{"id":"ee62068c-dd11-4fa5-a469-bc80e82ffd46","arxiv_id":"2412.11697","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The second semi-classical coefficient for Toeplitz operators on strictly pseudoconvex CR manifolds equals e^{-2(n+1)f}((1/2)R_scal + (n+1)Delta_b f) times the integral of chi(t)t^{n-1}dt, divided by (2pi)^{n+1}.","lead":"This paper computes the second coefficient in the large-k asymptotic expansion of Toeplitz operators on compact strictly pseudoconvex CR manifolds, expressing it through the Tanaka-Webster scalar curvature and the CR sublaplacian of a volume-dependent function. The formula gives the CR analogue of the well-known second Bergman kernel coefficient and is a checkable local invariant for CR geometry.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.2 rests on the unproved substitution of dV_ξ by dV in [12, §3.3]; the key identity (3.1.20) is asserted, not derived, and Lemma 3.2 does not supply the required second-order stationary phase calculation.","rationale":"The reader's weakest assumption identifies exactly the same load-bearing point: Theorem 1.2 is proved through Theorem 3.3, whose decisive step is a black-box appeal to [12, Section 3.3] under a volume replacement. My reading agrees that this is the most insecure link in the chain. The formula is plausible, and the cross-check in Remark 3.5 with the Bergman kernel coefficient of [9] is genuine independent support, but that consistency check does not establish the local stationary phase identity (3.1.20). I found no independent fatal flaw: the surrounding arguments in Theorem 3.8 and the proof of Theorem 1.2 are coherent once (3.1.20) is granted, and the local symbol bookkeeping in Lemma 3.2 is at least internally consistent. The correct disposition is therefore conditional acceptance pending a full derivation or an explicit published reference for the general-volume stationary phase computation. Since the reader already assigned CONDITIONAL, my stress-test does not change the verdict; it reinforces it.","tokens_in":25365,"tokens_out":10578,"duration_ms":108252,"concrete_test":"Independently derive (3.1.20) from the identity Π∘Π=Π: insert the Boutet de Monvel–Sjöstrand parametrix (1.1.3) with s0 satisfying the Taylor data in Lemma 3.2, change variables so that the stationary point is (u,σ)=(0,1), apply Hörmander's stationary phase expansion up to the term corresponding to s1, and compare the resulting s1(0,0) with (3.1.20)–(3.1.26). If the computation reproduces (3.1.26), the missing step is supplied and the concern is resolved; if an extra term involving derivatives of f appears, then Theorem 1.2 needs revision.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central formula (1.1.11) for b_1^χ depends on Theorem 3.3, and the proof of Theorem 3.3 is the sentence “By replacing the volume dV_ξ with dV in the proof of [12, Section 3.3]”, followed by the displayed identity (3.1.20). This is a substitution claim, not a derivation. In [12] the leading Szegő symbol s0(x,x) is the constant 1/(2π^{n+1}), so many terms in the stationary phase expansion of the composition identity Π∘Π=Π simplify. For arbitrary Reeb-invariant dV=e^{2(n+1)f}dV_ξ, s0(x,x)=e^{-2(n+1)f}/(2π^{n+1}) and s0(x,y) has nontrivial first and second derivatives, which Lemma 3.2 tabulates. However, Lemma 3.2 does not feed those derivatives through the second-order stationary phase operator L^{(1)}; identity (3.1.20) is simply asserted. Since formula (3.1.26) for s1(0,0) is exactly the output of that computation, an undetected f-dependent term in (3.1.20) would propagate directly into the claimed b_1^χ. The paper itself flags the issue in Remark 1.4(ii) (“we need to generalize the result in [12]”), but the generalization is not shown. The consistency check in Remark 3.5 with the known Bergman kernel coefficient is good supporting evidence, but it does not replace the missing calculation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies the semi-classical asymptotic expansion of the kernel of χ_k(A) on a compact strictly pseudoconvex embeddable CR manifold X of dimension 2n+1, where A=Π(-iT)Π is the Toeplitz operator associated with a Reeb vector field T and a Reeb-invariant volume form dV. Building on the full expansion obtained by Herrmann–Hsiao–Marinescu–Shen, the authors compute the second coefficient b_1^χ(x) explicitly in terms of the Tanaka–Webster scalar curvature R_scal, the CR sublaplacian Δ_b f of f=(1/(2(n+1))) log(dV/dV_ξ), and the integral of χ(t)t^{n-1}. The proof proceeds through a structural result (Theorem 1.3) expressing the symbols b_j^χ as polynomials in derivatives of χ and a generalization of the Hsiao–Shen computation of the second coefficient of the Szegő kernel to arbitrary Reeb-invariant volumes (Theorem 3.3/3.4). A consistency check with the known Bergman kernel expansion in the circle-bundle case is given in Remark 3.5.","tokens_in":25725,"tokens_out":11560,"duration_ms":101962,"significance":"If the main formula is correct, it provides the first explicit computation of the sub-leading coefficient in the CR Toeplitz expansion for general Reeb-invariant volumes, a natural analogue of the second Bergman kernel coefficient. The result is local and geometric, and the paper also proves a useful structural statement about the dependence of the symbols on χ (Theorem 1.3). A notable strength is the cross-check in Remark 3.5, where the authors derive the Bergman kernel coefficient and recover a known formula, which gives nontrivial evidence for the correctness of the formula. The main weakness is that the proof of the key Szegő-kernel generalization is not carried out: the identity (3.1.20) is asserted as a direct replacement of the volume in a previously published computation, without showing the necessary stationary-phase calculation.","major_comments":[{"comment":"The proof of Theorem 3.3 consists of the sentence 'By replacing the volume dV_ξ with dV in the proof of [12, Section 3.3], we have (3.1.20)', followed by the formula (3.1.22) claimed to follow from Lemma 3.2. This is not a derivation. Unlike the case dV=dV_ξ, where s0(x,x) is constant and many terms in the stationary-phase expansion simplify, here s0 has non-trivial first derivatives (3.1.11) and the weight e^{2(n+1)f} is present. Applying the second-order operator L^(1) to s0(p,u)s0(u,p)e^{2(n+1)f(u)}σ^n produces terms involving derivatives of s0 and f that are not all tabulated in Lemma 3.2; in particular, mixed second derivatives of s0 are not listed. Since (3.1.22) feeds directly into s1(0,0) and into the claimed formula for b_1^χ, this missing calculation is load-bearing. Remark 3.5 is good consistency evidence, but it does not substitute for the computation.","section":"Section 3, Theorem 3.3, Eqs. (3.1.20)–(3.1.22)"},{"comment":"Theorem 3.4 asserts that s1(x,x) in (3.1.29) is well-defined as a smooth function on X and that the formula holds for any phase satisfying (1.1.2)–(1.1.3), (3.1.27)–(3.1.28). No proof of this invariance is given; the statement appears immediately after Theorem 3.3 with no argument. Since the local computation in Theorem 3.3 is made for a special phase with T=∂_{x_{2n+1}} and λ(x)=1+O(|x|^3), the pointwise result at arbitrary x requires either a separate proof of independence or an explicit statement that the local computation can be transported by CR diffeomorphisms and admissible phase changes. Without this, the passage from Theorem 3.3 to Theorem 3.4, which Remark 1.4(ii) cites as the needed generalization, is not established.","section":"Section 3, Theorem 3.4"}],"minor_comments":[{"comment":"There are several typographical errors: 'vainishes' in Theorems 3.1 and 3.4; in Remark 3.5, 'Le dVΘ' should read 'Let dVΘ'; in (3.1.7), the index pattern ∂⁴φ/(∂z_j∂z̄_ℓ∂z_j∂z̄_ℓ) repeats j and ℓ improperly and should be clarified.","section":"Throughout"},{"comment":"The notation χk(A): is used before being defined; please define it explicitly as the operator with the kernel χk(A)(x,y) e^{ikm0 y_{2n+1}} τ(y_{2n+1}) or give the precise expression used in the subsequent computations.","section":"Section 3, Eq. (3.1.63)"},{"comment":"In the proof of Theorem 3.7, the claim that the remainder Fk(x,y′) remains O(k^{−∞}) after multiplication by e^{−ikm0 φ(x,y′)} is plausible but the argument is compressed; the reader would benefit from the explicit definition of Fk and a statement of the uniformity in (x,y′).","section":"Section 3, Proof of Theorem 3.7"},{"comment":"The phrase 'partial refinement of the main result obtained in [7]' is vague; the proof relies on [7, Lemma 4.2, Lemma 4.3, Theorem 4.11], which are cited but not summarized. Since [7] is a preprint, the authors should state which parts of the proof of Theorem 3.6 are new and ensure the cited results are accessible to the reader.","section":"Section 3, Theorem 3.6"}],"recommendation":"major_revision","confidential_remarks":"The main formula is likely correct given the consistency check in Remark 3.5, but the proof gap in Theorem 3.3/3.4 is substantial: the key stationary-phase computation is asserted rather than carried out. I would require the detailed computation before accepting the manuscript. The heavy reliance on [12] and [7], both with authors overlapping the present paper, is worth the editor's attention, particularly because [7] is a preprint."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a legitimate and useful extension of Hsiao–Shen. It computes the second coefficient b_1^χ in the semi-classical Toeplitz expansion for any Reeb-invariant volume form, adding the explicit volume correction (n+1)Δ_b f to the known dV = dV_ξ formula. The statement is clean, the formula is plausible, and the consistency check with the Bergman kernel coefficient in the circle-bundle case lines up with [9]. Anyone working on CR Szegő/Toeplitz asymptotics should read it.\n\nWhat’s actually new: the dV = dV_ξ case is already in [12], and the circle-bundle specialization appears in [7]/[9]. The genuinely new content is Theorem 1.2/1.5 for arbitrary Reeb-invariant dV, with the f-dependent correction, plus the χ-structure theorem (Theorem 1.3) that decouples the dependence on the cutoff function. These are natural and nontrivial. The citation pattern is fine; relying on [12] and [7] is appropriate since those are published results being generalized, not vague references.\n\nWhere the proof is soft: the proof of Theorem 3.3, which feeds directly into Theorem 1.2, contains the line \"By replacing the volume dV_ξ with dV in the proof of [12, Section 3.3]\" and then asserts the key identity (3.1.20). Lemma 3.2 does tabulate the derivatives of s0, and the final expression (3.1.22) is stated after a short \"From (3.1.10) and (3.1.11)\", but the actual stationary phase computation is not shown. This is a real presentational gap. I would not call it fatal: the computation is standard, the missing algebra is the kind a referee can reproduce, and the agreement with the Bergman kernel coefficient is genuine supporting evidence. Still, a careful referee would reasonably ask for the computation to be written out or for a published derivation to be cited.\n\nBottom line: this is a solid paper with one terse step, not a shaky one. If it crossed my desk as a new submission, I would send it out and the natural decision would be conditional accept, with the stationary phase computation made explicit. I note the arXiv header says it is already published in Analysis and Mathematical Physics, so this is more about how to use it going forward. I would cite it and would be happy to discuss it in a reading group.","headline":"A solid, useful extension of Hsiao–Shen with a real but fillable gap in the stationary phase computation.","tokens_in":26297,"tokens_out":2569,"would_cite":true,"duration_ms":26482,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["32Vxx","32A25","53D50"],"pacs":[],"model":"deepseek-v4-flash","headline":"Second coefficient in CR Toeplitz expansion computed explicitly","keywords":["CR manifolds","Szegő kernels","Toeplitz operators","semi-classical expansion","Tanaka-Webster scalar curvature","CR sublaplacian","strictly pseudoconvex","asymptotic expansion"],"falsifier":"On a strictly pseudoconvex CR manifold where the Szegő kernel or the spectrum of the Toeplitz operator is explicitly known—for example the sphere $S^{2n+1}$ with the standard CR structure and a non-trivial $S^1$-invariant volume form, so that $f\\not\\equiv 0$—one can compute $\\chi_k(A)(x,x)$ directly from the spectral decomposition of $A$ and compare the coefficient of $k^n$ with formula (1.1.11). A disagreement would disprove the formula; an independent case with $\\Delta_b f\\neq 0$ would confirm the volume-correction term.","tokens_in":25125,"feed_emoji":"🧮","tokens_out":14717,"duration_ms":121535,"temperature":0.7,"pith_summary":"Toeplitz operators on compact strictly pseudoconvex embeddable CR manifolds admit a semi-classical expansion of their kernel along the diagonal; the paper determines the second coefficient of that expansion. The second coefficient is the CR counterpart of the second Bergman kernel coefficient that is central in complex geometry, and it is the first coefficient after the leading term that carries geometric information. The formula expresses it as a product of an exponential weight, a combination of half the Tanaka–Webster scalar curvature and $(n+1)$ times the CR sublaplacian of a volume-form log-ratio, and the moment $\\int \\chi(t)t^{n-1}dt$ of the cut-off function. The result extends the known computation of the second Szegő kernel coefficient, which was restricted to the contact volume form, to any Reeb-invariant volume form.","feed_headline":"Explicit second coefficient for CR Toeplitz kernels","feed_subtitle":"It matches known Kähler/CR cases and extends Szegő kernel asymptotics to any Reeb-invariant volume.","key_machinery":"The argument is carried by the Fourier integral representation of the Szegő kernel, written as $\\Pi(x,y)=\\int_0^\\infty e^{it\\varphi(x,y)}s(x,y,t)dt$ with a classical symbol $s\\sim \\sum_{j\\ge0}s_j(x,y)t^{n-j}$ (see (1.1.3)). The load-bearing step is a change of volume: replacing the contact volume $dV_\\xi$ by an arbitrary Reeb-invariant $dV=e^{2(n+1)f}dV_\\xi$ and asserting that the stationary phase computation of the second Szegő coefficient from [12] carries over with the extra factor $e^{2(n+1)f}$ inserted. This yields the identity (3.1.20), where the operator $L^{(1)}$ in the standard stationary phase formula acts on $s_0(0,x)s_0(x,0)e^{2(n+1)f(x)}\\sigma^n$. The paper then proves a uniqueness statement (Theorem 3.7) for the coefficients of the expansion, and composes $\\chi_k(A)$ with a cut-off $\\tau_k(A)$ to extract $a_{1,0}(p,p)=s_1(p,p)$, transferring the Szegő kernel computation to the Toeplitz kernel.","core_discovery":"The central claim is stated in Theorem 1.2. For any cut-off $\\chi\\in C_c^\\infty(\\mathbb{R}_+)$, the diagonal kernel of $\\chi_k(A)$ expands as $\\chi_k(A)(x,x)\\sim \\sum_{j\\ge 0} b_j^\\chi(x)k^{n+1-j}$, and the $j=1$ coefficient is\n$$b_1^\\chi(x)=\\frac{1}{2\\$pi^{{n+1}}$}$e^{{-2(n+1)f(x)}}$\\left(\\frac{1}{2}R_{\\mathrm{scal}}(x)+(n+1)\\Delta_b f(x)\\right)\\int \\chi(t)$t^{{n-1}}$dt,$$\nwhere $f(x)=\\frac{1}{2(n+1)}\\log\\frac{dV(x)}{dV_\\xi(x)}$, $R_{\\mathrm{scal}}$ is the Tanaka–Webster scalar curvature of the contact form $\\xi$, and $\\Delta_b$ is the CR sublaplacian. The proof works by generalizing the computation of the second Szegő kernel coefficient from the volume $dV_\\xi$ to an arbitrary Reeb-invariant volume $dV=e^{2(n+1)f}dV_\\xi$; in that setting the leading Szegő symbol is $s_0=(2\\pi^{n+1})^{-1}e^{-2(n+1)f}$, and the stationary phase identity (3.1.20) yields $s_1(0,0)=(2\\pi^{n+1})^{-1}e^{-2(n+1)f}\\left(\\frac{1}{2}R_{\\mathrm{scal}}(0)+(n+1)\\Delta_b f(0)\\right)$. A structural refinement (Theorem 1.3) shows each symbol $b_j^\\chi(x,y,t)$ has the form $\\sum_s a_{j,s}(x,y)\\chi^{(s)}(t)t^{n+s-j}$, which is used to isolate the $j=1$ term and connect it to $s_1$.","pith_inferences":["The explicit dependence on $f$ suggests a variational reading: shifting the volume form shifts $b_1^\\chi$ by a term in the image of $\\Delta_b$, so one could ask whether a 'balanced' volume form making the correction vanish exists on a given CR manifold; the paper does not address this.","The same volume-replacement method likely produces higher Szegő coefficients as differential polynomials in $f$ and pseudo-Hermitian curvature tensors; checking the next order in the circle-bundle example against the known Bergman kernel expansion would test the method's reach.","Theorem 1.3's separation of $\\chi$ from the geometry means that spectral quantities built from $\\chi_k(A)$, such as weighted traces of functions of $A$, can be computed by integrating the fixed geometric coefficients against moments of $\\chi$; this is a natural next target."],"forward_implications":["For $dV=dV_\\xi$ the formula reduces to $b_1^\\chi(x)=\\frac{1}{2\\pi^{n+1}}\\frac{1}{2}R_{\\mathrm{scal}}(x)\\int\\chi(t)t^{n-1}dt$, recovering the second Szegő coefficient known from [12] and the Sasakian computations.","In the circle-bundle case over a Kähler manifold the same formula reproduces the known second Bergman kernel coefficient (Remark 3.5), bridging the CR expansion and the complex-geometric one.","The coefficient is local: each $a_j(x)$ in the refined expansion depends only on the germ of the CR structure, the contact form, and the volume form at $x$ (Remark 1.6); with $dV=dV_\\xi$ the $a_j$ are pseudo-Hermitian invariants.","The structural form (1.1.13) separates the geometry from the cut-off $\\chi$, so integrals of the kernel against arbitrary smooth weights can be computed from the fixed coefficients $a_{j,s}$ and moments of $\\chi$."],"supporting_citations":[{"why":"Supplies the Fourier integral representation (1.1.3) of the Szegő kernel that the symbol expansion starts from.","marker":"[1]"},{"why":"Provides known Szegő kernel coefficients for quasi-regular Sasakian manifolds with $S^1$ action, used as a comparison case.","marker":"[5]"},{"why":"Establishes the full asymptotic expansion (1.1.7)–(1.1.10) of $\\chi_k(A)$ and the self-adjointness of $A$; the present paper refines its result.","marker":"[7]"},{"why":"Gives the second Bergman kernel coefficient in the complex geometric setting, which the circle-bundle reduction of Remark 3.5 matches.","marker":"[9]"},{"why":"Provides the theorem on changing phases in Szegő kernel oscillatory integrals, used in the proof of Theorem 3.6.","marker":"[11]"},{"why":"Contains the stationary phase computation of the second Szegő coefficient for $dV=dV_\\xi$ that the paper generalizes to arbitrary Reeb-invariant volumes.","marker":"[12]"}],"fun_headline_variants":["Explicit second coefficient in CR Toeplitz expansion","Second coefficient of CR Toeplitz kernels computed","Toeplitz expansion: second term explicit for CR manifolds","CR Toeplitz second coefficient matches Kähler cases"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The formula depends on the assertion that the stationary phase computation of the second Szegő coefficient in [12], originally done for the contact volume $dV_\\xi$, extends verbatim to every Reeb-invariant volume $dV=e^{2(n+1)f}dV_\\xi$ including the extra weight $e^{2(n+1)f}$ in the phase integral and the modified leading symbol $s_0$; if that extension fails, the identity (3.1.20) for $s_1$ and hence Theorem 1.2 would need correction.","fun_headline_variants_meta":{"raw":{"variants":["Explicit second coefficient in CR Toeplitz expansion","Second coefficient of CR Toeplitz kernels computed","Toeplitz expansion: second term explicit for CR manifolds","CR Toeplitz second coefficient matches Kähler cases"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000295,"raw_usage":{"total_tokens":1788,"prompt_tokens":1093,"completion_tokens":695,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":709,"completion_tokens_details":{"reasoning_tokens":626}},"tokens_in":709,"tokens_out":695,"duration_ms":5831,"temperature":1.0,"reasoning_tokens":626,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T14:39:24.345946+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"On a strictly pseudoconvex CR manifold where the Szegő kernel or the spectrum of the Toeplitz operator is explicitly known—for example the sphere $S^{2n+1}$ with the standard CR structure and a non-trivial $S^1$-invariant volume form, so that $f\\not\\equiv 0$—one can compute $\\chi_k(A)(x,x)$ directly from the spectral decomposition of $A$ and compare the coefficient of $k^n$ with formula (1.1.11). A disagreement would disprove the formula; an independent case with $\\Delta_b f\\neq 0$ would confirm the volume-correction term.","supporting_citations":[{"cited_title":"Boutet de Monvel and J","cited_arxiv_id":null,"evidence_quote":"Supplies the Fourier integral representation (1.1.3) of the Szegő kernel that the symbol expansion starts from."},{"cited_title":"Herrmann, C.-Y","cited_arxiv_id":null,"evidence_quote":"Provides known Szegő kernel coefficients for quasi-regular Sasakian manifolds with $S^1$ action, used as a comparison case."},{"cited_title":"Hsiao, On the coefficients of the asymptotic expansion of the kernel of Berezin-Toeplitz quantization, Ann","cited_arxiv_id":null,"evidence_quote":"Gives the second Bergman kernel coefficient in the complex geometric setting, which the circle-bundle reduction of Remark 3.5 matches."},{"cited_title":"Hsiao and G","cited_arxiv_id":null,"evidence_quote":"Provides the theorem on changing phases in Szegő kernel oscillatory integrals, used in the proof of Theorem 3.6."},{"cited_title":"Hsiao and W.-C","cited_arxiv_id":null,"evidence_quote":"Contains the stationary phase computation of the second Szegő coefficient for $dV=dV_\\xi$ that the paper generalizes to arbitrary Reeb-invariant volumes."}],"review_version":1}