{"id":"57aea50f-d3f2-4259-bd2f-344d01efb20f","arxiv_id":"2411.18238","paper_version":2,"verdict":"UNVERDICTED","confidence":"MODERATE","novelty_score":1.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"An expository monograph on the fractional Laplacian, presenting equivalent definitions, explicit calculations, Liouville results, and a regularity theory for global and interior solutions.","lead":"This book-length draft introduces the fractional Laplacian through many equivalent definitions, worked examples, Liouville theorems, and a broad regularity theory in Lebesgue, Sobolev, and Besov spaces. It is an expository reference rather than a research announcement, so its value is pedagogical organization rather than new mathematics.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 4.49 states a Littlewood–Paley/Bessel equivalence for p≥1, but its proof relies on Mikhlin multipliers valid only for 1<p<∞; the p=1 endpoint is false because the Littlewood–Paley square function is not L^1-bounded.","rationale":"I read the book as an expository monograph whose central promise is that the multiple definitions of the fractional Laplacian are equivalent and that the subsequent regularity theory is reliable. The equivalence part in Chapter 1 is carefully proved and I found no fatal gap there. The self-containedness concern raised by the reader is real but is a scope/prerequisite issue rather than a correctness defect; unproved Mikhlin and interpolation results are standard and are cited, so they do not by themselves undermine the mathematics. The more serious defect is the p≥1 range in Theorem 4.49. The proof explicitly invokes the Mikhlin multiplier theorem, Theorem 4.48, which the authors themselves state only for 1<p<∞, and the claimed endpoint p=1 contradicts a classical fact about the Littlewood–Paley square function. This is not merely a missing proof: as stated, the theorem is false for p=1. Because Theorem 4.49 is the stated link between Bessel and Besov spaces, the book should be revised to restrict (4.109) to p>1 and to re-check any downstream statement that relies on the p=1 case. This is a concrete, localized correction rather than a wholesale rejection, so I recommend a conditional verdict.","tokens_in":107536,"tokens_out":19290,"duration_ms":174833,"concrete_test":"Test the endpoint p=1 of (4.109). Take n=1, 0<s<1, and u=B(s/2)*f with f an explicit L^1 function whose Littlewood–Paley square function is known not to be integrable, e.g. f=Σ_{j≥1} 2^{-j} χ_{[2^{-j},2^{-j+1}]}. Compute S u(x)=(Σ_{j≥0} 2^{2js}|φ_j*u(x)|^2)^{1/2} and both sides of (4.109). If ∫ S u dx=∞ while ∥u∥_{L^1_s}=∥f∥_{L^1}<∞, the claimed equivalence for p=1 is false and the theorem must be restricted to p>1 or supplied with a genuinely new endpoint argument.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The load-bearing flaw is in Section 4.6, Theorem 4.49, which asserts the two-sided estimate (4.109) for all p≥1: the Bessel potential norm L^p_s(R^n) is equivalent to the L^p norm of the dyadic square function. The proof reduces this to the vector-valued Mikhlin multiplier theorem, Theorem 4.48, which is stated and valid only for p∈(1,∞). At p=1 the proof has no justification, and the endpoint claim is in fact false. On the shell |ξ|≈2^j the multiplier 2^{js}⟨ξ⟩^{-s}φ_j(ξ) is comparable to the Littlewood–Paley projection φ_j(ξ); composing with B(s/2) therefore yields a square function essentially equal to the Littlewood–Paley square function of any f∈L^1. It is classical (e.g., [Ste70, Chapter IV]) that this square function is not bounded from L^1 to L^1, only from L^1 to weak-L^1. Hence the right-hand inequality in (4.109) fails for p=1: there exist f∈L^1 for which the square function of u=B(s/2)*f is not integrable while ∥u∥_{L^1_s}=∥f∥_{L^1}<∞. Since Theorem 4.49 is the principal bridge between Bessel potential spaces and Besov spaces, any later statement or application that invokes it with p=1 inherits a genuine mathematical gap, not merely an unproved background theorem.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This book manuscript (arXiv:2411.18238) is an expository treatment of the fractional Laplacian and its surrounding regularity theory. Chapter 1 collects several equivalent definitions of (-Δ)^s on a common core of functions and proves many of the equivalences; Chapter 2 gives explicit fractional Laplacians for power functions, half-space and ellipsoid-supported functions, and radial powers; Chapter 3 proves Liouville-type results via Fourier and distribution methods; Chapters 4–5 develop global and interior regularity theory in Lebesgue spaces using Riesz and Bessel potentials, Sobolev spaces, and Besov spaces; several appendices collect auxiliary tools. The preface advertises the text as self-contained, requiring only fundamental calculus and basic measure theory.","tokens_in":107796,"tokens_out":8222,"duration_ms":78416,"significance":"If corrected, this would be a useful didactic monograph. The systematic comparison of definitions, the explicit computations in Chapter 2, and the clean organization of the potential-space regularity theory are genuine strengths, and many arguments are proved in detail rather than merely quoted. The main obstacle is a false endpoint statement in the central Bessel–Besov comparability theorem, which currently invalidates the theorem in the stated generality. Since the surrounding regularity theory is mostly developed for 1<p<∞, the flaw is likely repairable, but it must be fixed and all downstream uses must be audited.","major_comments":[{"comment":"Theorem 4.49 is stated for all p≥1, but its proof invokes the vector-valued Mikhlin multiplier theorem (Theorem 4.48), which is valid only for p∈(1,∞). The endpoint p=1 is not a harmless limiting case: the right-hand side of (4.109) is the Littlewood–Paley square function, which is not bounded from L^1 to L^1 (only from L^1 to weak-L^1; see e.g. [Ste70, Chapter IV]). Hence there exist f∈L^1 such that, with u=B(s/2)∗f, the square function of u is not integrable while ∥u∥_{L^1_s}=∥f∥_1<∞. The theorem must therefore be restricted to p∈(1,∞), or the p=1 case must be replaced by a genuine weak-type statement; every later use of Theorem 4.49 at p=1 (for instance, in the Bessel–Besov comparison and in Appendix G) then needs to be re-examined. For 1<p<∞ the argument is standard and appears sound.","section":"§4.6, Theorem 4.49, Eq. (4.109)"}],"minor_comments":[{"comment":"The claimed prerequisites are not consistent with the text: the footnote on page 5 assumes familiarity with Sobolev spaces, and the proof of Lemma 4.6 invokes the Marcinkiewicz interpolation theorem while Theorem 4.49 relies on the Mikhlin multiplier theorem without proof. Please either state the actual prerequisites or supply the missing background.","section":"Preface and §4.1"},{"comment":"The second bullet says s∈(0,1/2), but the proof and Corollary 4.9 include the endpoint s=1/2. The statement should read s∈(0,1/2].","section":"Theorem 3.1"},{"comment":"The running header “–DRAFT–(containserrors...)” should be removed from the submission version, and the section title “Baloney around the regularity theory in Lebesgue spaces” is too informal for a journal text.","section":"Header and title of §4.1"},{"comment":"The phrase “aponsomemeasurableclass” in Definition 4.7 is a typo; it should be “a positive measurable class of functions” or similar.","section":"Definition 4.7"},{"comment":"There are numerous typographical artifacts in the LaTeX (e.g., “Fprexample” in a footnote, “R∋𝑥↦−→” in Proposition 4.1, and the duplicated “we have that th” in Theorem 4.49). A careful proofreading pass is needed.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is explicitly a draft and the header invites error reports; that is appropriate for a preprint, but the version submitted to a journal should be cleaned. The p=1 flaw is real and currently affects a named theorem, but the surrounding theory for 1<p<∞ appears solid, so I view the paper as salvageable through revision rather than as a reject. I did not verify every explicit formula in Chapter 2; those computations would deserve a line-by-line check in a final revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know that this draft of a book on the fractional Laplacian is genuinely useful as an entry point: it collects the standard definitions, gives careful proofs of the equivalences, and works out a large set of explicit examples. The organization is sensible and the constant bookkeeping is honest. But it is not ready as a reference until one serious error is fixed.\n\nThe problem is Theorem 4.49, the main bridge between Bessel potential spaces and Besov spaces. It asserts the two-sided equivalence with the dyadic square function for every p ≥ 1. The proof invokes the vector-valued Mikhlin multiplier theorem, which is valid only for 1 < p < ∞. At p = 1 the endpoint is not just an unproved case; it is false. On the dyadic shell the multiplier 2^{js} φ_j(ξ) \\langle ξ\\rangle^{-s} is comparable to the Littlewood–Paley projection, and the Littlewood–Paley square function is not bounded on L^1, only to weak-L^1. So the right-hand inequality in (4.109) fails: there are f ∈ L^1 with u = B(s/2) * f in L^1_s but the square function not in L^1. Any later statement that uses Theorem 4.49 with p = 1 inherits the gap. This is a load-bearing flaw, not an unproved background theorem.\n\nThe reader's report is otherwise fair. The prerequisites claim is indeed overstated: the preface promises only calculus and measure theory, but the book soon assumes Sobolev spaces and uses Marcinkiewicz and Mikhlin without proof. That is a presentation issue, fixable by being explicit about the audience. The self-citations are not a problem; much of the material is also grounded in Stein and other classical sources.\n\nThe rest of the book, from what I checked, is coherent and the calculations are carefully done. The Kelvin transform section and the explicit examples are the strongest parts.\n\nThis is a book for graduate students and for researchers who want a map of the nonlocal regularity landscape. It deserves a serious referee, but the referee needs to push for a correction to Theorem 4.49 — either restrict to 1 < p < ∞, or state the p = 1 result with the correct weak-type bound. Until then, I would not cite it as an authoritative reference.\n\nRecommendation: engage with it, but require the endpoint fix before publication.","headline":"A useful graduate-level introduction to the fractional Laplacian, but Theorem 4.49 has a false p=1 endpoint that must be fixed before the book can be trusted as a reference.","tokens_in":108332,"tokens_out":2754,"would_cite":false,"duration_ms":26116,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35R11","46E35","47G20","35S05"],"pacs":[],"model":"deepseek-v4-flash","headline":"The book's central claim is that all standard definitions of the fractional Laplacian coincide on a suitable common core of functions.","keywords":["fractional Laplacian","nonlocal operators","Riesz potentials","Bessel potential spaces","Besov spaces","regularity theory","Liouville theorem","Sobolev spaces"],"falsifier":"Compute the fractional Laplacian of a fixed Schwartz function at one point twice: once from the singular-integral definition and once from the Fourier-multiplier formula, using independent numerical quadratures or a computer algebra system. If the two values differ by more than round-off, the claimed equivalence of definitions fails.","tokens_in":1487,"feed_emoji":"📘","tokens_out":2281,"duration_ms":99191,"temperature":0.7,"pith_summary":"This book aims to give a self-contained initiation into the fractional world, centered on the fractional Laplacian. Its load-bearing claim is that the many definitions of this operator—as a singular integral, a Riesz potential, a Fourier multiplier, a heat-semigroup average, and a divergence of a nonlocal gradient—are equivalent on a suitable common core of functions. The first half of the book proves that equivalence and computes explicit examples; the second half develops global and interior regularity theory in Lebesgue, Sobolev, and Besov spaces. A reader who accepts the equivalence obtains a reliable toolbox for studying equations driven by nonlocal operators.","feed_headline":"The fractional Laplacian's eleven definitions are one operator","feed_subtitle":"A research-level introduction proves the integral, Fourier, heat-semigroup, and gradient forms coincide and builds a regularity theory.","key_machinery":"The load-bearing object is the fractional Laplacian $(-\\Delta)^s$, with normalizing constant $c_{n,s} = -2^{2s}\\Gamma((n+2s)/2)/(\\pi^{n/2}\\Gamma(-s))$, defined first as a singular integral with a principal value. The argument is carried by Theorem 1.2, which shows that the integral form, the Riesz-potential form, the Fourier-multiplier form with symbol $(2\\pi|\\xi|)^{2s}$, the heat-semigroup form, and the nonlocal-divergence-of-gradient forms all coincide on the common core of Schwartz-class functions and weaker classes. Later chapters add the Bessel kernel, whose Fourier symbol is $(1+4\\pi^2|\\xi|^2)^{-s}$, as the pivot that turns the equivalence into a regularity theory in Bessel potential, Sobolev, and Besov spaces.","core_discovery":"On the book's own terms, the central discovery is that the fractional Laplacian is one operator wearing many disguises: Theorem 1.2 lists eleven formulas that all define the same object on functions that are smooth enough near the point and decay appropriately at infinity. The proofs derive every formula from the singular-integral definition by matching normalizing constants, using Fourier analysis, Riesz and Bessel potentials, the heat semigroup, and nonlocal gradient and divergence calculus. From this unity the book builds a regularity theory: global solutions of $(-\\Delta)^s u = f$ with $f\\in L^p(\\mathbb{R}^n)$ land in Bessel potential spaces, and hence in Sobolev and Besov spaces, with explicit estimates; interior estimates follow by cutoffs. The book also proves Liouville-type rigidity and records the few explicit calculations that can be done in closed form.","pith_inferences":["If the equivalence is accepted, the higher-order finite-difference representation in formula (1.8) could be pushed further to give a unified treatment of fractional Laplacians of order larger than 2, recovering poly-Laplacians as limiting cases.","The book's explicit formulas for power functions and functions supported on balls could serve as benchmark tests for numerical schemes for nonlocal equations, since exact values are rare.","The Bessel and Besov bridge suggests a natural extension to endpoint or mixed-norm estimates for singular nonlocal equations, tracking low and high frequencies separately."],"forward_implications":["Any one of the equivalent definitions can be used as the starting point for a given problem, so integral, Fourier, or semigroup arguments can be mixed freely.","As $s\\to 0$ the fractional Laplacian converges to the identity, and as $s\\to 1$ it converges to $-\\Delta$, so the family genuinely interpolates between local and nonlocal calculus.","A global solution $u\\in L^p(\\mathbb{R}^n)$ of $(-\\Delta)^s u=f\\in L^p(\\mathbb{R}^n)$ gains Sobolev and Besov regularity with estimates controlled by $\\|u\\|_{L^p}+\\|f\\|_{L^p}$.","The same regularity transfers to bounded domains by cutoff and localization, giving interior estimates for solutions of fractional equations on domains.","The Liouville theorem holds: under a slow-growth condition, entire $s$-harmonic functions are affine when $s>1/2$ and constant when $s\\le 1/2$."],"supporting_citations":[{"why":"supplies the systematic comparison of equivalent definitions of the fractional Laplacian that Theorem 1.2 extends.","marker":"[Kwa17]"},{"why":"provides the higher-order finite-difference representation used for the full range of $s$ and for recovering the classical Laplacian.","marker":"[AJS18]"},{"why":"underpins the Riesz and Bessel potential estimates, the interpolation argument, and the singular-integral background used in Chapter 4.","marker":"[Ste70]"},{"why":"supplies the Bessel-potential-space framework and the multiplier background for the regularity theorems.","marker":"[Abe12]"},{"why":"provides the Fourier-distribution approach to the Liouville-type theorem that Chapter 3 presents.","marker":"[CDL15]"},{"why":"furnishes the explicit hypergeometric calculations for the fractional Laplacian of functions supported on intervals used in Chapter 2.","marker":"[Dyd12]"},{"why":"supplies the point-inversion transformation formulas and the singular-integral estimates used in the explicit examples and in Corollary 4.20.","marker":"[DV24]"}],"fun_headline_variants":["Eleven definitions, one operator: the fractional Laplacian","Fractional Laplacian: all definitions, one operator","One operator, eleven faces: fractional Laplacian","The fractional Laplacian: many forms, one truth","Fractional Laplacian unifies eleven definitions"],"cache_read_input_tokens":110464,"weakest_assumption_plain":"The load-bearing premise is that the book's proof chain is self-contained as promised; the regularity chapters rely on Sobolev-space fluency and on two unproved harmonic-analysis results, so the advertised prerequisite of only calculus and basic measure theory is not actually sufficient for the later parts.","fun_headline_variants_meta":{"raw":{"variants":["Eleven definitions, one operator: the fractional Laplacian","Fractional Laplacian: all definitions, one operator","One operator, eleven faces: fractional Laplacian","The fractional Laplacian: many forms, one truth","Fractional Laplacian unifies eleven definitions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000182,"raw_usage":{"total_tokens":1339,"prompt_tokens":1000,"completion_tokens":339,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":616,"completion_tokens_details":{"reasoning_tokens":260}},"tokens_in":616,"tokens_out":339,"duration_ms":3447,"temperature":1.0,"reasoning_tokens":260,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T11:22:20.255624+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the fractional Laplacian of a fixed Schwartz function at one point twice: once from the singular-integral definition and once from the Fourier-multiplier formula, using independent numerical quadratures or a computer algebra system. If the two values differ by more than round-off, the claimed equivalence of definitions fails.","supporting_citations":[],"review_version":1}