{"id":"42801617-537b-4a49-a6ba-5f021432458c","arxiv_id":"2412.11209","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"For a broad class of nonlinear wave equations on the Kerr-AdS exterior, exponentially decaying event horizon data is shown to yield future-global smooth solutions with exponential decay, even beyond the Hawking-Reall bound.","lead":"This mathematics paper proves that on a rotating black hole in anti-de Sitter space, specially chosen disturbances that fade very fast along the event horizon produce smooth wave solutions that also fade exponentially, even in a regime where generic waves are expected to grow. The result matters because such exceptional solutions may help physicists understand the stability of black holes with a negative cosmological constant.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Bootstrap improvement in Theorem 5.2 does not close for arbitrary finite D_k; the proof requires an unstated smallness condition on the scattering data that Theorem 4.2 does not impose.","rationale":"The reader's weakest assumption was local well-posedness of the finite-in-time characteristic problems, which the paper explicitly delegates to an adaptation of [19]. That gap is real and acknowledged, but it is plausibly repairable by a standard argument. The more load-bearing issue is the bootstrap/Grönwall closure in Section 5.3: as written, the improvement from (5.4) to (5.8) requires the right-hand side of (5.14) to be satisfiable, and direct inspection shows the Grönwall factor 1 + a/(2Υ−a) grows with M·D_k, forcing a smallness condition on D_k (or a lower bound on t*_0 depending on D_k) that never appears in Theorem 4.2. This affects the theorem's quantifiers and cannot be repaired by cosmetic changes; the statement must either impose smallness or allow t*_0 to depend on the data. The boundary-attainment gap noted in Section 2.4.1 is also relevant for α > 5/4, but it is secondary to the bootstrap issue. I therefore disagree with the reader's identification of the weakest assumption, while the overall verdict remains CONDITIONAL pending a corrected statement and proof.","tokens_in":39462,"tokens_out":21285,"duration_ms":203200,"concrete_test":"In Theorem 5.2, write the bootstrap assumption with constant M and derive the inequality for improvement explicitly at τ = t*_0. Replace h_H+ by a one-parameter family A·h_0 with D_k scaling as A², and solve M/2 ≥ C_0 (1 + C_0(3M e^{−B_k t*_0}A²D_{k,0} + C_{κ,k})/(2Υ−...)) e^{(2Υ−B_k)t*_0} for M. If the maximal admissible A is finite for fixed t*_0, B_k, Υ, then Theorem 4.2's 'all finite D_k' claim fails and a smallness hypothesis must be added.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Theorem 4.2 asserts existence for every smooth scattering data with D_k < ∞ and Υ_{κ,k} ≥ B_k, where B_k depends only on k and κ; no smallness of D_k is stated. The only closing mechanism is the bootstrap improvement in Section 5.3. The Grönwall estimate (5.12) gives ||ψ_i||²_{H^k(Σ_τ)} ≤ C exp(−2Υτ)D_k (1 + a/(2Υ−a)) with a = C(3C exp(−B_k τ)D_k + C_{κ,k}). To reach the improved bound (5.13), the bootstrap constant M must satisfy M/2 ≥ C_0 (1 + a/(2Υ−a)) e^{(2Υ−B_k)τ}. Because a depends linearly on M D_k, this inequality has a solution for M only if D_k is below a threshold depending on t*_0, B_k, and the energy constants; for fixed t*_0 and arbitrarily large D_k, no choice of M can close the bootstrap. The proof instead asserts (5.14) holds for C and B_k sufficiently large, but increasing B_k also increases the exponential factor e^{(2Υ−B_k)τ} ≥ e^{B_k τ} (since Υ ≥ B_k), so it cannot compensate for large D_k. Thus the proof silently requires D_k to be sufficiently small, or t*_0 to be chosen large depending on D_k. This is load-bearing because the theorem's central quantifier covers all finite D_k; without smallness, quadratic nonlinearities may blow up before the backward evolution reaches t*_0.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs exponentially decaying solutions of the massive nonlinear wave equation □_g ψ + αψ = F(ψ,∂ψ) on Kerr-AdS exteriors by prescribing exponentially decaying scattering data on the future event horizon and solving backwards, with Dirichlet boundary conditions at the conformal boundary. The main result, Theorem 4.2, asserts that for k ≥ 9 and any smooth scattering data with finite weighted norm D_k and associated decay rate Υ_{κ,k} ≥ B_k (B_k sufficiently large depending on k and κ), there exists a global-to-the-future solution in C H^k_{KAdS} satisfying the exponential decay estimate (4.2) and the L∞ bounds (4.3). The proof proceeds through finite-in-time approximating problems, a bootstrap argument for uniform exponential decay, and a convergence argument. The paper also discusses implications for quasinormal modes and for the Einstein equation in harmonic gauge.","tokens_in":39712,"tokens_out":14729,"duration_ms":113049,"significance":"If valid, the result would be significant: it provides an infinite-dimensional class of exponentially decaying classical solutions to a semilinear wave equation on Kerr-AdS, including outside the Hawking-Reall bound where forward mode solutions grow exponentially. The physical-space approach avoids frequency analysis and is insensitive to trapping and superradiance. The explicit four-level hierarchy of elliptic estimates (Propositions 3.6–3.12), the L∞ Sobolev embedding (Theorem 3.14), and the precise energy estimates (Proposition 3.16) are valuable technical contributions. The paper is also commendably explicit about its assumptions and about two admitted gaps: local well-posedness of the approximating problems and attainment of the Dirichlet boundary condition. However, as detailed in the major comments, the proof of the main theorem does not close as written for the stated quantifiers.","major_comments":[{"comment":"The bootstrap improvement does not close for arbitrary finite D_k. In (5.11), the coefficient a(C_{κ,k}) contains the term C·3C exp(−B_k τ)D_k, so the condition 2Υ_{κ,k} > a and the factor (1 + a/(2Υ_{κ,k}−a)) in (5.12) can only be controlled if D_k is sufficiently small or if B_k (or t*_0) is chosen depending on D_k. Theorem 4.2 fixes t*_0 and makes B_k depend only on k and κ, so the estimate (5.13) does not follow for all data satisfying (4.1). This is load-bearing because Theorem 4.2 is the main result.","section":"Section 5.3, Eqs. (5.10)–(5.13)"},{"comment":"In passing from the first to the second displayed inequality of (5.10), the pointwise term ||ψ_i||^4_{H^{k-1}(Σ_τ)} is dropped without justification. If this term is retained and bounded using the bootstrap assumption (5.4), it contributes a factor of order D_k^2 exp(−2B_k τ), which makes the improved estimate (5.13) even more restrictive and again requires a smallness condition or a τ-dependent largeness condition. The derivation of (5.11) from (5.10) is therefore not established.","section":"Section 5.3, Eq. (5.10)"},{"comment":"Existence of the approximating solutions ψ_i to the finite-in-time characteristic initial-boundary value problems (5.1) is assumed from an unspecified 'adaptation' of [19], which treats spacelike initial data. No proof or precise statement is given for the characteristic setting with the nonlinearity (2.17). Since the iteration, bootstrap, and convergence arguments in Sections 5.2–5.4 all rely on these ψ_i, the proof of Theorem 4.2 lacks a necessary foundation.","section":"Sections 1.3.3 and 5.1"},{"comment":"The paper explicitly notes that it does not prove that the constructed solutions attain the Dirichlet boundary condition r^{3/2−s}ψ|_I = 0, because the estimates only give r^{−1/2} decay. For α in (5/4, 9/4), this decay is insufficient to ensure the boundary condition in the classical sense, yet the boundary condition is used throughout Section 3 to drop boundary terms (for example in Propositions 3.6–3.12 and Proposition 3.16) and in the Sobolev embedding Theorem 3.14 via (3.30). Thus Theorem 4.2 is not fully proven for the stated mass range.","section":"Section 2.4.1 and Theorem 4.2"}],"minor_comments":[{"comment":"The phrase 'for B_n = C_{κ,n}' after the estimates (5.9) is unclear and likely a typo; the decay rate B_k should be the same as in Theorem 5.1.","section":"Theorem 5.2, after (5.9)"},{"comment":"The displayed inequality (5.14) is garbled and unreadable; it should be rewritten with explicit algebraic conditions.","section":"Eq. (5.14)"},{"comment":"The Arzelà-Ascoli argument is applied on the non-compact domain [t*_0, ∞) × {r ≥ r_+} × S²; a diagonal subsequence argument on a compact exhaustion should be stated.","section":"Section 5.4"},{"comment":"The multi-index notation in the H^k_{KAdS} norm is ambiguous: ρ is written with two components while elsewhere multi-indices have three components; the intended indexing should be clarified.","section":"Definition 2.1, Eq. (2.9)"},{"comment":"The notation F^{μν} e_μψ e_νψ is ambiguous because the same indices are used for the tensor components and the frame vectors; this should be disambiguated.","section":"Section 2.4.2, Eq. (2.17)"}],"recommendation":"major_revision","confidential_remarks":"This is a promising manuscript with a detailed energy method, but the proof of the main theorem has three significant gaps: the bootstrap does not close for the stated quantifiers, local well-posedness of the characteristic finite-time problems is not proved, and the Dirichlet boundary condition is not attained for α > 5/4. These gaps are potentially fixable within the manuscript's scope by adding a smallness condition on D_k (or allowing B_k to depend on D_k and t*_0), proving or properly citing a characteristic well-posedness theorem, and either restricting the mass range or supplying an argument that the constructed solutions satisfy the boundary condition. I therefore recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a real construction, not a mirage. Transferring the Dafermos–Holzegel–Rodnianski backwards scattering method to Kerr-AdS with nonlinearities is a worthwhile contribution, and the hierarchy of elliptic estimates in Section 3 is laid out carefully. The paper is honest about its two main gaps: local well-posedness of the finite-in-time characteristic problems is adapted from Holzegel's spacelike result rather than proven, and attainment of the Dirichlet boundary condition is not shown for α>5/4. Those are standard in this literature and openly flagged.\n\nThe soft spot that worries me more is in the bootstrap. Theorem 4.2 asserts existence for every smooth scattering data with finite D_k and Υ≥B_k, with no smallness of D_k. But the closing of the bootstrap in Theorem 5.2 requires the Grönwall constant a = C(3C exp(−B_k τ)D_k + C_{κ,k}) to be less than 2Υ and the ratio in (5.12) to beat the improved constant. For τ near t*_0, the exponential exp((B_k−2Υ)τ) cannot compensate for D_k arbitrarily large: the condition involves D_k times the bootstrap constant, and increasing B_k weakens the exponential factor. So the proof silently requires D_k sufficiently small (or t*_0 chosen large depending on D_k). That changes the quantifier in the main theorem. This is load-bearing because the nonlinearity is quadratic and may blow up before the backward evolution reaches t*_0.\n\nI want to stress that the core architecture—backwards evolution, redshift/blueshift, elliptic hierarchy, nonlinear estimate Proposition 3.17—is plausible and well explained. The gap is in the iteration closing, not in the overall strategy. It may well be repairable by a two-constant bootstrap and a statement with t*_0 depending on the data. But as written, Theorem 4.2 is not proven as stated.\n\nWho gets value: people working on AdS stability and scattering constructions will want to read this despite the gap. It deserves a serious referee; an editor should send it to review rather than desk-reject. The referee should focus on the bootstrap quantifier and the well-posedness adaptation.","headline":"A serious, well-written scattering construction whose main theorem currently overshoots its proof: the bootstrap closing needs a smallness condition on the data that Theorem 4.2 does not state.","tokens_in":40307,"tokens_out":5325,"would_cite":false,"duration_ms":47252,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35L05","35Q75","83C30","83C57"],"pacs":[],"model":"deepseek-v4-flash","headline":"By prescribing exponentially decaying data on the event horizon and evolving backwards, the paper constructs exponentially decaying solutions of a nonlinear massive wave equation on Kerr–Anti-de Sitter exteriors, including fast-spinning…","keywords":["Kerr–Anti-de Sitter","nonlinear wave equation","Klein–Gordon equation","backwards scattering","event horizon data","Dirichlet boundary conditions","exponential decay","blueshift"],"falsifier":"Run the linear ($\\mathcal{F}=0$) backwards evolution on Schwarzschild–AdS with smooth horizon data decaying like $e^{-B t_*}$ for a rate $B$ just above the constant $C_\\kappa$ appearing in the redshift-vector energy estimate; if the $k$-th order hypersurface energy fails to decay exponentially, the central estimate (4.2) is false. An analytic check of whether the characteristic-data local well-posedness asserted in Section 5.1 really follows from the cited spacelike-data result would also settle the construction.","tokens_in":39147,"feed_emoji":"🕳️","tokens_out":11507,"duration_ms":96253,"temperature":0.7,"pith_summary":"On the exterior of a rotating anti-de Sitter black hole (Kerr–AdS), general solutions of the massive wave equation decay at best logarithmically, and outside the Hawking–Reall bound the linear problem even admits exponentially growing modes. The paper nevertheless constructs a large class of smooth solutions that decay exponentially to the future. It does so by prescribing exponentially decaying scattering data on the future event horizon, imposing Dirichlet data at the conformal boundary, and evolving backwards in time. The main theorem yields such solutions for regularity $k \\ge 9$ whenever the data satisfy a weighted exponential-decay condition with rate $B_k$ sufficiently large compared with the surface gravity, and it gives explicit exponential energy and pointwise bounds. No symmetry assumptions or null condition are required, so the class carries the full functional degrees of freedom of the scalar equation and is positioned for adaptation to the Einstein vacuum equation in a suitable gauge.","feed_headline":"Exponentially decaying waves exist on Kerr–AdS black holes","feed_subtitle":"A backwards construction finds decaying nonlinear waves even where growing modes exist.","key_machinery":"The argument is carried by a backwards scattering iteration. Exponentially decaying data are posed on the future event horizon, and the redshift vector field $N=T+\\xi(r)Y$—a uniformly timelike vector field equal to $\\partial_{t_*}$ at infinity—is used as a multiplier to produce coercive energy estimates; the assumed decay rate $B_k$ is chosen larger than the blueshift growth constant $C_\\kappa$ so that a Grönwall argument closes. A four-level hierarchy of elliptic $L^2$ estimates (Propositions 3.7, 3.9, 3.10, 3.12) controls higher spatial derivatives on the far region, on the near region, and then on the whole exterior, and Sobolev embeddings convert these into the pointwise bounds. The theorem is assembled from finite-in-time problems (5.1) with truncated data, solved via a bootstrap that improves the exponential decay constant, with convergence by Arzelà–Ascoli.","core_discovery":"The central claim is that the backwards scattering construction works on Kerr–AdS: given smooth scattering data $h_{H^+}$ on the future event horizon that decay exponentially in $t_*$ at a sufficiently fast rate, and vanishing Dirichlet data on the timelike conformal boundary $\\mathcal{I}$, there exists a classical solution $\\psi$ of $\\Box_g\\psi+\\alpha\\psi=\\mathcal{F}(\\psi,\\partial\\psi)$ on the exterior, globally to the future of some slice, with energy decaying as $\\|\\psi\\|^2_{H^k_{KAdS}(\\Sigma_{t_*})} \\le C e^{-B_k t_*} D_k$ together with the $L^\\infty$ bounds (4.3) for $r^{1/2}D^{\\sigma}\\psi$. This is proved for $k\\ge 9$, for masses $\\alpha<9/4$ satisfying the Breitenlohner–Freedman bound, and for nonlinearities quadratic in unit derivatives of the form (2.17). The construction does not assume the Hawking–Reall bound, so it covers fast-spinning exteriors in which the forward linear problem has exponentially growing modes; the resulting solutions are exceptional rather than generic, since even the linear forward problem decays only logarithmically in general, and prescribing data that vanish after finite time gives nontrivial solutions outside the span of quasinormal modes.","pith_inferences":["If Theorem 4.2 is correct, exponentially decaying solutions are a non-generic but infinite-dimensional family inside the space of finite-energy solutions; characterizing their Cauchy data as a set is an open problem that numerical searches could probe.","The linear solutions outside the quasinormal-mode span suggest that mode decompositions miss part of the Kerr–AdS solution space, which may matter for stability arguments that rely on quasinormal-mode completeness.","The exponential-rate threshold $B_k>C_\\kappa$ points to a possible phase transition: polynomial horizon decay would likely produce horizons that are singular in the backwards evolution, analogous to the asymptotically flat weak-null-singularity picture; extending the bootstrap to slower decay rates would test this.","In the extremal limit $\\kappa\\to 0$ the blueshift constant vanishes, so the method may admit slower (possibly polynomial) horizon decay there; this is a concrete extension the paper leaves implicit."],"forward_implications":["Exponentially decaying nonlinear waves exist on every subextremal Kerr–AdS exterior allowed by (1.2), including angular momenta violating the Hawking–Reall bound where the forward linear problem has growing modes.","For the linear equation $\\mathcal{F}=0$, the construction produces exponentially decaying solutions that do not lie in the span of quasinormal modes; data vanishing after finite time yield solutions that vanish to the future of a late slice.","The class of solutions is large enough to exhibit the full functional degrees of freedom of the problem, with no symmetry and no null condition imposed on the nonlinearity.","Quantitative control is achieved: the $H^k_{KAdS}$ energy decays like $e^{-B_k t_*}$ and $r^{1/2}$-weighted $L^\\infty$ norms of derivatives up to order $k-3$ and $k-4$ satisfy the same exponential bound.","The paper argues that the semilinear framework is designed to extend to quasilinear systems such as the Einstein vacuum equation in harmonic gauge, where the relevant masses lie in the allowed range."],"supporting_citations":[{"why":"It supplies the backwards scattering template from the asymptotically flat setting that the paper adapts to Kerr–AdS.","marker":"[6]"},{"why":"It provides the linear well-posedness result whose characteristic-data adaptation is assumed for the finite-in-time problems (5.1).","marker":"[19]"},{"why":"It establishes the sharp logarithmic decay of general forward linear waves under the Hawking–Reall bound, setting the contrast for the exceptional class.","marker":"[21]"},{"why":"It gives the quasimode lower bound showing that general linear solutions decay at best inverse logarithmically.","marker":"[22]"},{"why":"It constructs exponentially growing mode solutions outside the Hawking–Reall bound, the regime in which the new construction still applies.","marker":"[9]"},{"why":"It fixes the Breitenlohner–Freedman bound and positive-energy framework underlying the coercivity and Hardy-type estimates.","marker":"[3]"},{"why":"It provides the energy-momentum tensor identities, the T-energy coercivity argument, and the modified Hardy inequality that Section 3.1 adapts.","marker":"[20]"},{"why":"It supplies the L∞ Sobolev embedding theorem on the spheres used to derive the pointwise bounds on normalised derivatives.","marker":"[18]"}],"fun_headline_variants":["Decaying waves on Kerr–AdS even where modes grow","Backwards scattering yields decaying Kerr–AdS waves","Fast-spinning Kerr–AdS admits decaying nonlinear waves","Exponential decay on Kerr–AdS beyond Hawking–Reall","Backwards trick finds decaying waves on Kerr–AdS"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The construction assumes without proof that each finite-in-time nonlinear problem (5.1) with characteristic horizon data is locally well-posed, citing an adaptation of a linear well-posedness result that treats spacelike data; if that adaptation fails or needs extra conditions, the approximating sequence and hence the existence theorem collapse.","fun_headline_variants_meta":{"raw":{"variants":["Decaying waves on Kerr–AdS even where modes grow","Backwards scattering yields decaying Kerr–AdS waves","Fast-spinning Kerr–AdS admits decaying nonlinear waves","Exponential decay on Kerr–AdS beyond Hawking–Reall","Backwards trick finds decaying waves on Kerr–AdS"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000788,"raw_usage":{"total_tokens":3501,"prompt_tokens":997,"completion_tokens":2504,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":613,"completion_tokens_details":{"reasoning_tokens":2420}},"tokens_in":613,"tokens_out":2504,"duration_ms":16770,"temperature":1.0,"reasoning_tokens":2420,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T15:11:23.053055+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the linear ($\\mathcal{F}=0$) backwards evolution on Schwarzschild–AdS with smooth horizon data decaying like $e^{-B t_*}$ for a rate $B$ just above the constant $C_\\kappa$ appearing in the redshift-vector energy estimate; if the $k$-th order hypersurface energy fails to decay exponentially, the central estimate (4.2) is false. An analytic check of whether the characteristic-data local well-posedness asserted in Section 5.1 really follows from the cited spacelike-data result would also settle the construction.","supporting_citations":[{"cited_title":"Dafermos, G","cited_arxiv_id":null,"evidence_quote":"It supplies the backwards scattering template from the asymptotically flat setting that the paper adapts to Kerr–AdS."},{"cited_title":"Holzegel, Well-posedness for the massive wave equation on asymptotically anti-de Sitter space- times, Journal of Hyperbolic Differential Equations 9 (2012), no","cited_arxiv_id":null,"evidence_quote":"It provides the linear well-posedness result whose characteristic-data adaptation is assumed for the finite-in-time problems (5.1)."},{"cited_title":"Holzegel and J","cited_arxiv_id":null,"evidence_quote":"It establishes the sharp logarithmic decay of general forward linear waves under the Hawking–Reall bound, setting the contrast for the exceptional class."},{"cited_title":"5, 1057–1090","cited_arxiv_id":null,"evidence_quote":"It gives the quasimode lower bound showing that general linear solutions decay at best inverse logarithmically."},{"cited_title":"Dold, Unstable mode solutions to the Klein-Gordon equation in Kerr-anti-de Sitter spacetimes , Communications in Mathematical Physics 350 (2017), 639–697","cited_arxiv_id":null,"evidence_quote":"It constructs exponentially growing mode solutions outside the Hawking–Reall bound, the regime in which the new construction still applies."},{"cited_title":"Breitenlohner and D","cited_arxiv_id":null,"evidence_quote":"It fixes the Breitenlohner–Freedman bound and positive-energy framework underlying the coercivity and Hardy-type estimates."},{"cited_title":"1, 169–197","cited_arxiv_id":null,"evidence_quote":"It provides the energy-momentum tensor identities, the T-energy coercivity argument, and the modified Hardy inequality that Section 3.1 adapts."},{"cited_title":"Hebey, Nonlinear Analysis on Manifolds: Sobolev Spaces and Inequalities, Courant Lecture Notes in Mathematics, vol","cited_arxiv_id":null,"evidence_quote":"It supplies the L∞ Sobolev embedding theorem on the spheres used to derive the pointwise bounds on normalised derivatives."}],"review_version":1}