REVIEW 3 major objections 4 minor 29 references
Derivation of the GKP-Witten relation by symmetry without Lagrangian
T0 review · 3 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash
Pith's one-line read The GKP-Witten relation between bulk and boundary correlators is derived using only conformal symmetry and the operator product expansion, with no Lagrangian, for arbitrary spin and to all orders in the source.
desk verdict A genuinely new all-orders, arbitrary-spin version of the GKP-Witten dictionary from symmetries, but the central equality is conditional on a Lagrangian-style subtraction of contact terms the paper does not derive. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The central device is conformal smearing: a bulk field $\hat\sigma^a(X)=\int d^d y\, h(z,x-y)\,\hat\phi^a(y)$ with kernel $h(z,x)\propto (z/(x^2+z^2))^{d-\Delta_\phi}$, which converts boundary conformal transformations into bulk coordinate transformations. Combined with the embedding-space formula for spinning correlators and a diagonalization step that selects orthogonal combinations of bulk operators, this yields the 2-point and 3-point functions used to seed an inductive OPE argument. The induction itself is the load-bearing mechanism: once the 3-point case is fixed by symmetry, the OPE forces every higher bulk-boundary correlation function to have the same boundary limit as the corresponding CFT correlator, which assembles into Eq. (29).
What would settle it
Evaluate the $z\to0$ limit of the bulk-boundary 4-point function in an explicit CFT with $\Delta_\phi<d/2$ (for instance the free scalar field theory) keeping all contact terms, and check whether the coefficient of $z^{\Delta_p-L_p}$ equals the connected CFT correlator after subtracting only local terms at $x_1$. If a residual nonlocal piece survives, Eq. (29) fails; if it does not, the induction is confirmed beyond the 3-point order.
Extended reading notes
Core claim
On the paper's own terms, the discovery is that the GKP-Witten relation holds at the level of correlation functions for arbitrary spin primary operators and to all orders in an external source, with no bulk action. After constructing bulk operators $\mathcal{G}^i_{s_1}(X_1)$ from boundary primaries by conformal smearing, diagonalizing them so that each $\mathcal{G}^i$ couples only to one primary $\mathcal{O}^{p_i}$, and fixing the bulk-boundary 2- and 3-point functions by symmetry via the embedding-space formula, the paper proves by induction using the OPE that the $z_1\to0$ limit of $\Phi^i_{s_1}(X_1)$—the bulk operator with sources turned on, divided by $Z(J)$—equals $z_1^{d-\Delta_{p_i}-L_{p_i}}\Lambda_{\Delta_{p_i}}^{-1} J^{p_i,s_1}(x_1)$ plus $z_1^{\Delta_{p_i}-L_{p_i}}$ times the connected generating functional of boundary correlators, up to local contributions that are dropped by analogy with boundary counterterms. Consequently the relation is presented as a theorem of conformal symmetry plus OPE, valid for any CFT satisfying their conditions (notably $\Delta_\phi<d/2$ for the smearing construction), and it cannot serve as a test of whether a CFT is holographic.
Load-bearing premise
The load-bearing premise is that the delta-function (contact) terms appearing in the small-$z$ limit can be set aside without a Lagrangian, by analogy with boundary counterterm subtraction; if those terms cannot be removed on symmetry grounds alone, the exact boundary limit of the bulk operator would not be the GKP-Witten form (29).
Editorial extensions
If this is right
- The GKP-Witten dictionary is a consequence of conformal symmetry and the OPE, so it applies to any CFT satisfying the stated conditions, not only holographic ones.
- The relation cannot be used to decide whether a given CFT has a holographic dual, since generic CFTs also satisfy it.
- A bulk theory can be interacting, with nonzero connected bulk-boundary higher-point functions, while its 2-point function obeys free AdS equations of motion.
- Bulk reconstruction from boundary data can proceed without a bulk action; conformal smearing is needed only to justify the covariance constraint, not for the final theorem.
- The framework opens a non-Lagrangian route to higher-spin duals and to extending conformal bootstrap methods to bulk-boundary correlators.
Reading between the lines
- If the contact-term subtraction can be justified purely within the non-Lagrangian framework—for instance by deriving the local counterterms from the symmetries themselves—Eq. (29) would become an exact statement rather than one modulo local terms; the paper leaves this gap open.
- Extending the argument to correlators with more than one bulk operator, which the authors suggest, would amount to defining a full bulk field theory without an action; the success or failure of that extension would sharpen the claim that the extra dimension is kinematic.
- Testing the subtracted limit in an explicit solvable CFT, such as the O(N) vector model or the critical $\phi^4$ theory in three dimensions, would provide a concrete check of the induction beyond the 3-point function.
- Because the smearing kernel requires $\Delta_\phi<d/2$, the 'generic CFT' claim is conditional; CFTs with only operators above the unitarity bound would need a different construction.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper constructs bulk operators from a scalar CFT primary by conformal smearing, uses conformal covariance to fix bulk-boundary 2- and 3-point functions for arbitrary spin, and then uses the OPE to extend the result to n-point functions. It defines a bulk one-point function in the presence of arbitrary CFT sources and claims that its small-z limit reproduces the connected CFT generating functional, Eq. (29), without using a Lagrangian or the large-N expansion. The central derivation, however, is conditional on dropping delta-function contact terms after Eq. (20), with the subtraction justified by appeal to the standard semiclassical Lagrangian approach.
Significance. If the subtraction issue is resolved, the result is significant: it would show that the non-local content of the GKP-Witten relation follows from conformal covariance plus the OPE, without large N or an action, and that the GKP-Witten relation therefore does not by itself distinguish holographic from generic CFTs. The paper's strengths are its explicit treatment of arbitrary spin via embedding-space formulas, the clean OPE induction, and its transparent acknowledgment of the contact-term limitation. The derivation of the 2- and 3-point functions and the induction step are standard and clearly presented.
major comments (3)
- [Eq. (20) and derivation of Eq. (29)] The exact small-z limit of Phi_i_s1(X1) is not shown to equal Eq. (29), because the delta-function contributions are dropped after Eq. (20) with the justification that they are 'usually subtracted by boundary counter terms in the standard semiclassical approach [5,6]'. Since Refs. [5,6] are Lagrangian treatments, this is an external input, not a consequence of the conformal covariance constraint (10) or the OPE. Please supply a non-Lagrangian subtraction prescription, for example a normal-ordering or a defining smearing of the boundary limit that removes the local terms covariantly, or prove that the local terms do not contribute to the integrated correlator. As written, the no-Lagrangian claim in the Conclusion is conditional on this unproven assumption.
- [Eq. (16)] The diagonalization step assumes the existence of k independent bulk operators \bar{G}^i_s with an invertible coupling matrix c_ij. This is an additional assumption, not derived from the conformal smearing construction. If the matrix is not invertible, or if the smeared composite operators do not yield exactly k independent combinations, the construction of G^i_s and the assertion that it couples only to O^{p_i} fail. Please show that the conformal smearing construction indeed produces k independent bulk operators for each spin sector.
- [Eqs. (22)-(24)] The induction step applies the OPE and then invokes Eqs. (20) and (15), but the set of local terms generated in this process is not characterized. When the OPE coefficients contain derivatives and the distributional identity (14) is used, the contact terms need not be of the simple 2-point form already displayed in Eq. (15); the paper does not show that all such terms are of the type 'usually subtracted' by boundary counterterms. This gap affects the all-orders-in-J claim for arbitrary sources.
minor comments (4)
- [Title page] The heading contains the typo 'L agrangian'; it should read 'Lagrangian'.
- [Conclusion] The abbreviation 'SAs' is used without definition; spell it out at first occurrence.
- [Eq. (14)] The identity in Eq. (14) is distributional; please specify the range of alpha and the space of test functions on which it holds, and state any regularization needed for the delta-function term.
- [Reference [22]] Reference [22] is cited only by arXiv number; update to the published version if one exists.
Circularity Check
The dictionary entry is partly built into the diagonalized bulk operators, and the exact small-z limit is obtained only after importing a Lagrangian-style contact-term subtraction, so the 'symmetry-only' derivation is conditional.
-
self definitional
[Sec. 3.2 'Diagonalization of bulk operators' (Eqs. 16-17) and Sec. 3.5 'GKP-Witten relation' (Eqs. 25-29)]
"We then construct a set of bulk operators as G^i_s = sum_j c^{-1}_{ij} \bar G^j_s, which is diagonal as lim_{z1->0} <0|G^i_{s1}(X1)O^{p_j}_{s2}(x2)|0> = delta_{ij} F^{i,p_i}_{s1,s2}(X1,x2). ... We define, for a bulk spin L_{p_i}=|s_1| operator in the presence of sources in CFT, the analog of Phi_cl as Phi^i_{s1}(X1) := <0|G^i_{s1}(X1) exp[ integral d^d y sum_{p,s} J^{p,s}(y)O^p_s(y)]|0> / Z(J). ... Employing the above formula, we can establish the GKP-Witten relation."
The bulk operators G^i are defined by inverting the matrix c_{ij} precisely so that their boundary two-point limit is the CFT two-point function of the selected primary O^{p_i} (up to the explicit local term in F). The central formula (29) is then the sum over n of the boundary limits (22), which for n=2 is the defining property (17) and for n>=3 is that same property propagated through the OPE. Hence the leading dictionary entry, the coefficient z^{Delta-L}<O exp(J O)>, is not an independent prediction of the symmetries; it is inserted by the diagonalization. The derivation retains real content in showing that the same coefficient survives at all orders and in fixing the exponents through Delta and L, so the circularity is partial rather than total.
full rationale
The paper's construction has a self-definitional core: the bulk operator whose boundary limit appears in (29) is not an independent dynamical field; it is built by conformal smearing and then linearly recombined in Eqs. (16)-(17) so that its boundary two-point function is the CFT two-point function of a chosen primary. Once that diagonalization is imposed, the OPE induction in Eqs. (20)-(24) is a genuine argument that the n-point boundary limits are the corresponding CFT correlators, and the powers (Delta-L and d-Delta-L) are fixed by conformal weights. The calculation is therefore not vacuous. However, the conclusion that the GKP-Witten relation is derived 'using only symmetries without a Lagrangian' is stronger than what is demonstrated. First, the leading dictionary entry is imposed by the diagonalization rather than predicted. Second, the exact small-z limit contains delta-function (contact) terms; after Eq. (20) the paper states 'We therefore neglect these types of contributions hereafter except those in the 2-pt function', justified only by the statement that such terms are 'usually subtracted by boundary counter terms' in the Lagrangian approach [5,6]. That subtraction is an external renormalization input, not a consequence of the conformal covariance constraint (10) or the OPE. Thus Eqs. (28)-(29) hold only after importing a Lagrangian-based subtraction prescription, which qualifies the no-Lagrangian claim. The self-citation [16] is used only to guarantee existence of the conformal smearing and the constraint (10); no uniqueness theorem or fitted result is imported from the authors' prior work, so this does not add circularity. Overall score 5: the central statement has independent content (all-order OPE consistency and spin dependence), but its leading term is definitional and the exact statement is conditional on the contact-term subtraction.
Assumptions & free parameters
free parameters (2)
- c_ij normalization matrix
- Sigma0 smearing normalization =
sqrt(Gamma(d) Gamma(d - Delta_phi) / (N pi^d Gamma(d/2 - Delta_phi) Gamma(d/2)))
assumptions (6)
- domain assumption Boundary CFT has conformal symmetry with the standard transformation of primaries (Eq. 7) and the bulk operators transform as (Eq. 10).
- domain assumption There exists a scalar primary phi-hat^a with Delta_phi < d/2 satisfying (4), so the smearing kernel (6) is well-defined.
- standard math The OPE converges and can be used inside bulk-boundary correlators, and the radial ordering by |x| is consistent.
- ad hoc to paper There exist k independent bulk operators G-bar^i_s with invertible coupling matrix c_ij (Eq. 16).
- ad hoc to paper Delta-function contributions in the small-z limit may be neglected or subtracted without a Lagrangian (after Eq. 20).
- standard math Spinning embedding-space forms of the 2-pt and 3-pt functions (Eqs. 12, 18) are correct, per [17,18].
invented entities (1)
-
Bulk operator sigma-hat^a(X) and G^i_s(X) constructed by conformal smearing
Cite this review
Pith. "Pith review of Derivation of the GKP-Witten relation by symmetry without Lagrangian." pith.science (2026). https://pith.science/paper/KZUA76T4
@misc{pith2026241116269,
author = {Pith},
title = {Pith review of: Derivation of the GKP-Witten relation by symmetry without Lagrangian},
year = {2026},
howpublished = {\url{https://pith.science/paper/KZUA76T4}},
note = {Machine review of arXiv:2411.16269}
}
abstract
We derive the GKP-Witten relation in terms of correlation functions by symmetry without referring to a Lagrangian or the large $N$ expansion. By constructing bulk operators from boundary operators in conformal field theory (CFT) by the conformal smearing, we first determine bulk-boundary 2-pt functions for an arbitrary spin using both conformal and bulk symmetries, then evaluate their small $z$ behaviors, where $z$ is the $(d+1)$-th coordinate in the bulk. Next, we explicitly determine small $z$ behaviors of bulk-boundary-boundary 3-pt functions also by the symmetries, while small $z$ behaviors of correlation functions among one bulk and $n$ boundary operators with $n\ge 3$ are fixed by the operator product expansion (OPE). Combining all results, we construct the GKP-Witten relation in terms of these correlation functions at all orders in an external source $J$. We compare our non-Lagrangian approach with the standard approach employing the bulk action. Our results indicate that the GKP-Witten relation holds not only for holographic CFTs but also for generic CFTs as long as certain conditions are satisfied.
Reference graph
Works this paper leans on
-
[16]
Special flow equation and GKP-Witten relation
S. Aoki, J. Balog, T. Onogi, and S. Yokoyama, “Special flow equation and the GKP–Witten relation,” PTEP 2023 no. 1, (2023) 013B03 , arXiv:2204.06855 [hep-th] . 6
work page Pith review arXiv 2023
-
[1]
and ( 2), as lim z1→0 Φ i s1 (X1) = z d−∆ pi −Lpi 1 Λ ∆ pi J pi,s 1 (x1) + · · · (29) + z ∆ pi −Lpi 1 ⟨0|Opi s1 (x1) exp [ ∫ ddy ∑ p,s J p,s (y)Op s (y) ] |0⟩c for i = 1 , 2, · · · , k , where · · · represent local contribu- tions, which contain J(x1), and are usually subtracted by boundary counter terms in the standard Lagrangian approach. Some compariso...
-
[2]
Gauge theory correlators from noncritical string theory,
S. S. Gubser, I. R. Klebanov, and A. M. Polyakov, “Gauge theory correlators from noncritical string theory,” Phys. Lett. B 428 (1998) 105–114 , arXiv:hep-th/9802109
arXiv 1998
-
[3]
Anti-de Sitter space and holography,
E. Witten, “Anti-de Sitter space and holography,” Adv. Theor. Math. Phys. 2 (1998) 253–291 , arXiv:hep-th/9802150
arXiv 1998
-
[4]
AdS dynamics from conformal field theory,
T. Banks, M. R. Douglas, G. T. Horowitz, and E. J. Martinec, “AdS dynamics from conformal field theory,” arXiv:hep-th/9808016
-
[5]
The Large N limit of superconformal field theories and supergravity,
J. M. Maldacena, “The Large N limit of superconformal field theories and supergravity,” Adv. Theor. Math. Phys. 2 (1998) 231–252 , arXiv:hep-th/9711200
arXiv 1998
-
[6]
Variational Methods in AdS/CFT
T. Andrade, M. Banados, and F. Rojas, “Variational Methods in AdS/CFT,” Phys. Rev. D 75 (2007) 065013 , arXiv:hep-th/0612150
work page Pith review arXiv 2007
-
[7]
Holographic and Wilsonian Renormalization Groups,
I. Heemskerk and J. Polchinski, “Holographic and Wilsonian Renormalization Groups,” JHEP 06 (2011) 031 , arXiv:1010.1264 [hep-th]
arXiv 2011
Show all 29 references
-
[8]
Holographic Geometry of Entanglement Renormalization in Quantum Field Theories,
M. Nozaki, S. Ryu, and T. Takayanagi, “Holographic Geometry of Entanglement Renormalization in Quantum Field Theories,” JHEP 10 (2012) 193 , arXiv:1208.3469 [hep-th]
2012 arXiv
-
[9]
Lecture notes on holographic renormalization,
K. Skenderis, “Lecture notes on holographic renormalization,” Class. Quant. Grav. 19 (2002) 5849–5876 , arXiv:hep-th/0209067
2002 arXiv
-
[10]
DERIV ATION OF THE GKP-WITTEN RELATION FOR AN ARBITRAR Y SPIN In this section, we derive the GKP-Witten relation for spinning fields at all order in J, using the constraint ( 10)
in the rest of this paper. DERIV ATION OF THE GKP-WITTEN RELATION FOR AN ARBITRAR Y SPIN In this section, we derive the GKP-Witten relation for spinning fields at all order in J, using the constraint ( 10). 2-pt function The bulk-boundary 2-pt function G1,p 2 A1,s 2 (X1, x 2) :...
-
[11]
Geometries from field theories,
S. Aoki, K. Kikuchi, and T. Onogi, “Geometries from field theories,” PTEP 2015 no. 10, (2015) 101B01 , arXiv:1505.00131 [hep-th]
2015 arXiv
-
[12]
Flow equation for the large N scalar model and induced geometries,
S. Aoki, J. Balog, T. Onogi, and P. Weisz, “Flow equation for the large N scalar model and induced geometries,” PTEP 2016 no. 8, (2016) 083B04 , arXiv:1605.02413 [hep-th]
2016 arXiv
-
[13]
Surface/State Correspondence as a Generalized Holography,
M. Miyaji and T. Takayanagi, “Surface/State Correspondence as a Generalized Holography,” PTEP 2015 no. 7, (2015) 073B03 , arXiv:1503.03542 [hep-th]
2015 arXiv
-
[14]
Continuous Multiscale Entanglement Renormalization Ansatz as Holographic Surface-State Correspondence,
M. Miyaji, T. Numasawa, N. Shiba, T. Takayanagi, and K. Watanabe, “Continuous Multiscale Entanglement Renormalization Ansatz as Holographic Surface-State Correspondence,” Phys. Rev. Lett. 115 no. 17, (2015) 171602 , arXiv:1506.01353 [hep-th]
2015 arXiv
-
[15]
Holographic computation of quantum corrections to the bulk cosmological constant,
S. Aoki, J. Balog, and S. Yokoyama, “Holographic computation of quantum corrections to the bulk cosmological constant,” PTEP 2019 no. 4, (2019) 043B06 , arXiv:1804.04636 [hep-th]
2019 arXiv
-
[17]
Flow equation, conformal symmetry, and anti-de Sitter geometry,
S. Aoki and S. Yokoyama, “Flow equation, conformal symmetry, and anti-de Sitter geometry,” PTEP 2018 no. 3, (2018) 031B01 , arXiv:1707.03982 [hep-th]
2018 arXiv
-
[18]
Therefore eq
should agree with the above expression, we see that F (x) ∝ xα as x → 0, where 2 α = τp2 + τp3 − τpi with τp = ∆ p + Lp. Therefore eq. ( 20) holds for any operator orderings. Bulk-boundary n-pt functions We consider the Bulk-boundary correlation functions, defined by Gi,p 2, ··...
-
[19]
AdS geometry from CFT on a general conformally flat manifold,
S. Aoki and S. Yokoyama, “AdS geometry from CFT on a general conformally flat manifold,” Nucl. Phys. B 933 (2018) 262–274 , arXiv:1709.07281 [hep-th]
2018 arXiv
-
[20]
Higher Spin Gauge Theory and Holography: The Three-Point Functions,
S. Giombi and X. Yin, “Higher Spin Gauge Theory and Holography: The Three-Point Functions,” JHEP 09 (2010) 115 , arXiv:0912.3462 [hep-th]
2010 arXiv
-
[21]
Spinning Conformal Correlators,
M. S. Costa, J. Penedones, D. Poland, and S. Rychkov, “Spinning Conformal Correlators,” JHEP 11 (2011) 071 , arXiv:1107.3554 [hep-th]
2011 arXiv
-
[22]
We assume that 4 eq
for n = 3. We assume that 4 eq. ( 22) holds for n. Using the OPE, we have Gi,p 2, ··· ,p n+1 s1,s 2, ··· ,s n+1(X1, x 2, · · · , x n, x n+1) = ∑ p,s Cpnpn+1p snsn+1s (xn − xn+1, ∂ xn+1) (23) × Gi,p 2, ··· ,p n− 1,p s1,s 2, ··· ,s n− 1,s (X1, x 2, · · · , x n−1, x n+1). Therefo...
-
[23]
Spinning AdS Propagators,
M. S. Costa, V. Gon¸ calves, and J. a. Penedones, “Spinning AdS Propagators,” JHEP 09 (2014) 064 , arXiv:1404.5625 [hep-th]
2014 arXiv
-
[24]
Holographic representation of local bulk operators,
A. Hamilton, D. N. Kabat, G. Lifschytz, and D. A. Lowe, “Holographic representation of local bulk operators,” Phys. Rev. D 74 (2006) 066009 , arXiv:hep-th/0606141
2006 arXiv
-
[25]
Higher Spins in AdS and Twistorial Holography,
S. Giombi and X. Yin, “Higher Spins in AdS and Twistorial Holography,” JHEP 04 (2011) 086 , arXiv:1004.3736 [hep-th]
2011 arXiv
-
[26]
AdS/CFT correspondence for the O(N ) invariant critical ϕ 4 model in 3-dimensions by the conformal smearing,
S. Aoki, K. Kawana, and K. Shimada, “AdS/CFT correspondence for the O(N ) invariant critical ϕ 4 model in 3-dimensions by the conformal smearing,” arXiv:2406.15033 [hep-th]
-
[27]
Tensor representations of conformal algebra and conformally covariant operator product expansion,
S. Ferrara, A. F. Grillo, and R. Gatto, “Tensor representations of conformal algebra and conformally covariant operator product expansion,” Annals Phys. 76 (1973) 161–188
1973
-
[28]
Nonhamiltonian approach to conformal quantum field theory,
A. M. Polyakov, “Nonhamiltonian approach to conformal quantum field theory,” Zh. Eksp. Teor. Fiz. 66 (1974) 23–42
1974
-
[29]
Duality in quantum field theory,
G. Mack, “Duality in quantum field theory,” Nucl. Phys. B 118 (1977) 445–457
1977
Reviewed August 12, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.