{"id":"c8db53d7-5857-44b2-88c8-bf6c8bb650e2","arxiv_id":"1908.05088","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For exponential maps with Julia set the whole plane, every piecewise-smooth Jordan curve contains a point whose orbit is dense.","lead":"This mathematics paper proves that for exponential maps whose Julia set is the whole complex plane, every piecewise-smooth loop contains at least one point whose forward orbit is dense. This answers a natural question about how orbits of exponential maps can cover the plane, with consequences for the boundaries of 'nice sets' used in inducing schemes.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 2.6's angle-density assertion is the linchpin, and the proof does not justify that f^{k+1}∘γ actually crosses the claimed translates of T.","rationale":"The reader identified Proposition 2.6 as the weakest load-bearing premise, and this stress test agrees. The proof of the proposition is not a complete derivation: the assertion that f^{k+1}∘γ crosses every 2πi-translate of T is exactly the point where a rigorous argument would need to show that the curve cannot slip through the complementary gaps of the Cantor set of hairs. Because the rest of the paper (Lemma 2.7, Proposition 2.8, Corollaries 2.13 and 2.14) is a chain of consequences of this angle-density statement, a failure there would invalidate the main theorem. The paper has genuine supporting material: the use of Viana's smoothness of hairs, standard facts about repelling periodic points, and Montel's theorem are reasonable, and the overall strategy is plausible. However, the note is written as a sequence of sketches: Lemma 2.2 is 'Evident', Lemma 2.4 is 'Elementary', and Proposition 2.6 is a compressed argument. The conditional verdict is therefore appropriate. I do not see a demonstrated internal inconsistency, only missing justification at the decisive point, so I would not reject the paper; I would keep it conditional pending a fully detailed proof of Proposition 2.6 (and the related domain issue in Lemma 2.7).","tokens_in":6292,"tokens_out":37017,"duration_ms":360736,"concrete_test":"Construct a C1 curve γ:[0,1]→S starting on the topmost hair of T and running through the gap between T and ∂S to the strip boundary, with {0}=γ^{-1}(T). Then check whether f^{k+1}∘γ actually intersects every 2πi-translate of T as Proposition 2.6 asserts; if such a γ avoids the other hairs, the 'crosses every translate' step is false. In the same model, verify whether the curve obtained by replacing γ with an iterate f^n∘γ can be kept inside S for the iterates where Lemma 2.4 is invoked; if not, Lemma 2.7's derivation is unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The main theorem rests on Proposition 2.6: every C1 curve γ⊂S with {0} a connected component of γ^{-1}(T) must, under iterates, intersect T in every angle in [0,π]. The proof is a sketch: after Lemma 2.5, it asserts that f^{k+1}∘γ crosses every 2πi-translate of T and that the angles of intersection are 3/C-dense. This is a nontrivial statement about intersections of two 1-dimensional sets. T is a Cantor set of quasihorizontal hairs in S, and it lies at positive distance from ∂S; the complement of T in S contains open gaps. A C1 curve starting on a hair of T can run through such a gap to ∂S without meeting any other hair, so the asserted 'crosses every translate' has no evident topological basis. The proof does not show why f^{k+1}∘γ is forced to meet all those translates, nor why the resulting angles fill a full interval. If this step fails, Lemma 2.7 (circle approximation), Proposition 2.8, Corollary 2.13, and hence Corollary 2.14 collapse. A related gap appears in Lemma 2.7: it applies Proposition 2.6 by saying 'we may assume γ intersects T at γ(0)', tacitly replacing γ by an iterate f^n∘γ; the new curve is not shown to satisfy f^j(γ([0,δ]))⊂S for the iterates where Lemma 2.4 is used, so the expansion estimates lack a valid domain. These gaps are likely repairable, but as written the central claim depends on unproved intersection and density statements.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies exponential maps f(z)=λe^z and claims that when the Julia set is the whole complex plane, every piecewise-smooth Jordan curve contains at least one point whose forward orbit is dense in C. The main theorem is proved by a sequence of lemmas: constructing small Jordan curves around arbitrary nonzero points by iterating a curve that meets the invariant set T, then using Montel's theorem and density of repelling periodic points to reduce the problem to a local curve-alignment argument. The paper also derives a corollary that bounded nice sets with piecewise-smooth boundary cannot intersect the Julia set in this setting.","tokens_in":6599,"tokens_out":2075,"duration_ms":20733,"significance":"If correct, the main theorem is a striking and elegant dynamical-density result: it says that the exponential map, when its Julia set is the whole plane, has dense orbits on every sufficiently regular closed curve. This would strengthen the author's earlier results on dense orbits of lines and would have concrete consequences for the regularity of boundaries of nice sets used in inducing and thermodynamic-formalism arguments. The argument is purely dynamical and does not fit parameters to data; it is a genuine theorem attempt with a clearly stated falsifiable conclusion. The paper is short but engages with a substantive open problem, and the claimed corollary about bounded nice sets is a useful contribution to the area.","major_comments":[{"comment":"The proof of Lemma 2.2 is given as 'Evident', but the lemma is load-bearing: it provides the four circular arcs whose images under f^2 form a small Jordan curve around an arbitrary nonzero z, and this construction is used in Lemma 2.3 and Proposition 2.8. The condition that C1 and C2 do not meet, while the union of C1, C2, and the 2πi-translate of C3 and C4 forms a Jordan curve surrounding y, is not self-evident and deserves a derivation or a figure-free argument.","section":"§2, Lemma 2.2"},{"comment":"Lemma 2.4 is stated with proof 'Elementary', but equation (2.1) is essential: it bounds the argument of Df^n(z) by a sum of arguments of the forward orbit, and this bound is what makes the curves Ta quasi-horizontal and gives the tangent-angle estimates used in Proposition 2.6 and Lemma 2.7. Without a proof, the reader cannot verify the key estimate |arg(Df^n(z))| < 1/50, which is quantitatively important for the density claims. The proof should be written out.","section":"§2, Lemma 2.4"},{"comment":"Proposition 2.6 is the core of the paper, but its proof is a sketch. The step 'f^{k+1}∘γ|[0,ε] crosses each 2jπi-translate of T' is asserted without justification: T has a Cantor structure with open gaps, and a C^1 curve starting on a hair of T can run through a gap to ∂S without meeting other hairs, so the claimed crossing property does not follow from the preceding lemmas alone. The assertion that the angles of intersection are 3/C-dense in a full interval of length π/2 is also not proved. Since Lemma 2.7, Proposition 2.8, and Corollaries 2.13–2.14 all depend on this proposition, this gap is load-bearing and must be filled by a complete argument.","section":"§2, Proposition 2.6"},{"comment":"Lemma 2.7 applies Proposition 2.6 by saying 'we may assume that γ intersects T transversally at γ(0)', which tacitly replaces γ by an iterate f^n∘γ. The new curve is not shown to satisfy the hypothesis of Lemma 2.4 on the relevant subinterval, namely that f^j(γ([0,δ]))⊂S for j=1,...,n. Without this domain condition, the expansion estimates from Lemma 2.4 cannot be invoked, and the subsequent derivative bounds and length estimates lack a valid basis. The phrases 'one can check' and 'for large n' also hide several nontrivial uniformity assertions that should be made explicit.","section":"§2, Lemma 2.7"}],"minor_comments":[{"comment":"The corollary following Definition 1.1 is stated without a number; for cross-referencing convenience it should be numbered, for instance as Corollary 1.5.","section":"§1, Definition 1.1 and Corollary"},{"comment":"The proof of Lemma 2.5 refers to curves W^±_{k,a} accumulating on Ta from above and below; a brief explanation of why γ, with {0} a connected component of γ^{-1}(T), must leave S within a chosen translate of a hair would make the argument easier to follow, especially because the choice of the shift σ^k is not explicitly connected to γ.","section":"§2, Lemma 2.5"},{"comment":"The notation An(γ) is defined as a subset of [0,π) but the proposition concludes equality to [0,π]; the endpoint discrepancy (π is excluded in the definition) is harmless but should be clarified.","section":"§2, Proposition 2.6"},{"comment":"Reference [1] is dated 1988 and is cited for C∞ parametrization of hairs; the author may wish to cite Viana's original paper with full bibliographic details, and to indicate whether the C∞ regularity applies uniformly to the parametrizations used in the proof.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper is elegant and the claimed theorem is significant, but the proof as written is too skeletal in exactly the places where the main claim lives. In particular, Proposition 2.6 is the linchpin and is not convincingly proved. I would encourage the editor to invite a revision with full proofs of Lemmas 2.2, 2.4, 2.5, and especially Proposition 2.6. If the missing intersection-density argument can be supplied without changing the statement, the result would be a strong addition to the literature. I do not see grounds for rejection, but the current version is not yet publishable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take on Dobbs's note. The main theorem is a genuinely nice result: for exponential maps with Julia set the whole plane, every piecewise-smooth Jordan curve contains a point with dense orbit. That generalizes his earlier work where a specific oblique line had dense third iterate, and the consequence for boundaries of nice sets is directly relevant to inducing schemes. The paper is well written and the motivation is clear.\n\nThe bad news: the proof, as written, is a sketch. Proposition 2.6 is the linchpin and it is not proved. The step where f^{k+1}∘γ 'crosses every 2jπi-translate of T' is exactly the kind of intersection statement that needs an argument. A C1 curve that starts on a hair of T can leave the set through a gap without meeting any other hair, so 'crosses' doesn't follow from anything in the text. The angle-density claim is similarly asserted. The stress-test note is on target here. Lemma 2.7 also has a domain problem: it applies Proposition 2.6 to an iterate without verifying the new curve remains in S for the iterates used in Lemma 2.4.\n\nSome smaller gaps: Lemma 2.2 is 'Evident' but actually requires a short argument; Lemma 2.4 is 'Elementary' but it drives the expansion estimates, so it should be proved. None of this is fatal in the sense of an obvious counterexample; the result is probably true. But the current version is not a complete proof. I could imagine an expert filling the gaps in Proposition 2.6, but it might also take real work.\n\nWho is this for? People working on transcendental dynamics, especially on hairs and on nice sets for inducing schemes. They will care about the statement even if the proof needs polish. I would send it to a serious referee, but I would ask for complete proofs of Proposition 2.6, Lemma 2.4, and the relevant part of Lemma 2.7 before acceptance. I would not cite it as a proven theorem until the gaps are closed.","headline":"A genuinely new result about dense orbits on Jordan curves for exponential maps, but the key proposition is a sketch that needs real work before the proof can be trusted.","tokens_in":7158,"tokens_out":2924,"would_cite":false,"duration_ms":28416,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["37F10"],"pacs":[],"model":"deepseek-v4-flash","headline":"For exponential maps with Julia set the whole plane, every piecewise-smooth Jordan curve contains a point whose forward orbit is dense in the entire complex plane.","keywords":["exponential maps","Julia set","dense orbit","Jordan curves","nice sets","transcendental dynamics","hairs","escaping set"],"falsifier":"Exhibit a $C^1$ curve $\\gamma$ in $S$ for which $\\{0\\}$ is a connected component of $\\gamma^{-1}(T)$ but the union of the angle sets $A_n(\\gamma)$ is not dense in $[0,\\pi]$; Proposition 2.6 would fail, and with it the construction of small Jordan curves and Theorem 1. A direct numerical test is to take a horizontal segment in the right half-plane whose left endpoint lies on $T$, iterate it many times, record the angles at which the iterated curve crosses $T$, and look for a fixed angular gap that survives as the segment length and number of iterates grow.","tokens_in":6038,"feed_emoji":"🌀","tokens_out":7762,"duration_ms":75320,"temperature":0.7,"pith_summary":"The paper proves a dense-orbit theorem for the exponential family $f(z)=\\lambda e^z$. If the Julia set of such a map is the whole complex plane, then on every piecewise-smooth Jordan curve there is at least one point whose forward orbit is dense in the whole plane. A companion result says that if a Jordan curve surrounds a point of the Julia set, then either the curve's orbit is dense or the curve is non-smooth on an uncountable set of points. The paper draws the corollary that no bounded nice set with piecewise-smooth boundary can intersect the Julia set. Nice sets are open sets whose forward images never meet their own boundary, and they are the standard building blocks in inducing constructions for ergodic and geometric properties of the dynamics.","feed_headline":"Every smooth closed curve meets a dense orbit","feed_subtitle":"In the whole-plane Julia case, each piecewise-smooth Jordan curve contains a point whose orbit is everywhere dense.","key_machinery":"The central object is $T$, the set of points whose forward orbits stay inside $S$, the union of two horizontal strips in the right half-plane. $T$ splits into uncountably many smooth, nearly horizontal curves $T_a$, indexed by binary sequences; each is called a hair and admits a $C^\\infty$ parametrisation. The load-bearing mechanism is Proposition 2.6: for a $C^1$ curve $\\gamma$ that enters $S$ and meets $T$ only at its endpoint, the angles at which the iterates $f^n\\circ\\gamma$ cross $T$ are dense in $[0,\\pi]$. That angular spreading is then used to make an iterate of a subarc $C^1$-close to a circle of any desired radius, and four such subarcs close up into a small Jordan curve surrounding any prescribed point.","core_discovery":"Theorem 1 is established by showing that any sufficiently regular closed curve can be used as a kind of template: from a small arc of the curve one can manufacture, through finitely many iterates, a tiny Jordan curve around any prescribed non-zero point. Choosing a dense sequence of target points and shrinking the Jordan curves to zero, an intersection argument places a point on the original curve whose orbit visits every neighbourhood of every point of the plane. The core geometric engine is Proposition 2.6, which asserts that a $C^1$ arc entering the right half-plane and touching the invariant set $T$ only at its endpoint will, under iteration, hit $T$ at angles that fill the entire interval $[0,\\pi]$. From that angle-density statement the paper derives the ability to approximate circles by iterated arcs, which is exactly what is needed to surround arbitrary points.","pith_inferences":["Going beyond the paper, if the Proposition 2.6 sketch is completed, the construction appears local enough to yield a dense-orbit point on any $C^1$ curve that enters the right half-plane with a single tangency to $T$, not only on closed Jordan curves.","The same angle-density mechanism may extend to other transcendental entire functions whose Julia set is the whole plane and which admit an invariant family of smooth horizontal hairs; the exponential family is the cleanest case.","One could probe the theorem numerically on a large circle: record how many iterates are needed before the constructed point enters every grid cell, and watch how that time grows as the grid is refined, to test the quantitative strength behind the qualitative density statement."],"forward_implications":["Every piecewise-smooth Jordan curve in the plane contains a point whose forward orbit under the exponential map is dense, whenever the Julia set is the whole plane.","A Jordan curve whose interior meets the Julia set either has a dense orbit as a set or has uncountably many non-smooth points.","No bounded nice set with piecewise-smooth boundary can intersect the Julia set of such an exponential map.","The known result that the third iterate of an oblique line is dense is extended from straight lines to arbitrary piecewise-smooth Jordan curves."],"supporting_citations":[{"why":"It supplies the $C^\\infty$ parametrisation of the hairs $T_a$, which the proof uses to treat $T$ as a union of smooth curves.","marker":"[1]"},{"why":"It establishes that the third iterate of an oblique line is dense under the exponential map, the result this paper generalises to arbitrary piecewise-smooth Jordan curves.","marker":"[4]"},{"why":"It gives the same oblique-line density statement in expository form, serving as a motivating precedent for the dense-orbit conclusion.","marker":"[5]"},{"why":"It provides the framework of nice sets and the bounded nice set example used in the paper's corollary about boundaries of nice sets.","marker":"[2]"}],"fun_headline_variants":["Dense orbit on every smooth curve","Each Jordan curve hides a dense orbit","Curves always meet a dense orbit","Whole-plane Julia: dense orbit on each curve"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole construction rests on Proposition 2.6, the sketched claim that a $C^1$ curve entering $S$ and meeting $T$ only at its endpoint produces iterated intersections with $T$ at every angle in $[0,\\pi]$, together with the unproved 'elementary' Lemma 2.4 that controls the arguments of the iterates.","fun_headline_variants_meta":{"raw":{"variants":["Dense orbit on every smooth curve","Each Jordan curve hides a dense orbit","Curves always meet a dense orbit","Whole-plane Julia: dense orbit on each curve"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00021,"raw_usage":{"total_tokens":1307,"prompt_tokens":738,"completion_tokens":569,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":354,"completion_tokens_details":{"reasoning_tokens":516}},"tokens_in":354,"tokens_out":569,"duration_ms":5848,"temperature":1.0,"reasoning_tokens":516,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:25:26.130199+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit a $C^1$ curve $\\gamma$ in $S$ for which $\\{0\\}$ is a connected component of $\\gamma^{-1}(T)$ but the union of the angle sets $A_n(\\gamma)$ is not dense in $[0,\\pi]$; Proposition 2.6 would fail, and with it the construction of small Jordan curves and Theorem 1. A direct numerical test is to take a horizontal segment in the right half-plane whose left endpoint lies on $T$, iterate it many times, record the angles at which the iterated curve crosses $T$, and look for a fixed angular gap that survives as the segment length and number of iterates grow.","supporting_citations":[{"cited_title":"Viana da Silva","cited_arxiv_id":null,"evidence_quote":"It supplies the $C^\\infty$ parametrisation of the hairs $T_a$, which the proof uses to treat $T$ as a union of smooth curves."},{"cited_title":"Line, spiral, dense","cited_arxiv_id":null,"evidence_quote":"It establishes that the third iterate of an oblique line is dense under the exponential map, the result this paper generalises to arbitrary piecewise-smooth Jordan curves."},{"cited_title":"Dense yet elementary","cited_arxiv_id":null,"evidence_quote":"It gives the same oblique-line density statement in expository form, serving as a motivating precedent for the dense-orbit conclusion."},{"cited_title":"Nice sets and invariant densities in complex dynamics","cited_arxiv_id":null,"evidence_quote":"It provides the framework of nice sets and the bounded nice set example used in the paper's corollary about boundaries of nice sets."}],"review_version":1}