{"id":"82c8029a-e9a0-4ea6-90f9-e447e3f1e2f1","arxiv_id":"2506.07269","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Positive solutions of -Δu=f(u) with f(s)=s^{2*-1}L(s), L slowly varying and satisfying a derivative-decay condition, have a uniform L∞ bound independent of the solution.","lead":"This mathematics paper proves that for a broad family of nearly critical nonlinearities in semilinear elliptic equations, every positive solution has a size bound that does not depend on the solution. The result covers slowly varying functions such as powers of logarithms and oscillatory logarithmic factors, extending earlier bounds that worked mostly for subcritical exponents or a single model nonlinearity.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: Theorem 1.1 is internally coherent; condition (1.5) is explicitly load-bearing but assumed, not a hidden flaw.","rationale":"The reader's weakest assumption is (1.5), and I agree that it carries the proof, but the theorem explicitly assumes it, and the appendix shows the excluded cases. The proof's main steps all check out: Lemma 4.1 is a clean Karamata/Pohozaev argument; the contradiction argument's exponents are consistent for any fixed q in the open interval (N/2,N); the lower radius estimate and Morrey embedding are standard. The minor unflagged items (q range, Theorem 3.1 sketch) do not undermine the central claim. The one concrete erratum I noticed is in Example A.1(6): exp(α log^β log s) is slowly varying only when the log-difference vanishes, which requires β<2 (the proof restricts to β<1); the claim for all β>0 is false. Since Corollary 1.2 and Appendix B impose 0<β<1 for L6, this is peripheral. Therefore the reader's CONDITIONAL verdict should remain as is.","tokens_in":17767,"tokens_out":27347,"duration_ms":293581,"concrete_test":"Independently rederive (4.5) and the final exponent chain for an admissible oscillatory example such as L12(s)=exp(α[γcos(log log(K+s))+log log(K+s)]), verifying that |L'|∈RV_{-1}, that (1.5) holds, and that each power in the chain from R_n^{2-N/q}≥C M h(M)^{-1/q} to M|L'(M)|/L^{N/2}(M)≤C is consistent.","verdict_should_be":"UNCHANGED","load_bearing_attack":"I read the proof of Theorem 1.1 in good faith and did not find a load-bearing gap. The asymptotic (4.5) follows from g(t)=t^{2*} |L'(t)| in RV_{2*-1} via Karamata's theorem; Pohozaev's identity together with the uniform boundary C^1 estimate (4.2) gives the weighted L^1 bound (4.1); and the contradiction in Section 4 is algebraically sound for any fixed q in (N/2,N). The only truly sensitive point is hypothesis (1.5), precisely as the reader says: the final display M|L'(M)|/L^{N/2}(M)≤C contradicts unboundedness only when the ratio diverges. But this is an explicit assumption, and the paper correctly notes that iterated-log factors fail it, so the statement is not overbroad. Peripheral issues exist: Theorem 3.1 is sketched via references, the admissible range for the regularity exponent q is not stated, and Example A.1(6) claims L6∈RV0 for all β>0 although the proof and Appendix B only support 0<β<1. None of these affects the central theorem.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves uniform L∞ a priori bounds for positive weak solutions of −Δu = f(u) in a bounded C^{2,α} domain Ω ⊂ R^N, N > 2, where f(s) = s^q L(s) with L slowly varying at infinity. The main result, Theorem 1.1, treats the critical case q = 2∗−1 under hypotheses (f1)∞–(f3)∞ plus the quantitative condition (1.5). The proof combines a uniform boundary C^1 estimate via moving planes, Pohozaev's identity, Karamata theory for regularly varying functions, elliptic regularity, Sobolev embedding, and Morrey's theorem to derive a weighted L^1 estimate (Lemma 4.1) and then a lower bound on the radius of a ball where a solution is at least half its maximum, yielding a contradiction with (1.5). Theorem 1.3 gives an alternative proof for q < 2∗−1. Appendices A and B list fourteen slowly varying factors and verify which satisfy the hypotheses of the theorems.","tokens_in":17991,"tokens_out":25949,"duration_ms":273596,"significance":"If correct, Theorem 1.1 supplies uniform a priori bounds for a broad class of slightly subcritical nonlinearities at the critical exponent, including logarithmic and oscillatory slowly varying factors, thereby extending earlier results in [6] and addressing a question raised in [22]. The proof is internally coherent and uses standard tools, and condition (1.5) is explicit and falsifiable; the authors also honestly identify that iterated-logarithm factors fail this condition, so the statement is not overbroad. The detailed verification of examples in the appendices is a useful contribution, even though some of the parameter ranges there need clarification.","major_comments":[],"minor_comments":[{"comment":"The regularity exponent q is introduced only with q > N/2, but the Sobolev conjugate 1/q∗ = 1/q − 1/N and the conclusion q∗ > N require q < N. Please state explicitly at the start of the argument that q is fixed in (N/2, N).","section":"Section 4 (proof of Theorem 1.1)"},{"comment":"Theorem 3.1 is a central ingredient in Lemma 4.1, but its proof is only a sketch referring to [6] and [8]; please either provide a complete proof or state the precise results and conditions from those references, including the additional monotonicity assumption used for nonconvex domains.","section":"Section 3 (Theorem 3.1)"},{"comment":"Example A.1(6) asserts L6 ∈ RV0 for all β > 0, while the verification in Appendix B and the discussion of (f3) restrict to 0 < β < 1; please clarify the valid range or correct the inconsistency.","section":"Appendix A, Example A.1(6)"},{"comment":"The proof bounds the integral over {u > s0}, but (4.1) is stated for the full integral over Ω; please justify the small-value contribution or restrict the statement to the truncated integral that is actually used.","section":"Lemma 4.1, Step 4"},{"comment":"There are a few typographical issues in the displayed estimates in the proof of Theorem 1.1 (for example, the superscripts in the line involving R^{2−N/q}_n are split across lines); please proofread the equations.","section":"Throughout"}],"recommendation":"minor_revision","confidential_remarks":"For the editor: the manuscript is within the scope of the journal, and the main theorem appears sound after local clarifications. The references to the authors' own earlier work [6] and [18] are used for technical steps but not in a circular way. I recommend minor revision, contingent on the clarifications listed in the report."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here is my read. Theorem 1.1 is real progress on the Gidas-Spruck a priori bound problem at the critical exponent. Previous work covered a single logarithmic model; this one covers every slowly varying L with L' < 0 eventually, |L'| in RV_{-1}, and the ratio condition (1.5). The proof is a clean, coherent assembly of standard tools: moving planes for the boundary estimate, Pohozaev for the weighted L1 bound, Karamata theory to identify the asymptotic of 2*F(s)-s f(s), and Morrey embedding to bound the radius of the half-max ball. I checked the key asymptotic (4.5) and the final contradiction; both are sound. The condition (1.5) is explicitly load-bearing, and the paper is honest about its limits, noting that iterated-log factors fail it and that oscillatory L with infinite amplitude are out of scope. The soft spots are minor. In the proof of Theorem 1.1, the exponent q must be chosen in (N/2, N) for the argument to work: that range makes 2*(q-1)-q positive and gives q* > N for Morrey. The paper never states this range, though it is implicit in the inequalities. Theorem 3.1 is only sketched, with the details deferred to references; this is standard for this area, but it is a point of reliance. Example A.1(6) claims L6 is slowly varying for all beta > 0, while the verification and Appendix B only support 0 < beta < 1; a small overclaim in an appendix, not central to the main theorem. Theorem 1.3 is explicitly an alternative proof of a known result, so the novelty and value sit on Theorem 1.1 and the worked examples in Appendices A and B. I also want to push back on any reading that (1.5) is a hidden flaw. It is a stated assumption, and the paper itself explains which natural examples fail it. The citation pattern is clean: [6] is the result being generalized, [18] is the author's own Caratheodory result used as a black box, and the dependence is transparent. No circularity, no fitting. For whom: this is for people working on a priori bounds and related Liouville theorems for superlinear elliptic problems. It deserves a serious referee; the main theorem is a genuine, checkable advance. I would send it to peer review and would cite it in my own work in this area.","headline":"Theorem 1.1 genuinely extends the critical-exponent a priori bound from one log model to a broad class of slowly varying nonlinearities, and the proof is honest and checkable.","tokens_in":683,"tokens_out":1577,"would_cite":true,"duration_ms":28596,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35B45","35B09","35B33"],"pacs":[],"model":"deepseek-v4-flash","headline":"Uniform $L^\\infty$ a priori bounds hold for all positive weak solutions of the slightly subcritical problem (1.1) under condition (1.5) on the slowly varying factor.","keywords":["a priori bounds","positive solutions","slightly subcritical nonlinearity","regularly varying functions","Pohozaev identity","moving planes method","critical Sobolev exponent"],"falsifier":"The theorem would be refuted by exhibiting, on some bounded domain, an unbounded sequence of positive solutions to (1.1) with $f(s)=s^{2^*-1}L(s)$ where $L$ satisfies (f1)–(f3) and (1.5). A concrete place to look is the unit ball for one of the listed oscillatory factors, for instance $L_{12}(s)=\\exp\\{\\alpha(\\gamma\\cos(\\log\\log(K+s))+\\log\\log(K+s))\\}$ with $\\alpha<0$ and $|\\gamma|<1$; finding radial positive solutions with $\\|u_n\\|_\\infty\\to\\infty$ would invalidate Theorem 1.1.","tokens_in":17537,"feed_emoji":"📐","tokens_out":10213,"duration_ms":99943,"temperature":0.7,"pith_summary":"This paper proves that positive solutions of the semilinear elliptic problem $-\\Delta u=f(u)$ in a bounded domain $\\Omega\\subset\\mathbb{R}^N$, with $u=0$ on $\\partial\\Omega$, admit a uniform $L^\\infty$ bound when $f(s)=s^{2^*-1}L(s)$ is slightly subcritical and $L$ is a slowly varying factor satisfying a quantitative decay condition: $s|L'(s)|/L^{N/2}(s)\\to+\\infty$ as $s\\to\\infty$. If the proof is right, this supplies a priori estimates for nonlinearities with logarithmic and oscillatory corrections, extending earlier results that stopped short of the critical exponent. The paper also proves the same kind of bound for the entire subcritical range $q\\in[1,2^*-1)$ under milder hypotheses. The route goes through a weighted $L^1$ estimate obtained from the Pohozaev identity, followed by elliptic regularity and a lower bound on the radius of a ball where a solution exceeds half of its maximum.","feed_headline":"One decay condition rules out blow-up in critical elliptic equations","feed_subtitle":"Slowly varying factors satisfying s|L'(s)|/L^{N/2}(s)→∞ yield uniform bounds for every positive solution.","key_machinery":"The load-bearing object is the slowly varying factor $L$, meaning $L(\\tau s)/L(s)\\to 1$ as $s\\to\\infty$ for every $\\tau>0$. The workhorse identity is the asymptotic equivalence $2^*F(s)-sf(s)\\sim \\frac{1}{2^*}s^{2^*+1}|L'(s)|$ as $s\\to\\infty$, where $F(s)=\\int_0^s f(t)\\,dt$; it follows from the uniform convergence theorem for regularly varying functions applied to $g(s)=s^{2^*}|L'(s)|$, which lies in $RV_{2^*-1}$ because $|L'|\\in RV_{-1}$. This identity turns the Pohozaev identity, combined with uniform boundary estimates, into the weighted estimate $\\int_\\Omega u^{2^*+1}|L'(u)|\\,dx\\le C$. A second mechanism, Morrey's theorem, gives a lower bound on the radius of a ball where a solution exceeds half its maximum; inserting that ball into the weighted estimate produces the ratio bound that contradicts (1.5).","core_discovery":"The central claim is Theorem 1.1: let $q=2^*-1$ and assume $f(s)=s^{2^*-1}L(s)$ with $L\\in RV_0$ slowly varying, eventually decreasing, $|L'|\\in RV_{-1}$, and $\\lim_{s\\to\\infty} s|L'(s)|/L^{N/2}(s)=+\\infty$. Then there is a constant $C$ depending only on $f$, $\\Omega$, and $N$ such that every positive weak solution $u$ of (1.1) satisfies $\\|u\\|_\\infty\\le C$. The proof works by contradiction: an unbounded solution sequence would first yield, through the Pohozaev identity and boundary estimates, the uniform weighted estimate $\\int_\\Omega u^{2^*+1}|L'(u)|\\,dx\\le C$; a radius estimate from Morrey's theorem would then squeeze from this the ratio bound $\\|u_n\\|_\\infty |L'(\\|u_n\\|_\\infty)|/L^{N/2}(\\|u_n\\|_\\infty)\\le C$, contradicting (1.5).","pith_inferences":["The ratio in (1.5) is plausibly a genuine threshold rather than a technical convenience: slowly varying factors that decay too weakly, such as iterated logarithms, may admit unbounded solution sequences, and the paper's exclusion of them could reflect an actual boundary of the theorem.","The same two-step architecture—a Pohozaev-weighted estimate followed by a Morrey radius estimate—could transfer to other critical problems with regularly varying nonlinearities, such as systems or quasilinear equations, once an analogous weighted integral can be derived.","A direct numerical or perturbative check in a ball, using one of the oscillatory factors inside the class, could map how large the solutions need to be before the uniform bound sets in; for the excluded iterated-log case, searching for radial unbounded solutions would test whether the condition is necessary."],"forward_implications":["For every $f(s)=s^{2^*-1}L(s)$ satisfying (1.5), all positive weak solutions of (1.1) are bounded in $L^\\infty$ by a constant independent of the solution, so no interior blow-up can occur.","Corollary 1.2 yields uniform a priori bounds for the concrete family $L_i$ of slowly varying factors listed in the paper (indices 1, 2, 4–7, 11–14, with the stated parameter restrictions), including multiparameter oscillatory factors.","For the strictly subcritical range $q\\in[1,2^*-1)$, Theorem 1.3 gives the same uniform bound under slow variation alone, without the derivative condition (1.5).","The iterated-logarithm factor $L_3(s)=(\\log_m(K+s))^\\alpha$ with $m\\ge2$ fails condition (1.5), so the theorem marks a boundary of the method: such nonlinearities would require a stronger condition or a different argument."],"supporting_citations":[{"why":"Supplies the moving-planes boundary uniform estimates and the Pohozaev identity that anchor Steps 1 and 2 of the proof.","marker":"[8]"},{"why":"Provides the technique of estimating from below the radius of a ball where a solution exceeds half of its maximum.","marker":"[6]"},{"why":"Provides the global Hölder estimates used to upgrade boundary information to uniform $C^{1,\\nu}$ control.","marker":"[15]"},{"why":"Supplies the Schauder and $W^{2,p}$ estimates used to control solutions in a neighborhood of the boundary.","marker":"[12]"},{"why":"Provides the Sobolev embeddings and Morrey's theorem used in the radius estimate.","marker":"[3]"},{"why":"Foundational uniform convergence theorem for regularly varying functions, used to derive the asymptotic equivalence for $2^*F(s)-sf(s)$.","marker":"[13]"},{"why":"Source for the properties of regularly varying functions quoted as Theorem 3.2 and Proposition 3.4.","marker":"[21]"}],"fun_headline_variants":["Decay condition kills blow-up in critical elliptic problems","Uniform bounds for all positive solutions under one decay condition","Slightly subcritical, no blow-up: L-decay suffices","One decay condition guarantees uniform a priori bounds","L-decay prevents blow-up in elliptic equations"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof stands or falls on condition (1.5), namely that the ratio $s|L'(s)|/L^{N/2}(s)$ diverges to $+\\infty$; if that ratio fails to diverge, the contradiction step that forces a uniform bound never materializes.","fun_headline_variants_meta":{"raw":{"variants":["Decay condition kills blow-up in critical elliptic problems","Uniform bounds for all positive solutions under one decay condition","Slightly subcritical, no blow-up: L-decay suffices","One decay condition guarantees uniform a priori bounds","L-decay prevents blow-up in elliptic equations"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00106,"raw_usage":{"total_tokens":4443,"prompt_tokens":937,"completion_tokens":3506,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":553,"completion_tokens_details":{"reasoning_tokens":3429}},"tokens_in":553,"tokens_out":3506,"duration_ms":33772,"temperature":1.0,"reasoning_tokens":3429,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T05:39:45.847745+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"The theorem would be refuted by exhibiting, on some bounded domain, an unbounded sequence of positive solutions to (1.1) with $f(s)=s^{2^*-1}L(s)$ where $L$ satisfies (f1)–(f3) and (1.5). A concrete place to look is the unit ball for one of the listed oscillatory factors, for instance $L_{12}(s)=\\exp\\{\\alpha(\\gamma\\cos(\\log\\log(K+s))+\\log\\log(K+s))\\}$ with $\\alpha<0$ and $|\\gamma|<1$; finding radial positive solutions with $\\|u_n\\|_\\infty\\to\\infty$ would invalidate Theorem 1.1.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the moving-planes boundary uniform estimates and the Pohozaev identity that anchor Steps 1 and 2 of the proof."},{"cited_title":"Castro and R","cited_arxiv_id":null,"evidence_quote":"Provides the technique of estimating from below the radius of a ball where a solution exceeds half of its maximum."},{"cited_title":"Linear and quasilinear elliptic equa- tions","cited_arxiv_id":null,"evidence_quote":"Provides the global Hölder estimates used to upgrade boundary information to uniform $C^{1,\\nu}$ control."},{"cited_title":"Gilbarg and N","cited_arxiv_id":null,"evidence_quote":"Supplies the Schauder and $W^{2,p}$ estimates used to control solutions in a neighborhood of the boundary."},{"cited_title":"Karamata","cited_arxiv_id":null,"evidence_quote":"Foundational uniform convergence theorem for regularly varying functions, used to derive the asymptotic equivalence for $2^*F(s)-sf(s)$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Source for the properties of regularly varying functions quoted as Theorem 3.2 and Proposition 3.4."}],"review_version":1}