{"id":"1dfe7603-97ed-4371-8c90-17b28c45e36e","arxiv_id":"1908.03037","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"For exponential polynomials satisfying certain argument conditions and a technical condition in the opposite-argument case, the complement of the Julia set intersect the fast escaping set has finite Lebesgue measure.","lead":"This paper gives conditions under which the Fatou set and the complement of the fast escaping set of an exponential polynomial have finite Lebesgue measure. The conditions force the function to grow at least like exp(|z|^alpha) everywhere except on a set of finite measure.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Section 4 side-length bound for the tiling is internally inconsistent: with |z|^{d-1} the tiles are too large for Lemma 3.3 and the annulus argument; the proof appears to require |z|^{-(d-1)}.","rationale":"The reader's weakest assumption was condition (3) in Theorem 1.3, but the more immediate problem is the proof's tiling construction in Section 4. The side-length condition as displayed contradicts the injectivity radius from Lemma 3.3 and the explicit containment S subset D(z0,(sigma/2)|z0|^{-(d-1)}) later in the same proof. Because every subsequent density estimate and the final annulus summation depend on the tiles being small enough for injectivity and localized enough for the annulus inclusion, the proof does not currently establish the theorem. I do not see evidence that the theorem itself is false; the issue appears to be a sign error in the exponent, likely a typo. However, the submitted manuscript's proof is invalid without correction, so the appropriate verdict is conditional acceptance pending verification and correction of the side-length bound. I disagree with the reader's identification of the weakest assumption because the failure is in the geometric tiling rather than in condition (3).","tokens_in":13992,"tokens_out":21653,"duration_ms":217131,"concrete_test":"Check the original TeX or PDF of Section 4 for the exponent in the side-length condition. If it literally reads |z|^{d-1}, compute the diameter of a tile at distance R and compare it with the injectivity radius 2 sigma R^{-(d-1)}; verify that S is not contained in D(z0,(sigma/2)|z0|^{-(d-1)}) and that a tile intersecting ann(r) can extend outside {r/2 <= |z| <= 3r}. Then replace the exponent by -(d-1) and re-run the estimates (5) and (6); if the corrected tiling restores the inclusions and the density bound, the proof is fixable, but the manuscript as submitted needs this correction.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The most load-bearing issue is in Section 4, where the square tiling is defined. The side length s of each square S is required to satisfy sigma/(4 sqrt 2) min_{z in S}|z|^{d-1} <= s <= sigma/sqrt 2 max_{z in S}|z|^{d-1}, with sigma as in Lemma 3.3. For a square at distance R from the origin, this forces s of order R^{d-1}. However, Lemma 3.3 only guarantees injectivity in disks of radius 2 sigma R^{-(d-1)}, which is far smaller than such a tile. The proof then states 'By Lemma 3.3, f is injective in S' and also 'S subset D(z0,(sigma/2)|z0|^{-(d-1)})'. The latter inclusion is compatible only with s of order |z0|^{-(d-1)}, not |z0|^{d-1}. If the displayed inequality is taken literally, a single tile can have diameter R^{d-1}, so the later inclusion union_{S0 intersecting ann(r)} S0 subset {r/2 <= |z| <= 3r} is false, and the annulus summation does not control the measure of C\\(A(f) cap J(f)). The proof is therefore incomplete as written. The likely correction is to replace |z|^{d-1} by |z|^{-(d-1)} in the side-length condition; with that fix, Lemma 3.3 applies and the density and annulus estimates go through. This is an internal inconsistency, not a disagreement with consensus.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies exponential polynomials of the form f(z) = sum_{j=1}^N Q_j(z) exp(b_j z^d + P_j(z)) with d >= 3, deg(P_j) < d, and distinct nonzero b_j. The main results, Theorems 1.1 and 1.3, give conditions on the arguments of the b_j — and, in the 'opposite arguments' case, an additional polynomial condition (3) — under which C \\ (A(f) ∩ J(f)) has finite Lebesgue measure. Since F(f) ⊂ C \\ J(f) and C \\ I(f) ⊂ C \\ A(f), this implies that the Fatou set and the non-escaping set have finite Lebesgue measure. The proof has three parts: Section 2 establishes that |f| and |f'| grow at least like exp(|z|^alpha) for some alpha > 0 outside an exceptional set E_1 of finite Lebesgue measure; Section 3 proves injectivity of f on small disks of radius comparable to |z|^{-(d-1)} near the complement of a slightly larger exceptional set; Section 4 uses a McMullen-type tiling and density argument to show that the iterated preimages of tiles capture almost all of C outside a finite-measure set. Example 1.2 shows that condition (3) is sharp: without it, exp(iz) sinh(z^3) has a superattracting fixed point whose basin has infinite measure.","tokens_in":14347,"tokens_out":19842,"duration_ms":196808,"significance":"If the proof is completed, the paper gives a substantial and fairly general sufficient condition for finiteness of the Fatou set and the complement of the escaping set for exponential polynomials, going beyond the special functions treated by Schubert, Hemke, and Zhang-Yang and complementing the positive-measure results of Sixsmith and Bergweiler-Chyzhykov. The paper is self-contained: all auxiliary lemmas, including the key growth estimate (Lemma 2.4), the injectivity criteria (Lemmas 3.2-3.5), and the final density argument, are proved in the text. The sharpness example is instructive and correctly handled in Section 5. The main caveat is a load-bearing inconsistency in the tiling construction in Section 4, which I describe below; I view it as correctable, but it must be fixed before the proof is valid.","major_comments":[{"comment":"The displayed side-length condition for the squares in the tiling is internally inconsistent with the rest of the proof. Read literally, the inequalities give a side length s of order |z|^{d-1} for a square at distance |z| from the origin. However, Lemma 3.3 only guarantees injectivity on disks of radius 2 sigma |z|^{-(d-1)}, and the proof later asserts that a square S with centre z0 satisfies S subset D(z0, (sigma/2)|z0|^{-(d-1)}). These statements cannot hold simultaneously for large |z|. The construction described immediately afterward — start with fixed-size squares and subdivide until the upper bound is met — also fails for the literal reading, because the lower bound grows with |z|. The intended condition must have the reciprocal exponent: the min and max should appear in the denominator, i.e. s should be comparable to |z|^{-(d-1)}. This correction is load-bearing: the density estimate (6), the bound on the perimeter of f(S), the inclusion used for the Koebe distortion argument, and the annulus sum all rely on tiles having diameter comparable to |z|^{-(d-1)}. I regard this as a fixable typographical/sign error rather than a fatal flaw, but it must be corrected and the surrounding constants (for example the factor 4 sigma sqrt(2) in the perimeter estimate) adjusted accordingly.","section":null}],"minor_comments":[{"comment":"The displayed formula for f'(z) is missing a closing parenthesis at the end; it reads '... exp(bjzd +Pj(z).' and should close the exponential argument as exp(b_j z^d + P_j(z)).","section":null},{"comment":"The symbol S is used both for the collection of squares and for a generic square in the collection; using a script letter for the collection would reduce ambiguity.","section":null},{"comment":"In the bound for meas(f(S) \\ bigcup_{S' in pack(f(S))} S'), the term meas(f(S) cap B0) is not explicitly included; it is harmless because B0 is bounded and max_{z in S} |f'(z)| is large for the squares that matter, but it should be mentioned for completeness.","section":null},{"comment":"Reference [4] contains a typo in the author name: 'Bergweilwer' should be 'Bergweiler'.","section":null}],"recommendation":"major_revision","confidential_remarks":"To the editor: the manuscript is within the journal's scope and the main result is likely correct. The Section 4 exponent error appears to be a typo rather than a conceptual gap, so I recommend major revision rather than rejection. The proof is otherwise careful and self-contained, and I have no concerns about novelty or citation practices."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know: this paper proves genuinely new measure finiteness results for a broad class of exponential polynomials, but the proof of the main theorem has a load-bearing gap in the Section 4 tiling.\n\nWhat is new: Theorems 1.1 and 1.3 cover sums Q_j(z) exp(b_j z^d + P_j(z)) with d≥3, mixed argument directions, and a condition (3) for opposite-argument pairs. This generalizes Hemke's Q1 e^P + Q2 e^{-P}, Zhang–Yang's P(e^z)/e^z, and McMullen/Schubert's sin/sinh cases. The counterexample exp(iz)sinh(z^3) is clean and shows the condition is needed.\n\nThe paper does real work: Section 2 proves the lower bound |f(z)| ≥ exp(|z|^α) outside a finite-measure exceptional set; that bound is proved, not assumed. The injectivity lemmas are standard, and Section 5 verifies the counterexample in detail. The writing is careful and honest about what condition (3) does.\n\nThe soft spot is exactly where the stress-test notes: the tiling side length. The displayed condition in Section 4 says side length s satisfies σ/(4√2) min|z|^{d−1} ≤ s ≤ σ/√2 max|z|^{d−1}. For a square near radius R, this forces s ≈ R^{d−1}, which is huge. Lemma 3.3 only guarantees injectivity in disks of radius 2σ|z|^{-(d−1)}, i.e., tiny for large |z|. The later claim 'S ⊂ D(z0, (σ/2)|z0|^{-(d−1)})' cannot hold for such a square. The annulus argument also relies on squares intersecting ann(r) lying in {r/2 ≤ |z| ≤ 3r}, which fails if squares have diameter R^{d−1}. So the proof of Theorem 1.3 is incomplete as written. The most likely fix is to replace |z|^{d−1} by |z|^{-(d−1)} in that one displayed chain; then Lemma 3.3 applies and the density/annulus estimates seem to go through. I haven't checked every constant, but nothing else looks broken.\n\nThe result is probably true and worth pursuing. My recommendation: do not desk reject. Send it to a referee for complex dynamics; the referee will likely demand the tiling fix, and after that the paper should be publishable. For a reading group, it's a useful example of spotting a typo-scale flaw that derails a proof, and of how the correct exponent should be negative.\n\nYes, serious thinker: the paper is coherent and the gap looks like an honest typo, not a hidden circularity.","headline":"Genuinely new measure finiteness results for exponential polynomials, but Section 4's tiling side-length bound appears inverted and the proof as written doesn't close.","tokens_in":14902,"tokens_out":3248,"would_cite":false,"duration_ms":31337,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["37F10","30D05","30D20"],"pacs":[],"model":"deepseek-v4-flash","headline":"For a class of exponential polynomials, the Fatou set and the non-escaping set have finite Lebesgue measure.","keywords":["exponential polynomial","Fatou set","Julia set","fast escaping set","Lebesgue measure","iteration of entire functions","finite measure"],"falsifier":"For any function satisfying the hypotheses except condition (3), inspect the set where $|f(z)|\\(\\le 1\\)$. The paper shows that for $h(z)=\\exp(iz)\\sinh(z^3)$, the infinite-measure set $B=\\{re^{i\\theta}: |\\theta-\\pi/2|\\le 1/(r^2\\log r)\\}$ is mapped into a small disk around zero, so it lies in the attracting basin. If a function satisfying condition (3) also had an infinite-measure region where $|f|$ stays bounded, the theorem would fail; checking this bound along the curves $\\arg z\\approx \\pm\\pi/(2d)$ is a direct numerical test.","tokens_in":13797,"feed_emoji":"📏","tokens_out":7206,"duration_ms":79363,"temperature":0.7,"pith_summary":"The paper proves a finiteness result for the geometry of iteration: for a broad class of exponential polynomials $$f(z)=\\sum_{j=1}^N Q_j(z)\\exp(b_j z^d+P_j(z)),\\qquad d\\ge 3,\\quad \\deg P_j<d,$$ with the leading coefficients $b_j$ occupying distinct directions (ordered arguments, each gap at most $\\pi$, and either no opposite pair or an extra phase-matching condition), the complement of $A(f)\\cap J(f)$ has finite Lebesgue measure. Here $A(f)$ is the fast escaping set and $J(f)$ the Julia set. Since $A(f)$ is contained in the escaping set and $J(f)$ is the complement of the Fatou set, this immediately gives finite Lebesgue measure for the Fatou set and for the non-escaping set. The interest is that these sets can have infinite area for nearby examples such as $\\sin(z)$ or $\\sin(z^2)$, so the result identifies a precise threshold where the area becomes finite.","feed_headline":"Exponential polynomials with finite-area Fatou sets","feed_subtitle":"Under separated phase conditions, almost every point escapes to infinity; the technical exception is sharp.","key_machinery":"The load-bearing objects are the phase-difference polynomials $P_{j,k}(z)=(b_j-b_k)z^d+(P_j-P_k)(z)$ and the exceptional sets $E_l=\\bigcup_{j\\ne k}P_{j,k}^{-1}(U_l)$, where $U_l=\\{w:|\\Re w|<l|w|^{\\nu/d}\\}$ with $\\nu=d-\\tfrac52$. These sets have finite Lebesgue measure; outside them one summand dominates, so $f$ behaves like a single exponential and satisfies the lower bound $|f(z)|\\ge\\exp(|z|^\\alpha)$ for $\\alpha<\\nu$. Condition (3) enters exactly to keep this lower bound valid in the case of opposite leading arguments, where without it the function can be bounded on an infinite-area set. Injectivity on small disks centered outside $E_2$, obtained from derivative estimates and the standard distortion lemma, allows the construction of nested preimages inside a square tiling.","core_discovery":"The central claim is Theorem 1.3: under the hypotheses above, and with an additional condition (3) when $\\arg b_{j+1}=\\arg b_j+\\pi$ or $\\arg b_N=\\arg b_1+\\pi$, the Lebesgue measure of $\\mathbb{C}\\setminus(A(f)\\cap J(f))$ is finite. The proof works by showing that $|f(z)|\\ge \\exp(|z|^\\alpha)$ for some $\\alpha>0$ on the complement of a finite-area exceptional set, then using a square-mesh construction to build a nested family of sets inside $A(f)\\cap J(f)$ whose leftover density decays faster than any exponential. Theorem 1.1 is the special case with strictly separated arguments, and the theorem also covers the functions $Q_1(z)e^{P(z)}+Q_2(z)e^{-P(z)}$ studied earlier. The example $h(z)=\\exp(iz)\\sinh(z^3)$ is shown to violate the conclusion without condition (3): its attracting basin of zero has infinite measure.","pith_inferences":["The proof's exponent $\\nu=(d-5)/2$ is positive only for $d\\ge 3$, which suggests that extending the theorem to $d=2$ would require a genuinely new mechanism rather than a minor modification.","Condition (3) has a geometric reading: opposite exponentials must share a common polynomial $g$ in their phases, making their leading level sets parallel; the paper does not explore this geometric interpretation.","A natural next step would be to ask whether the same hypotheses give lower bounds on the Hausdorff or packing dimension of the Julia set, since finite-area complements plus fast escape often force substantial fractal structure.","The same square-mesh density construction may adapt to entire perturbations with slowly growing prefactors, provided the domination and injectivity estimates survive."],"forward_implications":["For every function in the stated class, the Fatou set has finite Lebesgue measure, so almost every point lies in the Julia set and in the fast escaping set.","The non-escaping set has finite Lebesgue measure, meaning almost every starting point escapes to infinity under iteration.","Theorem 1.1 covers sums of exponentials with strictly separated arguments, without any additional phase-matching condition.","The theorem subsumes the earlier finite-measure result for $f=Q_1 e^{P}+Q_2 e^{-P}$ with $\\deg P\\ge 3$.","The proof gives a quantitative annulus estimate: the part of a large annulus not already in $A(f)\\cap J(f)$ has measure at most $\\exp(-c r^\\alpha)$, which forces the total exceptional set to be finite."],"supporting_citations":[{"why":"Supplies the square-mesh construction and density argument used to build the nested subsets of the fast escaping Julia set.","marker":"[9]"},{"why":"Provides the criterion (Lemma 4.2) that places orbit points in the Julia set once the derivative ratio is bounded below.","marker":"[13]"},{"why":"Supplies the comparison lemma for iterates of the maximum modulus used to certify that the constructed points lie in the fast escaping set.","marker":"[3]"},{"why":"Establishes that the Fatou set of an exponential polynomial has no multiply connected components, ruling out that alternative in the Julia-set criterion.","marker":"[15]"},{"why":"Provides the distortion estimates for injective holomorphic maps used to control derivative ratios on nested squares.","marker":"[11]"},{"why":"Gives the earlier finite-measure result for $Q_1 e^P + Q_2 e^{-P}$, which the present theorem subsumes.","marker":"[8]"},{"why":"Provides the basic example where finite measure was already known for $\\sinh$, framing the problem treated here.","marker":"[12]"},{"why":"Introduces the fast escaping set $A(f)$, the central object of the theorem.","marker":"[5]"}],"fun_headline_variants":["Finite measure for Fatou and non-escaping sets","Almost every point escapes for exponential polynomials","Exponential polynomials with finite exceptional area","Fatou sets of finite area for exponential polynomials"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof's global lower bound $|f(z)|\\ge \\exp(|z|^\\alpha)$ outside a finite-area set rests on condition (3) when two leading exponents point in opposite directions; if that condition fails, the function can be bounded on an infinite-area set and the conclusion is false.","fun_headline_variants_meta":{"raw":{"variants":["Finite measure for Fatou and non-escaping sets","Almost every point escapes for exponential polynomials","Exponential polynomials with finite exceptional area","Fatou sets of finite area for exponential polynomials"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001414,"raw_usage":{"total_tokens":5642,"prompt_tokens":809,"completion_tokens":4833,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":425,"completion_tokens_details":{"reasoning_tokens":4775}},"tokens_in":425,"tokens_out":4833,"duration_ms":36663,"temperature":1.0,"reasoning_tokens":4775,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:27:22.818099+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For any function satisfying the hypotheses except condition (3), inspect the set where $|f(z)|\\(\\le 1\\)$. The paper shows that for $h(z)=\\exp(iz)\\sinh(z^3)$, the infinite-measure set $B=\\{re^{i\\theta}: |\\theta-\\pi/2|\\le 1/(r^2\\log r)\\}$ is mapped into a small disk around zero, so it lies in the attracting basin. If a function satisfying condition (3) also had an infinite-measure region where $|f|$ stays bounded, the theorem would fail; checking this bound along the curves $\\arg z\\approx \\pm\\pi/(2d)$ is a direct numerical test.","supporting_citations":[{"cited_title":"McMullen , Area and Hausdorff dimension of Julia sets of entire functions, Trans","cited_arxiv_id":null,"evidence_quote":"Supplies the square-mesh construction and density argument used to build the nested subsets of the fast escaping Julia set."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the criterion (Lemma 4.2) that places orbit points in the Julia set once the derivative ratio is bounded below."},{"cited_title":"Bergweiler , Lebesgue measure of Julia sets and escaping sets of certain entire functions, Fund","cited_arxiv_id":null,"evidence_quote":"Supplies the comparison lemma for iterates of the maximum modulus used to certify that the constructed points lie in the fast escaping set."},{"cited_title":"Zheng , On multiply-connected Fatou components in iteration of meromor- phic functions, J","cited_arxiv_id":null,"evidence_quote":"Establishes that the Fatou set of an exponential polynomial has no multiply connected components, ruling out that alternative in the Julia-set criterion."},{"cited_title":"Pommerenke , Univalent functions , Studia Mathematica/Mathematische Lehrb¨ ucher XXV, Vandenhoeck & Ruprecht, G¨ ottingen, 1975","cited_arxiv_id":null,"evidence_quote":"Provides the distortion estimates for injective holomorphic maps used to control derivative ratios on nested squares."},{"cited_title":"Hemke , Recurrence of entire transcendental functions with simple post- singular sets, Fund","cited_arxiv_id":null,"evidence_quote":"Gives the earlier finite-measure result for $Q_1 e^P + Q_2 e^{-P}$, which the present theorem subsumes."},{"cited_title":"Schubert, Area of Fatou sets of trigonometric functions, Proc","cited_arxiv_id":null,"evidence_quote":"Provides the basic example where finite measure was already known for $\\sinh$, framing the problem treated here."},{"cited_title":"Bergweiler and A","cited_arxiv_id":null,"evidence_quote":"Introduces the fast escaping set $A(f)$, the central object of the theorem."}],"review_version":1}