedure which uses an appropriate Legendre transform [ref] .

    Wilson Loops and Spherical Branes · Holographic Wilson loop · 0.80

  • We observe that this localization does not hold for $MM$ loop equation [ref] and ordinary Wilson loops $Ψ(x,y;A)=P expi∫_L_xy A_α dx_α$: ∫_L_xx dx_α<N/2 g^2 Tr(δS_YM/δA_α(x)Ψ(x,x;A))> =i ∫_L_xx dx_α ∫_L_xx dy_α δ^(4)(x-y) <Tr Ψ(x,y;A)>< Tr Ψ(y,x;A)> Indeed, while deforming an ordinary Wilson loop by a backtracking cusp leaves the loop invariant at classical level because of the zig-zag symmetry [ref] of Wilson loops, at quantum level it introduces an additional logarithmic divergence known as cusp anomaly [ref] .

    Yang-Mills mass gap, Floer homology, glueball spectrum, and conformal window in large-N QCD · A synopsis of the main argument and plan of the paper · 0.94

  • When $g_4 ≈0$ and $n_1 n_2=1$ one finds the familiar $N=4$ formula [ref] W≈2 I_1 (√( λ_YM ) ) √( λ_YM ) with λ_YM = g_YM^2 N On the other hand in the limit of large $λ_eff$ with $g_4 λ^2$ kept finite one finds W ≈4 e^√(λ_eff ) √(2π) λ_eff^34 ( δ+1 δ+2 ) This result provides us with a remarkable simple prediction for the AdS dual of the deformed theory.

    Wilson Loops and Chiral Correlators on Squashed Sphere · The $N=4$ deformed theory · 0.80

Cites it as support 3 papers
Cites it as background 60 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.