{"id":"554eae8c-a35d-46d7-a6f0-6a3c28cc65c7","arxiv_id":"1908.02264","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Existence of dilation-periodic, positive singular solutions of the critical sublaplacian equation on the Heisenberg group is proved for all sufficiently large periods T.","lead":"This paper constructs positive solutions to the critical Yamabe equation on the Heisenberg group that are singular at the origin and repeat self-similarly under dilations, analogous to the known Fowler solutions in Euclidean space. The result supplies a missing building block for the CR Yamabe problem and for gluing constructions of singular Webster metrics.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 5.6 is under-proved: the lemmas do not control J''_T(Ψ_λ) on ∂Ψ_λ/∂λ, so uniform invertibility on its orthogonal complement is not established.","rationale":"The reader's conditional verdict identifies precisely the weakest link. I reviewed Sections 4–6 and found no more fundamental flaw: the periodized family Ψ_λ lies in X_T, the gradient estimate in Proposition 4.1 follows the stated Lorentz/Hölder route, and the final reduced-functional argument follows the Ambrosetti–Malchiodi scheme modulo standard technical details. The unsupported step is Proposition 5.6. Lemmas 5.2–5.5 establish coercivity on V = {e, Ψ_λ}^⊥ and a negative/one-dimensional behavior for Ψ_λ, but the tangent direction e = ∂Ψ_λ/∂λ, which is exactly the direction eliminated in condition (17), is never analyzed. One needs at least an estimate for ⟨J''_T(Ψ_λ)e, e⟩, for ⟨J''_T(Ψ_λ)e, Ψ_λ⟩, or for the angle between e and Ψ_λ to conclude the block on e^⊥ is uniformly invertible. This gap is likely repairable by a standard nondegeneracy-transfer argument, but as written it leaves the contraction mapping in Lemma 6.1 without a proven bounded inverse. Therefore the concern is real and load-bearing, but it does not force a different verdict from the reader's CONDITIONAL; it reinforces the need for revision.","tokens_in":16066,"tokens_out":17224,"duration_ms":192182,"concrete_test":"Derive Proposition 5.6 from Lemmas 5.4–5.5 by writing A = J''_T(Ψ_λ) in the orthonormal basis e = ∂Ψ_λ/∂λ, f = Ψ_λ, and V = span{e,f}^⊥. Show that the displayed estimates imply a uniform lower bound on the smallest singular value of A|_{e^⊥}. In particular, compute or bound the 2×2 block [[⟨Ae,e⟩, ⟨Ae,f⟩], [⟨Af,e⟩, ⟨Af,f⟩]] and the angle between e and f. If this block is uncontrolled, exhibit a symmetric 3×3 matrix satisfying exactly the three lemma inequalities whose restriction to e^⊥ is singular; if no such counterexample exists, provide the missing estimate that excludes it.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The contraction argument in Lemma 6.1 rests on Proposition 5.6, which asserts uniform invertibility of A = J''_T(Ψ_λ) on {∂Ψ_λ/∂λ}^⊥. The preceding lemmas give three things: (i) coercivity of ⟨Au,u⟩ on V = span{∂Ψ_λ/∂λ, Ψ_λ}^⊥, (ii) |⟨AΨ_λ, Ψ_λ⟩| ≥ c||Ψ_λ||^2, and (iii) small cross terms |⟨AΨ_λ, v⟩| for v ∈ V. These estimates do not determine the action of A on e = ∂Ψ_λ/∂λ itself, nor on the two-dimensional block span{e, Ψ_λ}. In particular, they do not control ⟨Ae, e⟩, ⟨Ae, Ψ_λ⟩, or the angle between e and Ψ_λ. Without such control, a direction in e^⊥ mixing e and Ψ_λ can produce a small or zero eigenvalue in the e–Ψ block, and the claimed bounded inverse need not exist. The text says the conclusion follows by “elementary Hilbert space theory”, but that theory only applies once the missing block is estimated. This is exactly the tangent direction isolated in condition (17), and Proposition 5.6 is the load-bearing premise for the reduction in Lemma 6.1.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs positive solutions to the critical equation -Δ_{H^n}u = u^{(Q+2)/(Q-2)} on the Heisenberg group, singular at the origin, that satisfy the homogeneity condition u∘δ_T = T^{-(Q-2)/2}u with T the smallest period. The proof uses a Lyapunov-Schmidt reduction: a family Ψ_λ of approximate solutions is built by periodizing the Jerison-Lee bubble, the gradient of the functional is shown to be small on this family (Proposition 4.1), a nondegeneracy statement for the second differential on the orthogonal complement of the tangent is asserted (Proposition 5.6), and a contraction argument yields a solution of the auxiliary equation (Lemma 6.1). The bifurcation equation is then solved by finding a critical point of the finite-dimensional reduced functional. The paper also proves a Sobolev inequality on the periodic function space X_T using Lorentz spaces.","tokens_in":16255,"tokens_out":15784,"duration_ms":161318,"significance":"If the main theorem is correct, it gives the first analogue of Euclidean Fowler solutions in the Heisenberg group, providing singular periodic solutions that are natural building blocks for more general singular solutions. The use of Lorentz spaces to control the T-dependence of the Sobolev constant and the periodization construction are interesting and potentially reusable techniques. The argument is parameter-free and relies on external classification and nondegeneracy results rather than circular reasoning. However, the central invertibility step, Proposition 5.6, is not adequately proved, so the existence theorem is not fully established as written.","major_comments":[{"comment":"The proof of uniform invertibility of J''_T(Ψ_λ) on the orthogonal complement of ∂Ψ_λ/∂λ is incomplete. The estimates in Lemmas 5.3–5.5 are all derived under the two orthogonality conditions (17) and (18), i.e., on the subspace V = span{∂Ψ_λ/∂λ, Ψ_λ}^⊥. Lemma 5.5 provides (i) coercivity on V, (ii) a lower bound |d²J_T(Ψ_λ)[Ψ_λ,Ψ_λ]| ≥ c‖Ψ_λ‖², and (iii) smallness of the cross term |d²J_T(Ψ_λ)[Ψ_λ,u]| for u ∈ V. These estimates do not determine the quadratic form on the two-dimensional block span{∂Ψ_λ/∂λ, Ψ_λ}, nor on the component of Ψ_λ orthogonal to ∂Ψ_λ/∂λ. In particular, no estimate is given for ⟨J''_T(Ψ_λ)∂Ψ_λ/∂λ, ∂Ψ_λ/∂λ⟩, for ⟨J''_T(Ψ_λ)∂Ψ_λ/∂λ, Ψ_λ⟩, or for the angle between these two vectors. Without this information, a direction in the orthogonal complement of ∂Ψ_λ/∂λ mixing Ψ_λ with the tangent can have small or zero eigenvalue, and the claimed bounded inverse need not exist. The sentence 'It follows from the preceding lemmas and elementary Hilbert space theory' does not fill this gap, because the standard block-diagonalization argument requires control on the missing block. Since Proposition 5.6 is the load-bearing premise for the contraction argument in Lemma 6.1, the proof of Theorem 1.1 is incomplete as written.","section":"Section 5, Proposition 5.6"},{"comment":"The statement 'from the proof of Proposition 5.6 follows that J(ω_λ) has Morse index one on {λ∂ω_λ/∂λ}^⊥' is not a consequence of the lemmas as they stand. The Morse index of the linearized operator on the orthogonal complement of the kernel requires the full spectral decomposition of that operator, which is exactly what is missing from Proposition 5.6. While this assertion may become a corollary once the missing estimates on the tangent block are supplied, it is not justified by the present proof and should not be used as part of the argument for positivity or for the structure of the solution set.","section":"Section 6, proof of Theorem 1.1"}],"minor_comments":[{"comment":"There are several typos and spacing errors, e.g., 'similar to t he Fowler' in the abstract, 'structure structure' on page 2, and 'In the this case a lso' on page 2; these should be corrected.","section":"Abstract and Introduction"},{"comment":"The proof of the third inequality in Lemma 5.4 is summarized as 'by Lemma 5.3, by equations (15) and (16), and elementary linear algebra'; this step is quite compressed and would benefit from a detailed derivation, especially because it is used to control the cross terms in Lemma 5.5.","section":"Lemma 5.4"},{"comment":"The notation ~X_T and ~Z_T is introduced in Section 2 and used in Section 6, but the subscript T is sometimes omitted; for consistency, the spaces and curves should be written with the subscript throughout.","section":"Notation"},{"comment":"The statement of Proposition 3.2 says '∇u ∈ L^{2,∞}(H^n)' without specifying that this is the subriemannian gradient; the proof correctly uses subriemannian derivatives, so the statement should be adjusted for clarity.","section":"Proposition 3.2"},{"comment":"The final sentence 'The last assertion follows by construction' is vague; the authors should explicitly indicate that the constructed solution has smallest period T and briefly explain why no smaller period occurs.","section":"Proof of Theorem 1.1"}],"recommendation":"major_revision","confidential_remarks":"The main result is plausible and the construction is natural, but the gap in Proposition 5.6 is substantive and affects the core of the proof. The missing estimates on the tangent block appear fixable—for instance, by directly estimating the action of J''_T(Ψ_λ) on ∂Ψ_λ/∂λ and on the mixed terms with Ψ_λ, or by reparameterizing the approximate curve so that the tangent is better controlled. I would encourage the editor to send the paper back for a major revision rather than reject it, as the rest of the argument is coherent and the result would be a valuable contribution. The paper is an arXiv preprint from 2019; the authors should also verify whether any related work has appeared in the interim."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is the first construction of dilation-periodic singular solutions on the Heisenberg group, and the result should be true. But the key invertibility step, Proposition 5.6, is asserted rather than proved, and the gap is real. I agree with your stress-test note. What is new and good: the idea is to periodize the Jerison-Lee bubbles and run a Lyapunov-Schmidt reduction on the space X_T of functions with the right dilation homogeneity. The Lorentz-space Sobolev bound on X_T with explicit log T dependence (Section 3) is a useful tool, and the gradient estimate for the approximate solution (Section 4) is careful. Lemmas 5.2-5.5 are mostly standard cut-off and comparison arguments; I did not spot an error there. The soft spot: Proposition 5.6 does not follow from the lemmas. The lemmas establish coercivity on V = span{tau, Psi}^perp, a lower bound for <A Psi, Psi>, and small cross terms <A Psi, v> for v in V. That controls the two-by-two block except for the direction in tau^perp that is the projection of Psi. To control that direction you need either <A tau, tau>, <A tau, Psi>, or a bound on the angle between tau and Psi. None is provided. Without it, the Schur complement could vanish, so the asserted uniform invertibility on tau^perp is not established. 'Elementary Hilbert space theory' does not apply until that block is estimated. Since Lemma 6.1's contraction argument uses exactly this invertibility, the proof as written is incomplete. The gap looks fixable: one could impose a Pohozaev-type normalization that makes the tangent direction orthogonal to Psi, or prove coercivity on tau^perp directly using the nondegeneracy of the single bubble plus the smallness of the periodization error. But that is real additional work, not a typo-level fix. For whom: anyone working on CR Yamabe gluing or blow-up analysis in subelliptic PDE. The construction is a natural building block, and if the gap is repaired, it will be cited. I would not cite it as a black box until then. Recommendation: send to peer review, conditional. The result is worth referee time; a competent referee can check the gap and the author can repair it. Not a desk reject.","headline":"A genuinely new Fowler-type construction for the Heisenberg group whose main theorem is plausible, but Proposition 5.6's invertibility claim is under-proved and is the load-bearing step.","tokens_in":16855,"tokens_out":7621,"would_cite":false,"duration_ms":81381,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35R03","35H20","35J20","35J61"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper constructs positive singular solutions of the critical Yamabe equation on the Heisenberg group that are periodic under dilation, for every sufficiently large period.","keywords":["Heisenberg group","Fowler solutions","singular periodic solutions","CR Yamabe equation","Lyapunov-Schmidt reduction","critical exponent","dilation periodicity","Lorentz spaces"],"falsifier":"Compute the quadratic form $d^2J_T(\\Psi_\\lambda)[\\partial\\Psi_\\lambda/\\partial\\lambda,\\partial\\Psi_\\lambda/\\partial\\lambda]$ on the periodized bubble and check whether it stays bounded away from zero, uniformly in $\\lambda$, for large $T$; if it can approach zero or change sign, Proposition 5.6's bounded-inverse conclusion fails and the contraction argument in Lemma 6.1 has no fixed point. Alternatively, test the asserted homogeneous solution numerically by direct integration of the subelliptic equation on a fundamental domain $\\Omega_T$ for a sequence of large $T$.","tokens_in":15771,"feed_emoji":"","tokens_out":8828,"duration_ms":79483,"temperature":0.7,"pith_summary":"The paper aims to transplant Euclidean Fowler solutions to the Heisenberg group: it claims that the critical equation $-\\Delta_{\\mathbf{H}^n} u = u^{(Q+2)/(Q-2)}$ on $\\mathbf{H}^n\\setminus\\{0\\}$ has positive singular solutions satisfying the dilation-periodicity $u\\circ\\delta_T = T^{-(Q-2)/2}u$ for every sufficiently large period $T$, with $T$ the smallest period. The interest is that the Heisenberg sublaplacian is not rotationally invariant, so the standard Euclidean reduction to radial ODEs is unavailable; the paper instead builds the solutions as small perturbations of an infinite sum of dilated copies of the known global regular bubble. A Lyapunov-Schmidt reduction on the space of dilation-periodic functions turns the problem into a contraction argument. A sympathetic reader would take the main contribution to be a Fowler-type existence result in a noncommutative Carnot setting, opening a route toward more general singular solutions.","feed_headline":"Heisenberg group gains Fowler-type periodic singular solutions","feed_subtitle":"Fowler-type solutions in the Heisenberg group exist for every large period, extending the Euclidean construction.","key_machinery":"The load-bearing object is the family of approximate solutions $\\Psi_{\\lambda,T}=\\sum_{k\\in\\mathbb{Z}}\\omega_{\\lambda/T^k}$, the periodization of the global regular bubble $\\omega_\\lambda$ (the explicit positive solution of the critical equation on $\\mathbf{H}^n$). This sum lies in the dilation-periodic space $X_T=\\{u: u\\circ\\delta_T=T^{-(Q-2)/2}u\\}$ and forms a closed curve $Z_T$ parametrized by $\\lambda\\in(0,\\infty)$; the paper shows its members are almost critical points of the variational functional $J_T$ in the sense that $\\|\\nabla J_T(\\Psi_\\lambda)\\|\to0$ uniformly as $T\\to\\infty$. Around this curve the argument runs a Lyapunov-Schmidt reduction: a nondegeneracy estimate for the linearized operator $J_T''(\\Psi_\\lambda)$ on the orthogonal complement of the tangent direction is transferred from the known nondegeneracy of $J''(\\omega_\\lambda)$ on the whole Heisenberg group, and the auxiliary equation is solved by contraction. An explicit Sobolev constant on $X_T$, obtained through Lorentz-space convolution estimates, controls all error terms uniformly in $\\lambda$ and $T$.","core_discovery":"On the paper's own terms, the central discovery is Theorem 1.1: there is a threshold $T_0$ such that for every $T\\ge T_0$ the equation $-\\Delta_{\\mathbf{H}^n} u = u^{(Q+2)/(Q-2)}$ on $\\mathbf{H}^n\\setminus\\{0\\}$, with $Q=2n+2$, admits a positive solution $u$ satisfying $u\\circ\\delta_T = T^{-(Q-2)/2}u$ and having no smaller dilation period. The solution is built as $\\Psi_{\\lambda}+w(\\lambda)$, where $\\Psi_\\lambda$ is the periodized sum $\\sum_{k\\in\\mathbb{Z}}\\omega_{\\lambda/T^k}$ of bubbles and $w(\\lambda)$ is a correction found by the Lyapunov-Schmidt method. The sum is singular at the origin, and the periodicity is with respect to the natural anisotropic dilations $\\delta_T(z,t)=(Tz,T^2t)$ of the Heisenberg group.","pith_inferences":["The same periodization-plus-contraction scheme is likely to work for other Carnot groups with a dilation group and a known nondegenerate bubble, provided the linearized nondegeneracy is verified on the full orthogonal complement of the tangent direction.","A natural quantitative next step would be to track the threshold $T_0$: the proof uses uniform-in-$\\lambda$ estimates and does not identify the sharp period below which periodic solutions disappear, so the true existence boundary remains open.","The Lorentz-space Sobolev constant with explicit $(\\log T)$ dependence may be reusable as a black box for other periodic variational problems on stratified homogeneous groups, independent of this particular equation."],"forward_implications":["For every sufficiently large $T$, the Heisenberg group carries a positive singular solution of the critical CR Yamabe equation whose dilation-period is exactly $T$, not merely a divisor of it.","The constructed solutions scale like $|x|^{-(Q-2)/2}$ near the origin, matching the singular cylindrical solution and giving the Heisenberg analogue of Euclidean Fowler/Delaunay-type ends.","Because the curve $\\tilde Z_T=\\{\\Psi_\\lambda+w(\\lambda)\\}$ consists entirely of critical points, the result yields a one-parameter family of distinct periodic solutions rather than an isolated example.","The Morse-index argument places each solution at index one within the symmetric subspace, consistent with the known index of the bubble, so the solutions are natural building blocks for gluing constructions of more general singular solutions."],"supporting_citations":[{"why":"Supplies the explicit global regular solution $\\omega_\\lambda$ and its classification, from which the periodized family $\\Psi_\\lambda$ is built.","marker":"[JL]"},{"why":"Provides the nondegeneracy of the linearized operator $J''(\\omega_\\lambda)$ on the whole Heisenberg group, which Proposition 5.6 transfers to $J_T''(\\Psi_\\lambda)$.","marker":"[MU]"},{"why":"Gives the Lyapunov-Schmidt reduction and contraction argument used to solve the auxiliary and bifurcation equations.","marker":"[AM]"},{"why":"Provides the Euclidean Fowler solutions that the paper's theorem extends to the Heisenberg group.","marker":"[CGS]"},{"why":"Supplies the Young-O'Neil convolution inequality for Lorentz spaces used to prove the Sobolev estimate on $X_T$.","marker":"[ON]"},{"why":"The symmetric criticality principle that lets the author solve the problem on the $U(n)$-invariant subspace.","marker":"[Pal]"},{"why":"Used for the Morse-index-one property and part of the sign/positivity argument.","marker":"[BCD]"},{"why":"Bony's maximum principle supplies the strict positivity of the constructed solution.","marker":"[Bon]"}],"fun_headline_variants":["Fowler solutions on Heisenberg group: periodic and singular","Periodic singular solutions for critical Heisenberg equation","Lyapunov-Schmidt finds Fowler-type periodic singular solutions","Critical Heisenberg equation yields periodic singular solutions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument collapses if the asserted uniform invertibility of the linearized operator $J_T''(\\Psi_\\lambda)$ on the space orthogonal to the tangent direction $\\partial\\Psi_\\lambda/\\partial\\lambda$ fails; the proof's coercivity estimates are given on the subspace orthogonal to both that tangent and $\\Psi_\\lambda$, and on $\\Psi_\\lambda$ itself, so the invertibility in the remaining tangent direction is the load-bearing premise that is asserted rather than explicitly displayed.","fun_headline_variants_meta":{"raw":{"variants":["Fowler solutions on Heisenberg group: periodic and singular","Periodic singular solutions for critical Heisenberg equation","Lyapunov-Schmidt finds Fowler-type periodic singular solutions","Critical Heisenberg equation yields periodic singular solutions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000318,"raw_usage":{"total_tokens":1757,"prompt_tokens":864,"completion_tokens":893,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":480,"completion_tokens_details":{"reasoning_tokens":831}},"tokens_in":480,"tokens_out":893,"duration_ms":50195,"temperature":1.0,"reasoning_tokens":831,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:51:38.874668+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the quadratic form $d^2J_T(\\Psi_\\lambda)[\\partial\\Psi_\\lambda/\\partial\\lambda,\\partial\\Psi_\\lambda/\\partial\\lambda]$ on the periodized bubble and check whether it stays bounded away from zero, uniformly in $\\lambda$, for large $T$; if it can approach zero or change sign, Proposition 5.6's bounded-inverse conclusion fails and the contraction argument in Lemma 6.1 has no fixed point. Alternatively, test the asserted homogeneous solution numerically by direct integration of the subelliptic equation on a fundamental domain $\\Omega_T$ for a sequence of large $T$.","supporting_citations":[],"review_version":1}