{"id":"79fb9b8b-503b-4730-972a-cede2d5e2f93","arxiv_id":"1908.07383","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For transcendental self-maps of the punctured plane, the escaping set is either connected or has infinitely many components, and the first doubly connected Baker domain is constructed.","lead":"This paper proves that for holomorphic self-maps of the punctured plane, the set of points whose orbits tend to zero or infinity is either one connected piece or infinitely many separate pieces. It also constructs the first example of a doubly connected Baker domain, a type of escaping region that cannot occur for entire functions.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The core trichotomy appears sound, but Theorem 1.2 rests on a compressed transfer through the harmonic-measure lemma: Lemma 3.2 is cited from [OS16] and the reduction to the simply connected case in Theorem 1.3 is not written out.","rationale":"The reader's weakest assumption was Lemma 3.2 itself because it is imported without proof. I agree that Lemma 3.2 is the key external input, but I do not see a defect in the lemma; the more immediate risk is that its hypotheses are verified in the text only by the sentence 'we can assume...' after Baker's theorem. Because every later corollary essential to Theorem 1.2 passes through this point, a small gap here would be load-bearing. I therefore recommend keeping the conditional verdict rather than accepting outright. I also note Example 5 contains an explicit 'it can be shown' calculation; it is not central to the main theorems but should be completed in a final version.","tokens_in":16583,"tokens_out":36852,"duration_ms":387484,"concrete_test":"Rewrite the proof of Theorem 1.3 for the case N=0 (U doubly connected). Apply Lemma 3.2 to G0 = U_1 with the itinerary σ(e), obtaining a harmonic-measure-zero set H_1 ⊂ ∂U_1. Then check, using [Ran95, Theorem 4.3.8] for f: U → U_1, that f^{-1}(H_1) ∩ ∂U has harmonic measure zero relative to U and contains ∂U \\setminus \\widetilde{I}_e(f). If the containment and zero-measure pullback hold, Theorem 1.2 is established; if not, the dichotomy needs a different proof.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorems 1.1, 1.4, and 1.5 are largely independent of the disputed mechanism, but Theorem 1.2 is obtained from Corollary 3.7, which in turn uses Corollary 3.6 and Theorem 1.3. Theorem 1.3 is the only place where the paper invokes Lemma 3.2, a harmonic-measure lemma imported from [OS16] without proof. In the proof of Theorem 1.3, Lemma 3.2 requires the domains G_k = U_{n_k} to be disjoint and simply connected. Baker's theorem guarantees at most one doubly connected Fatou component, and the paper says that by [Ran95, Theorem 4.3.8] 'we can assume, therefore, that U_n is simply connected for n in N0.' This reduction is the load-bearing step: if U itself is the unique doubly connected component, Lemma 3.2 cannot be applied with G0 = U, and the argument must be shifted to U_{N+1}, then pulled back to ∂U. If that pullback fails, Corollary 3.6's conclusion that every bounded component of I_e meets J(f) is unsupported, and Theorem 1.2's exclusion of bounded components of I_e(f)∪{0,∞} collapses. The paper does not show the details of this shift, so the central dichotomy is conditional on this verification.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the dynamics of transcendental self-maps of the punctured plane C* and establishes a trichotomy for the connectivity of the escaping set I(f) and of I(f) ∪ {0,∞}. Theorem 1.1 states that I(f) and each little escaping set I_e(f) are either connected or have infinitely many components. Theorem 1.2 states that I(f) ∪ {0,∞} is connected or has exactly two components, one containing 0 and the other ∞. Together these yield cases (I1)-(I3), and the paper provides examples realizing each case. Theorem 1.3 gives a harmonic-measure result on the boundary of escaping wandering domains, which is used to prove Theorem 1.2. Theorem 1.4 constructs the first example of a transcendental self-map of C* with a doubly connected Baker domain, and Theorem 1.5 shows that if such a domain exists, then the closure of the Baker domain contains both 0 and ∞ and I(f) ∪ {0,∞} is connected.","tokens_in":16885,"tokens_out":5060,"duration_ms":52722,"significance":"The main results are natural analogues of the Rippon-Stallard theory for transcendental entire functions, and the trichotomy for I(f) ∪ {0,∞} is a clean and useful classification. The construction of a doubly connected Baker domain is an important new phenomenon, since Baker domains of transcendental entire functions are simply connected. The proofs of Theorems 1.1 and 1.5 are elegant, and the topological index argument in Lemma 5.3 is particularly clear. The paper also gives concrete examples that are well chosen and appear to correctly illustrate all three connectivity cases. The main reservation is that Theorem 1.2 relies on the proof of Theorem 1.3, and that proof is currently too compressed in a place that is genuinely load-bearing.","major_comments":[{"comment":"The reduction 'We can assume, therefore, that U_n is simply connected for n in N0' is not fully justified. If the unique doubly connected Fatou component is U itself, Lemma 3.2 cannot be applied with G0 = U, and the argument must be shifted to some later U_{N+1} and then pulled back to ∂U. The paper should spell out this shift, state precisely how the harmonic-measure-zero property transfers through the pullback when U is doubly connected, and verify that the resulting domains G_k = U_{n_k} are disjoint and simply connected. This step is essential for Corollary 3.6 and hence for Theorem 1.2.","section":"Section 3, proof of Theorem 1.3"},{"comment":"Lemma 3.2 is quoted from [OS16, Lemma 4.1] without proof, but it is the central mechanism of Theorem 1.3. Since the paper applies the lemma to Fatou components of a transcendental self-map of C*, not to the original entire-function setting, the hypotheses deserve a direct verification. Please include a proof or a detailed restatement of the lemma together with a check that the analyticity, continuity, and boundary-mapping hypotheses hold for the maps f^{n_k - n_{k-1}} restricted to the relevant Fatou components.","section":"Section 3, Lemma 3.2"},{"comment":"The claim that 'each of the components of the preimage of γ in H_{n0+1} is a curve γ'' that is unbounded in C*' is stated without proof. This is a load-bearing point for the conclusion that the closure of the Baker domain contains both 0 and ∞. The proof should justify why every such preimage component is unbounded in C* and why each complementary component of H_{n0} in U contains one of them, since this is not an immediate consequence of the preceding definitions.","section":"Section 5, proof of Theorem 1.5"},{"comment":"The key calculation for Example 5 is omitted: the text says 'it is a calculation that E ⊆ I(cosh z) and that E contains every preimage of the real line.' These are precisely the hypotheses needed for Proposition 4.1, and they are not evident. Please provide the estimates showing that the grid E lies in the escaping set of cosh and that every component of cosh^{-1}(E) is met by E, so that the connectivity conclusion for I∞(f) is actually established.","section":"Section 4, Example 5"}],"minor_comments":[{"comment":"There are several typos: 'satifies' should be 'satisfies', 'Berweiler' should be 'Bergweiler', and 'cuve' should be 'curve'.","section":"Throughout"},{"comment":"The wording 'apart possibly from a set of harmonic measure zero relative to V , points z∈∂V satisfy that, for any R>0, Re f~n(z)>R' is awkward; it should say that for all such z and all R>0 the stated inequalities hold for all sufficiently large n.","section":"Corollary 3.3"},{"comment":"The sentence 'by Picard's theorem, Ae(f) meets both sets H0 and H∞' is terse; it would help to state explicitly that transcendental self-maps of C* have no exceptional values and to explain why this implies that every nonempty backward-invariant escaping set meets both sides of the separating open sets.","section":"Proof of Theorem 1.2"},{"comment":"The deduction that the two halves of the imaginary axis lie in attracting Fatou components is plausible but would benefit from a sentence explaining that the attracting fixed point y0 of f^ on the positive real line gives an attracting fixed point i y0 for f, and similarly for -i y0.","section":"Example 2"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper does two genuinely new things. First, it proves the C* analogue of the Rippon–Stallard dichotomy: for a transcendental self-map of the punctured plane, the escaping set I(f) is either connected or has infinitely many components, and the same holds for the little escaping sets I_e(f). Second, it shows that the union I(f)∪{0,∞} is either connected or splits into exactly two components, one around each essential singularity. The proof strategy is natural—blowing-up property plus harmonic measure on wandering domains—and the adaptation to two essential singularities requires real work because of multiply connected Fatou components. The paper also constructs the first doubly connected Baker domain in C* and proves such a domain forces the closure to contain both 0 and ∞, giving a clean topological obstruction. These are substantive results in transcendental dynamics and deserve a serious referee.\n\nThe organization is clear: background lemmas are cited explicitly, the trichotomy (I1)–(I3) is well motivated, and the examples cover the three cases. The proof of Theorem 1.1 via a general completely invariant set is clean. The examples of type (I3) are convincing, and the use of a lift to prove connectivity of I∞(f) in Example 5 is elegant, though the crucial calculation that E⊆I(f~) is asserted rather than shown. I would like to see that filled in, but it is not load-bearing for the main theorems.\n\nThe soft spot I share with the stress-tester is Theorem 1.3's reliance on [OS16, Lemma 4.1]. That lemma is imported without proof, and the reduction to the simply connected case is compressed: Baker's theorem allows at most one doubly connected component, and the paper says we can assume all U_n are simply connected after pulling back by a suitable iterate. The pullback argument is plausible and standard, but it is not written out. If that step failed, Corollaries 3.6 and 3.7 and hence Theorem 1.2 would be unsupported. I do not see an actual gap—the logic appears sound—but a referee should push for details. The note about the requirement that the domains G_k in Lemma 3.2 be disjoint is handled by selecting n_k, and the argument can be shifted past the unique doubly connected component. Still, the paper should spell out why the harmonic measure zero property pulls back correctly when the starting domain U is itself the doubly connected one.\n\nSelf-citation is not a problem here: the cited results from [Mar18], [Mar19], and [EMS19] provide background structure, and the main new theorems are independent of those papers. The Example 5 issue is minor and localized. Overall, the paper is honest, well-written, and mathematically significant. I would recommend sending it to peer review with a request for a careful referee report on Section 3.","headline":"Solid generalization of Rippon–Stallard to the punctured plane, with a new doubly connected Baker domain; the main trichotomy holds up, though one imported lemma and an unshown example calculation merit attention.","tokens_in":17416,"tokens_out":728,"would_cite":true,"duration_ms":9859,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["37F10","30D05"],"pacs":[],"model":"deepseek-v4-flash","headline":"For every transcendental self-map of the punctured plane, the escaping set is either connected or has infinitely many components.","keywords":["holomorphic dynamics","escaping set","punctured plane","connectivity","Baker domain","Fatou components","transcendental self-map","harmonic measure"],"falsifier":"Exhibit a transcendental self-map $f$ of $\\mathbb{C}^*$ whose set $I(f)\\cup\\{0,\\infty\\}$ has a component entirely contained in $\\mathbb{C}^*$ (bounded away from both 0 and $\\infty$), or whose escaping set $I(f)$ has exactly two components. Theorems 1.2 and 1.1 respectively forbid both possibilities, so either would directly refute the paper's central claims.","tokens_in":16354,"feed_emoji":"♾️","tokens_out":4755,"duration_ms":49842,"temperature":0.7,"pith_summary":"This paper settles, for holomorphic self-maps of the punctured plane, the possible shapes of the escaping set: the points whose orbits accumulate only at 0 or infinity. It proves that the escaping set is either connected or splits into infinitely many components, and that when 0 and infinity are added, the enlarged set is either connected or has exactly two components, one attached to each singularity. This yields a clean trichotomy, with examples showing each case occurs. The paper also constructs the first example of a doubly connected Baker domain in the punctured plane, a phenomenon impossible for transcendental entire functions, and shows such a domain forces the escaping set plus singularities to be connected.","feed_headline":"Escaping set in punctured plane: connected or infinite pieces","feed_subtitle":"A trichotomy for I(f)∪{0,∞}, plus the first doubly connected Baker domain in C*, impossible for entire functions.","key_machinery":"The central object is the essential itinerary: the sequence $e=(e_n)\\in\\{0,\\infty\\}^{\\mathbb{N}_0}$ recording whether $|f^n(z)|\\le 1$ or $|f^n(z)|>1$. The paper partitions $I(f)$ into little escaping sets $I_e(f)$ and their finer immediate counterparts $\\widetilde{I}_e(f)$, which are the natural containers for escaping Fatou components. The load-bearing mechanism is a harmonic-measure result, quoted as Lemma 3.2, asserting that along a wandering orbit inside a simply connected domain, the boundary points whose iterates stay at spherical distance greater than $c\\rho$ from the limit point form a set of harmonic measure zero. Theorem 1.3 applies this to show that, for an escaping wandering domain, the boundary points not in the appropriate $\\widetilde{I}_e(f)$ have harmonic measure zero; Corollaries 3.6 and 3.7 then rule out bounded components of $I_e(f)\\cup\\{0,\\infty\\}$, which is exactly what proves Theorem 1.2. For the Baker-domain construction, the mechanism is a Carleman-type approximation theorem that produces a self-map $f(z)=\\exp(g(z+3R)+z^{-2})$ with an invariant doubly connected absorbing set; a lift argument and an index computation ($\\mathrm{ind}(f)=0$) then yield Theorem 1.5.","core_discovery":"For a transcendental self-map $f$ of $\\mathbb{C}^*$, the escaping set $I(f)$ cannot have finitely many components beyond one: Theorem 1.1 says $I(f)$ is either connected or has infinitely many components, and the same holds for each little escaping set $I_e(f)$ indexed by an essential itinerary. Theorem 1.2 says $I(f)\\cup\\{0,\\infty\\}$ is either connected or consists of exactly two components, one containing 0 and the other $\\infty$. Together these give the trichotomy (I1)-(I3), and the paper supplies functions realizing each case. In addition, Theorem 1.4 constructs a transcendental self-map of $\\mathbb{C}^*$ with a doubly connected Baker domain, disproving for this setting the classical theorem that Baker domains are simply connected; Theorem 1.5 shows that if such a domain exists, then $\\mathrm{ind}(f)=0$ and the closure of the domain contains both essential singularities, so $I(f)\\cup\\{0,\\infty\\}$ is connected.","pith_inferences":["The same harmonic-measure machinery should transfer to the little fast escaping sets $A_e(f)$ and $A(f)$, giving the same connected-or-infinitely-many dichotomy for fast escaping points; the paper's remark points in this direction.","A natural next question, not addressed here, is whether every possible combination of the (I1)-(I3) cases with the Julia-set trichotomy (J1)-(J3) can be realized; the examples suggest several combinations are attainable.","Since Baker's theorem limits the punctured plane to at most one doubly connected Fatou component, the new doubly connected Baker domain likely represents the only new topological type of Baker domain in this setting; a classification of all Baker domains in $\\mathbb{C}^*$ would be a testable extension.","For transcendental entire functions, whether a disconnected escaping set must have uncountably many components remains open; the punctured-plane result avoids exceptional points via Picard's theorem, so a similar dichotomy there may require new ideas rather than a direct transfer."],"forward_implications":["No transcendental self-map of $\\mathbb{C}^*$ can have an escaping set with exactly two, three, or any finite number greater than one of components; disconnectedness forces infinitely many components.","If $I(f)$ is disconnected but $I(f)\\cup\\{0,\\infty\\}$ is connected, then every component of $I(f)$ must be unbounded in $\\mathbb{C}^*$ and no component can be isolated away from both essential singularities.","If $I(f)\\cup\\{0,\\infty\\}$ is disconnected, its two components are distinguished by which essential singularity they contain, and each little escaping set $I_e(f)$ meets both of them.","Baker domains in the punctured plane are not governed by the simple-connectivity theorem that holds for entire functions: doubly connected Baker domains exist, and any such domain forces $\\mathrm{ind}(f)=0$ and the connectedness of $I(f)\\cup\\{0,\\infty\\}$.","The trichotomy (I1)-(I3) mirrors the known trichotomy for Julia sets in this setting, so the connectivity of the escaping set and the connectivity of the Julia set can be compared case by case."],"supporting_citations":[{"why":"Supplies the entire-function model: the escaping set is either connected or has infinitely many components, and the boundary of an escaping wandering domain contains escaping points; Theorems 1.1 and 1.3 adapt these arguments to $\\mathbb{C}^*$.","marker":"[RS11]"},{"why":"Its Lemma 4.1 is quoted as Lemma 3.2, the harmonic-measure zero statement that Theorem 1.3 and hence Theorem 1.2 depend on.","marker":"[OS16]"},{"why":"Established the basic structure of the escaping set for transcendental self-maps of $\\mathbb{C}^*$: nonempty little escaping sets, $J(f)=\\partial I_e(f)$, and unbounded components; these facts are used throughout.","marker":"[Mar18]"},{"why":"Proved that Fatou components of transcendental self-maps of $\\mathbb{C}^*$ are simply or doubly connected with at most one doubly connected component; used to reduce the harmonic-measure argument to simply connected domains.","marker":"[Bak87]"},{"why":"Proved Baker domains of transcendental entire functions are simply connected, the result whose failure in $\\mathbb{C}^*$ is demonstrated by the construction in Theorem 1.4.","marker":"[Bak84]"},{"why":"Provides the Carleman-type approximation theorem used to construct the transcendental self-map with a doubly connected Baker domain in the proof of Theorem 1.4.","marker":"[Gai87]"},{"why":"Gives lift techniques and prior constructions of Baker domains and escaping wandering domains for transcendental self-maps of $\\mathbb{C}^*$, including the absorbing-set and lift lemmas used in Section 5.","marker":"[Mar19]"}],"fun_headline_variants":["Escaping set: connected or infinitely many pieces","Punctured plane: escaping set's connectivity trichotomy","First doubly connected Baker domain found in C*","In C*, escaping set has 1 or infinitely many components","Baker domains can be doubly connected in punctured plane"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument leans on a quoted harmonic-measure lemma saying that, for a wandering orbit through a sequence of simply connected domains, almost every boundary point must keep returning to the same side of the unit circle as the orbit; if that lemma fails in the punctured-plane setting, the dichotomy for $I(f)\\cup\\{0,\\infty\\}$ would not be established.","fun_headline_variants_meta":{"raw":{"variants":["Escaping set: connected or infinitely many pieces","Punctured plane: escaping set's connectivity trichotomy","First doubly connected Baker domain found in C*","In C*, escaping set has 1 or infinitely many components","Baker domains can be doubly connected in punctured plane"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000203,"raw_usage":{"total_tokens":1382,"prompt_tokens":941,"completion_tokens":441,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":557,"completion_tokens_details":{"reasoning_tokens":364}},"tokens_in":557,"tokens_out":441,"duration_ms":4722,"temperature":1.0,"reasoning_tokens":364,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:20:41.614073+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit a transcendental self-map $f$ of $\\mathbb{C}^*$ whose set $I(f)\\cup\\{0,\\infty\\}$ has a component entirely contained in $\\mathbb{C}^*$ (bounded away from both 0 and $\\infty$), or whose escaping set $I(f)$ has exactly two components. Theorems 1.2 and 1.1 respectively forbid both possibilities, so either would directly refute the paper's central claims.","supporting_citations":[],"review_version":1}