{"id":"0db9f917-88f8-434d-86cf-dfc2edde5a2d","arxiv_id":"2502.00825","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":2.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A survey of regularity estimates for the Laplacian and p-Laplacian on metric measure spaces, with a proof overview of the author's second-order regularity theorem in bounded RCD spaces.","lead":"This paper surveys what is known about Hölder, Lipschitz, and second-order regularity for the Laplacian and the p-Laplacian on metric measure spaces. It is useful as an entry point to RCD-space regularity theory, and it sketches the proof of the author's recent p-Laplacian result.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The p<2 absorption bound after (6.13) fails on a one-dimensional Euclidean example: R=(2−p)(4−p)|Hess|², so λ_p<1 is impossible when p<3−√2; the p∈(1,3) range in Theorem 5.1 is unsupported as presented.","rationale":"The central claim is Theorem 5.1: second-order regularity for the p-Laplacian on bounded RCD(0,∞) spaces, with p∈(1,3). The proof strategy is coherent: regularize, prove ε-uniform second-order estimates, pass to limit. The decisive step is the uniform estimate (6.6), because Step 4 simply transfers it to the limit. In that estimate, the only genuinely nonlinear obstacle is the remainder R in (6.13) for p<2; for p≥2 the argument is standard. The text does not prove the p<2 absorption, but refers to [13, Prop. 4.3]. My explicit computation on the simplest possible RCD space, the interval, shows that the pointwise inequality quoted from that proposition cannot hold as stated: for u(x)=x²/2 and p=3/2, R→5/4 while |Hess|²=1. Thus the survey's proof overview contains a concrete mathematical error or a severe omission. This is not a mere 'proof deferred to another paper' concern: the asserted inequality appears to be false. The theorem may still be true—[13] could prove it by a different argument or over a smaller p-range—but the present text does not establish it for p∈(1,3). The reader's weakest_assumption already identified the absorption of R as the load-bearing point; I agree and sharpen it with an elementary falsifying instance. I do not see an independent obstacle in Step 1 or Step 4; they are routine conditional on (6.6). The use of the improved Bochner inequality (4.9) and Corollary 4.5 as black boxes is acceptable in a survey, though it adds to non-self-containedness. Given the stated goal of the paper is partly to give an overview of the proof of (⋆p), an incorrect or incomplete proof sketch is a substantive problem. The right editorial outcome is to require the author to either correct the p-range (e.g., to p>3−√2, if that matches [13]), state the precise hypotheses of [13, Prop. 4.3], or provide the missing argument. Until then, the paper should not be taken as having proven Theorem 5.1 as stated. I therefore recommend keeping the reader's CONDITIONAL verdict: the concern is real, but it is potentially fixable and does not by itself disprove the underlying result.","tokens_in":28282,"tokens_out":15635,"duration_ms":146630,"concrete_test":"Compute R in (6.13) for the one-dimensional RCD example X=[−1,1], u(x)=x²/2−1/6, p=3/2, ε=10⁻⁶, x=1: the two terms give ≈(2−p)²+2(2−p)=1.25 while |Hess(u)|²=1, falsifying R≤λ_p|Hess|² for λ_p<1. Then check the statement of [13, Prop. 4.3]: if it is exactly as quoted, the proof of Theorem 6.4 is invalid for p<3−√2; if it contains additional assumptions, the survey must state them and show they hold for the solutions uε constructed in §6.3. As a complementary check, rerun the absorption step for p=1.4 on the same example; the coefficient (2−p)(4−p)≈1.56 gives an even larger violation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 6.2's proof of Theorem 6.4 hinges on absorbing the remainder R in (6.13) into the left-hand side. For p≥2 the second term of R is non-positive and the first is absorbed because |p−2|<1. For p<2 the text invokes [13, Prop. 4.3]: R≤λ_p|Hess(u)|² with λ_p∈(0,1). On the bounded RCD(0,∞) interval X=[−1,1] with Lebesgue measure, take u(x)=x²/2 (recentered to zero mean), so |Du|=|x|, ∇|Du|=1, Δ∞u=x², |Hess|=1. Since M=+∞ for p<2, the parentheses in (6.13) become (p−2)²(Δ∞u)²/(|Du|²+ε)² + 2(2−p)|∇|Du||²|Du|²/(|Du|²+ε). At x=1, letting ε→0+, R→(2−p)²+2(2−p)=(2−p)(4−p). For p=3/2 this is 5/4>1=|Hess|², so no λ_p<1 exists. The same computation shows the failure whenever (2−p)(4−p)>1, i.e. p<3−√2. Therefore the displayed argument cannot yield the uniform estimate (6.6) for the full interval p∈(1,3) stated in Theorem 5.1. Either the survey misstates [13, Prop. 4.3], that proposition has additional hypotheses omitted here, or the p-range must be narrowed. In any case, as written, the central theorem is not established by Section 6 for p∈(1,3−√2).","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript is a survey of regularity theory for the Laplacian and p-Laplacian on metric measure spaces, with emphasis on RCD spaces. It first reviews the variational formulation, boundary value problems, and Hölder regularity under doubling/Poincaré assumptions. It then recalls second-order and Lipschitz regularity for the Laplacian under Ricci bounds, eigenfunction estimates, and unique continuation. The final two sections present recent results by Benatti and Violo: Theorem 5.1 (|Du|^{p−2}∇u ∈ H^{1,2}_C and |Du|^{p−1} ∈ W^{1,2} when Δ_p u ∈ L², p ∈ (1,3)) and related Lipschitz statements, together with a proof overview via ε-regularization. The survey is clearly written and the statements are consistent with the cited literature.","tokens_in":28704,"tokens_out":19710,"duration_ms":180122,"significance":"If correct, Theorem 5.1 is a substantial contribution: it extends to nonsmooth RCD spaces the Euclidean second-order regularity f ∈ L² ⇒ |∇u|^{p−2}∇u ∈ W^{1,2} for the p-Laplacian, and the range p ∈ (1,3) is plausible from the structure of the proof. The paper also provides a useful state-of-the-art account and lists open questions. The proof overview is not fully self-contained—several key claims (Prop. 6.3, Lemma 6.6, the p<2 absorption after (6.13), and the p<2 case of Theorem 6.9) are imported from the authors' preprint [13]—but this is typical for a survey. I checked the specific stress-test objection: the one-dimensional computation with R = (2−p)(4−p)|Hess|² does not apply to Eq. (6.13), because for p<2 the factor (1−2χ_{|Du|≤M}) is −1 when M = +∞, making the first contribution negative. Thus I find no mathematical contradiction in the manuscript, though the p<2 bound needs to be stated explicitly for clarity.","major_comments":[],"minor_comments":[{"comment":"The stress-test counterexample with R = (2−p)(4−p)|Hess|² is not applicable because the first term in R contains the factor (1−2χ_{|Du|≤M}), which is −1 for p<2 (M=+∞), so the first contribution is negative. Nevertheless, the bound R ≤ λ_p|Hess(u)|² for p<2 is only asserted with a reference to [13, Prop. 4.3]; since the full range p∈(1,3) of Theorem 5.1 relies on this absorption, the survey should state explicitly what that proposition gives (or at least its hypotheses) so that the reader can verify the argument.","section":"Section 6.2, after Eq. (6.13)"},{"comment":"The sentence 'Therefore by lower semicontinuity (see e.g. )' contains an empty citation; please provide a precise reference (e.g., a proposition in [13] or in [52]).","section":"Section 6.2, below Eq. (6.6)"},{"comment":"The notation T_{f,w}(w) and the subsequent expression T_{f,w}(U_1−U_2) appear to be a typo; the contraction argument should involve T_{f,w}(U_1)−T_{f,w}(U_2) (or the definition of the map should be clarified), otherwise the displayed estimate is confusing.","section":"Section 6.3, proof of Proposition 6.8"},{"comment":"The right-hand side of inequality (D) is missing the integration domain; it should read ∫_Ω |Du|^p dm (or the domain should be understood from context, but it is better to write it explicitly).","section":"Section 3.2, Proposition 3.5"},{"comment":"Please fix the typos: 'brtiefly' (Section 1), 'Corolalry' (Section 3.3), 'Haj/suppress lasz' in reference [58], and 'a version of the classical Weyl’s lemma' → 'a version of the classical Weyl lemma' (Section 4.2).","section":"Throughout"}],"recommendation":"minor_revision","confidential_remarks":"The manuscript is in large part a report on the author's own joint work [13], including the entire proof overview of Theorem 5.1. This is not circular, but the survey should make the provenance of the p<2 technical steps clearer, perhaps by adding a remark that the detailed proof of the p<2 absorption is contained in [13, Prop. 4.3] and is not reproduced in full. The paper is suitable for a survey/overview venue in math.AP."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The bottom line: this is a well-organized survey of Laplacian and p-Laplacian regularity in metric measure spaces, with the centerpiece being a proof sketch of a second-order regularity theorem for the p-Laplacian in bounded RCD(0,∞) spaces. That theorem is not new here—it comes from the author's companion paper with Benatti [13]—but the survey frames it usefully against the first-order theory, the Laplacian case, and open questions.\n\nWhat it does well: the paper gives a clean map of the area. Sections 3 and 4 summarize existence, boundary regularity, Harnack and Hölder estimates, Lipschitz regularity, eigenfunction bounds, and unique continuation, with the right pointers to the literature. The proof sketch in Section 6 is a genuine attempt to convey how the ε-regularization argument works, including the fixed-point construction in Section 6.3. The survey statements look consistent with the cited results, and the heavy citation to [13] is legitimate because the main theorem is from there.\n\nSoft spots: the paper is not self-contained, which is normal for a survey but more pronounced here because the proof of Theorem 5.1 repeatedly defers to [13], especially the p<2 absorption step and the approximation in Theorem 6.9. There is also a blank citation after \"lower semicontinuity (see e.g. )\" in Section 6.2. These are minor, not load-bearing. The \"sufficiently close to two\" phrasing in the introduction undersells the actual p∈(1,3) range; that is a harmless imprecision.\n\nI checked the stress-test objection about the absorption inequality after (6.13). It doesn't hold up. For p<2 the proof sets M=∞, so the indicator inside the first remainder term gives 1−2χ=−1, not +1. On the 1D example the remainder is then bounded by −(p−2)²+2(2−p)=p(2−p), which is strictly less than 1 for p∈(1,2). So the claimed λ_p∈(0,1) is consistent with that example. The concern appears to be a sign error, not a real gap in the survey's argument.\n\nWho this is for: anyone wanting an entry point into p-Laplacian regularity on RCD spaces, or a preview of the proof of the main theorem before reading [13]. It deserves a serious referee; a published survey would be useful. I'd recommend sending it to review, with the blank reference and the p-range wording fixed.","headline":"A useful survey of p-Laplacian regularity in RCD spaces; the central theorem is imported from the author's companion paper, and the stress-test objection about the absorption step misses a sign.","tokens_in":29235,"tokens_out":6025,"would_cite":false,"duration_ms":57420,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35B65","35J92","46E36","58J05"],"pacs":[],"model":"deepseek-v4-flash","headline":"On bounded RCD(0,∞) spaces, square-integrable p-Laplacian data forces the nonlinear flux |Du|^{p−2}∇u into the Sobolev class H^{1,2}_C(TX), with an explicit L2 estimate.","keywords":["nonlinear potential theory","metric spaces","Laplacian","p-Laplacian","elliptic PDEs","regularity estimates","Ricci curvature","RCD spaces"],"falsifier":"Perform the calculation of Section 6.2 on a concrete bounded RCD(0,∞) space, for instance a Euclidean cone or a product manifold, and check whether the displayed remainder $R$ after (6.13) obeys $R\\le \\lambda_p|\\mathrm{Hess}(u)|^2$ at almost every point with $\\lambda_p<1$ for all $p\\in(1,2)$. Finding a value of $p$ where this pointwise absorption fails would break the uniform estimate (6.6) and with it the claimed $H^{1,2}_C$ regularity.","tokens_in":27994,"feed_emoji":"📐","tokens_out":9178,"duration_ms":82062,"temperature":0.7,"pith_summary":"This survey argues that second-order elliptic regularity survives in metric measure spaces with a synthetic lower Ricci bound. Its central claim is a recent theorem: on a bounded RCD(0,∞) space, if a function $u$ lies in the domain of the $p$-Laplacian and $\\Delta_p u$ is square-integrable for $p\\in(1,3)$, then the nonlinear flux $|Du|^{p-2}\\nabla u$ has an $L^2$ covariant derivative and $|Du|^{p-1}$ belongs to $W^{1,2}(X)$. The estimate is quantitative, with a constant depending only on $p$. This is the $p$-analogue of the known linear fact that $\\Delta u\\in L^2$ forces $u\\in W^{2,2}$, and it matters because a general difference-quotient method is unavailable in nonsmooth settings. The paper also surveys the surrounding landscape: H\\\"older and boundary regularity under doubling plus Poincar\\'e inequalities, Lipschitz estimates under Ricci bounds, and a detailed outline of the proof via regularization and fixed-point arguments.","feed_headline":"L2 data for p-Laplacians buys Sobolev regularity in rough spaces","feed_subtitle":"A square-integrable right-hand side upgrades the p-Laplacian's flux to a Sobolev vector field.","key_machinery":"The proof is carried by four devices: the $p$-Bochner identity (5.1), which expresses divergence of $|\\nabla u|^{p-2}A(\\nabla|\\nabla u|^p)$ as a Hessian-square term plus curvature and Laplacian terms; the regularized $(\\varepsilon,p)$-Laplacian $\\Delta_{p,\\varepsilon}u := \\mathrm{div}((|Du|^2+\\varepsilon)^{(p-2)/2}\\nabla u)$, which is uniformly elliptic; the developed operator $D_{\\varepsilon,p}u := \\Delta u + (p-2)\\frac{\\mathrm{Hess}(u)(\\nabla u,\\nabla u)}{|\\nabla u|^2+\\varepsilon}$, which makes the freezing step algebraic; and the improved Bochner inequality (4.9) on RCD spaces, together with an absorption device that controls a remainder by $\\lambda_p|\\mathrm{Hess}(u)|^2$ with $\\lambda_p<1$. The overall mechanism is: regularize, prove uniform a priori bounds on the flux, solve the regularized equation by a fixed point using the compact inclusion $D_0(\\Delta)\\subset W^{1,2}$, then pass $\\varepsilon\\to 0^+$.","core_discovery":"The paper's central asserted content is Theorem 5.1: on a bounded RCD(0,∞) space, for $p\\in(1,3)$, if $u\\in D(\\Delta_p)$ and $\\Delta_p u\\in L^2(m)$, then $|Du|^{p-2}\\nabla u\\in H^{1,2}_C(TX)$ and $|Du|^{p-1}\\in W^{1,2}(X)$, with $\\int |\\nabla(|Du|^{p-2}\\nabla u)|^2\\,dm \\le C_p(\\|\\Delta_p u\\|_{L^2}^2 + \\||Du|^{p-1}\\|_{L^1})$. The space $H^{1,2}_C(TX)$ is the space of $L^2$ vector fields whose covariant derivative is again $L^2$. This is the natural nonlinear analogue of the linear inclusion $u\\in W^{1,2}$, $\\Delta u\\in L^2 \\Rightarrow u\\in W^{2,2}$; the flux vector field, not $u$ itself, gains the second derivative, and the paper notes that $u\\in W^{2,2}$ would be false even in the Euclidean setting.","pith_inferences":["If the absorption inequality used for $p<2$ can be proved directly from the Bochner inequality rather than by a deferred argument, the method would likely give local, not only global, $H^{1,2}_C$ regularity for the $p$-Laplacian; this is a testable extension the survey leaves open.","The frozen-coefficient fixed-point scheme used to solve the regularized equation looks transplantable to other divergence-form operators whose coefficients depend on $|\\nabla u|$, such as weighted or anisotropic $p$-Laplacians; on RCD spaces the same two-step contraction argument might yield second-order estimates for those operators.","Because Theorem 5.1 controls $|\\nabla(|Du|^{p-2}\\nabla u)|$ in $L^2$, it may feed into higher integrability and compactness arguments for $p$-harmonic approximations, and ultimately into boundary regularity or nodal-domain estimates for $p$-eigenfunctions in singular spaces, though the paper does not pursue these."],"forward_implications":["For $p\\in(1,3)$, the $p$-Laplacian behaves like its linear brother: an $L^2$ right-hand side upgrades the natural stress vector field $|Du|^{p-2}\\nabla u$ to a vector field with $L^2$ covariant derivative, and the estimate is quantitative in $p$.","$u$ itself does not gain two derivatives; the paper notes this would be false even in Euclidean space, so the theorem identifies the right object to regularize.","On bounded RCD(0,N) spaces with finite $N$, a datum in $L^q$ for $q>N$ forces solutions of the $p$-Poisson equation to be Lipschitz (Theorem 5.3).","$p$-electrostatic potentials solving $\\Delta_p u=0$ away from a compact set are locally Lipschitz outside the obstacle (Theorem 5.4).","For $K<0$ or finite dimension $N$, the $p$-range can be widened, so the regularity phenomenon is not tied to $p=2$ nor to nonnegative curvature (Remark 5.5)."],"supporting_citations":[{"why":"Supplies the companion proof of Theorem 5.1, Theorem 5.3 and Theorem 5.4, along with the deferred absorption inequality for p<2.","marker":"[13]"},{"why":"Provides the improved Bochner inequality (4.9) and the inclusion D(Δ)⊂W^{2,2} that the p-Laplacian argument uses as its linear backbone.","marker":"[44]"},{"why":"Defines the RCD(K,∞) condition through the weak Bochner inequality, fixing the class of spaces in which the main theorem is stated.","marker":"[4]"},{"why":"Establishes the self-improvement of the Bochner inequality and the Laplacian second-order regularity result that motivates the nonlinear version.","marker":"[89]"},{"why":"Gives the Euclidean model: Δp u∈L²_loc if and only if |∇u|^{p−2}∇u∈W^{1,2}_loc, which is the benchmark for Theorem 5.1.","marker":"[27]"},{"why":"Earlier Sobolev regularity for |∇u|^{p−1} under higher-integrability data, the result that Theorem 5.1 sharpens in the RCD setting.","marker":"[72]"}],"fun_headline_variants":["p-Laplacian flux gains Sobolev regularity from L^2 data","Sobolev-regular flux for p-Laplacian on RCD spaces","L^2 p-Laplacian data yields Sobolev flux in rough spaces","RCD spaces: L^2 p-Laplacian forces Sobolev flux","Second-order regularity of p-Laplacian flux"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the nonsmooth space satisfies the improved Bochner inequality and that, for every $p\\in(1,3)$, the remainder term in the $p$-Bochner computation can be pointwise absorbed into a fraction of $|\\mathrm{Hess}(u)|^2$; for $p<2$ that absorption is asserted with proof deferred to the companion paper, and if it fails the uniform bounds and the limit argument collapse.","fun_headline_variants_meta":{"raw":{"variants":["p-Laplacian flux gains Sobolev regularity from L^2 data","Sobolev-regular flux for p-Laplacian on RCD spaces","L^2 p-Laplacian data yields Sobolev flux in rough spaces","RCD spaces: L^2 p-Laplacian forces Sobolev flux","Second-order regularity of p-Laplacian flux"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000707,"raw_usage":{"total_tokens":3132,"prompt_tokens":840,"completion_tokens":2292,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":456,"completion_tokens_details":{"reasoning_tokens":2194}},"tokens_in":456,"tokens_out":2292,"duration_ms":14901,"temperature":1.0,"reasoning_tokens":2194,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T17:35:00.924077+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Perform the calculation of Section 6.2 on a concrete bounded RCD(0,∞) space, for instance a Euclidean cone or a product manifold, and check whether the displayed remainder $R$ after (6.13) obeys $R\\le \\lambda_p|\\mathrm{Hess}(u)|^2$ at almost every point with $\\lambda_p<1$ for all $p\\in(1,2)$. Finding a value of $p$ where this pointwise absorption fails would break the uniform estimate (6.6) and with it the claimed $H^{1,2}_C$ regularity.","supporting_citations":[{"cited_title":"Second-order estimates for the $p$-Laplacian in RCD spaces","cited_arxiv_id":"2401.09982","evidence_quote":"Supplies the companion proof of Theorem 5.1, Theorem 5.3 and Theorem 5.4, along with the deferred absorption inequality for p<2."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the improved Bochner inequality (4.9) and the inclusion D(Δ)⊂W^{2,2} that the p-Laplacian argument uses as its linear backbone."},{"cited_title":"Savar´e, Self-improvement of the Bakry- ´Emery condition and Wasserstein contraction of the heat flow in RCD(K,∞ ) metric measure spaces , Discrete Contin","cited_arxiv_id":null,"evidence_quote":"Establishes the self-improvement of the Bochner inequality and the Laplacian second-order regularity result that motivates the nonlinear version."},{"cited_title":"Cianchi and V","cited_arxiv_id":null,"evidence_quote":"Gives the Euclidean model: Δp u∈L²_loc if and only if |∇u|^{p−2}∇u∈W^{1,2}_loc, which is the benchmark for Theorem 5.1."},{"cited_title":"Lou , On singular sets of local solutions to p-Laplace equations , Chinese Ann","cited_arxiv_id":null,"evidence_quote":"Earlier Sobolev regularity for |∇u|^{p−1} under higher-integrability data, the result that Theorem 5.1 sharpens in the RCD setting."}],"review_version":1}