betaFeedback

WorksMathematics

Cofinite derivative zeros at the order-one threshold

Abstract

We give a complete probabilistic construction of a transcendental entire function for which every nonempty open set contains a zero of every sufficiently high derivative. The construction has log⁡M(r)=O(rlog⁡rlog⁡log⁡r)\log M(r)=O(r\log r\log\log r), hence order exactly one and infinite exponential type. No transcendental entire function of finite exponential type can have this zero property. We also establish a local limiting zero-counting measure for the constructed derivatives. The original existence assertion has prior solution claims, including Eric Hou's preprint. The order-one construction and growth obstruction are the proposed strengthening.

Problem, scope, and attribution

The prize catalogue lists JSP-000752 as Open on the source snapshot read on October 4, 2026 [1]. It asks for an entire function with dense derivative zero sets along every infinite subsequence of derivative orders. The natural, nontrivial interpretation requires a transcendental entire function: a nonzero polynomial would satisfy the literal version because all sufficiently high derivatives vanish identically. This is Erdős Problem 906 in Bloom’s database [3].

The prize rules permit a complete mathematical solution to be supplied before Lean formalization [2]. Its catalogue flag “Eligible to claim” has a different meaning: an award claim requires both an accepted mathematical solution and verified formalization. This manuscript supplies mathematical proof evidence for review; no award entitlement or maintainer acceptance is asserted.

The bounded random-series and zero-free-disk framework has a clear precedent in Hou [4], whose stated construction has order two. Earlier probabilistic proposals are linked in the discussion of Problem 906 [3]. The existence answer itself is therefore not claimed as a new breakthrough. The construction below replaces the Fock weights with slowly increasing derivative weights and proves a stronger growth conclusion. It uses an inverse-amplitude moment and Jensen’s inequality; no Laguerre asymptotics or external random-zero concentration theorem is needed. The order-one strengthening was not located in the cited prior sources; this is not a claim of verified first-discovery priority.

Write B(c,R)={z:∣z−c∣<R}B(c,R)=\{z:|z-c|<R\} and Mf(r)=max⁡∣z∣=r∣f(z)∣M_f(r)=\max_{|z|=r}|f(z)|. The zero property we will prove is

∀U⊂C nonempty and open,∃NU ∀n≥NU,∃z∈U:f(n)(z)=0.(1)\forall U \subset\mathbb{C}\ \text{nonempty and open},\quad\exists N_U\ \forall n \ge N_U,\quad\exists z \in U:f^{(n)}(z)=0. \tag*{(1)}

All derivative orders are nonnegative integers.

Lemma 1 (The required quantifier). An entire function has property (C) if and only if the union of the zero sets of f(nk)f^{(n_k)} is dense for every strictly increasing sequence (nk)(n_k) of derivative orders.

Proof. If (C) holds, every such sequence eventually exceeds NUN_U, so the union meets each nonempty open UU. Conversely, if (C) fails for one UU, infinitely many derivative orders have no zero in UU. Enumerating those orders gives a strictly increasing sequence whose union of zero sets misses UU. □\square

A sharp order threshold

Theorem 2. There is a transcendental entire function ff with property (C) such that

log⁡Mf(r)=O(rlog⁡rlog⁡log⁡r)(r→∞).(2)\log M_f(r)=O(r\log r\log\log r)\qquad(r\to\infty). \tag*{(2)}

It has order exactly one and infinite exponential type. Moreover, no transcendental entire function satisfying (C) has finite exponential type. Thus one is the smallest possible order, and this order is attained.

Finite exponential type here means that ∣g(z)∣≤CeT∣z∣|g(z)|\le Ce^{T|z|} for some finite C,T>0C,T>0 and all zz. The usual order is ρ(g)=lim sup⁡r→∞log⁡log⁡Mg(r)/log⁡r\rho(g)=\limsup_{r\to\infty}\log\log M_g(r)/\log r.

Proposition 3 (Finite exponential type is impossible). A transcendental entire function of finite exponential type cannot satisfy (C).

Proof. Suppose ∣g(z)∣≤CeT∣z∣|g(z)|\le Ce^{T|z|} with T>0T>0. Fix R>0R>0, and let

mn(R)=max⁡∣z∣≤R∣g(n)(z)∣.m_n(R)=\max_{|z|\le R}|g^{(n)}(z)|.

If (C) holds, for all n≥NRn\ge N_R choose an∈B(0,R)a_n\in B(0,R) with g(n)(an)=0g^{(n)}(a_n)=0. The line segment from ana_n to any zz in the closed disk lies in that disk. The fundamental theorem of calculus along this segment gives

∣g(n)(z)∣≤∣z−an∣mn+1(R)≤2Rmn+1(R).|g^{(n)}(z)|\le|z-a_n|m_{n+1}(R)\le2Rm_{n+1}(R).

Consequently

mn(R)≥mNR(R)(2R)−(n−NR)(n≥NR).(3)m_n(R)\ge m_{N_R}(R)(2R)^{-(n-N_R)}\qquad(n\ge N_R). \tag*{(3)}

The factor mNR(R)m_{N_R}(R) is strictly positive, since a derivative of a transcendental entire function is not identically zero. Cauchy’s estimate on a circle of radius tt about zz, ∣z∣≤R|z|\le R, gives

mn(R)≤CeTRn!eTttn.m_n(R)\le Ce^{TR}n!\frac{e^{Tt}}{t^n}.

Take t=n/Tt=n/T and use n!≤nnn!\le n^n. For n≥1n\ge1 this yields

mn(R)≤CeTR(eT)n.(4)m_n(R)\le Ce^{TR}(eT)^n. \tag*{(4)}

Choosing 0<R<(2eT)−10<R<(2eT)^{-1} contradicts (3)–(4), because (2R)−1>eT(2R)^{-1}>eT. □\square

The weighted random entire function

Set A=eeA=e^e and, for t≥0t\ge0, define

w(t)=log⁡(t+A)log⁡log⁡(t+A),Wn=w(n).(5)w(t)=\log(t+A)\log\log(t+A),\qquad W_n=w(n). \tag*{(5)}

Define b0=1b_0=1 and bn=∏j=1nWjb_n=\prod_{j=1}^{n}W_j for n≥1n\ge1. Let (ξj)j≥0(\xi_j)_{j\ge0} be independent random variables, each uniformly distributed with respect to area on the closed unit disk. Our entire function is

f(z)=∑j=0∞ξjbjj!zj.(6)f(z)=\sum_{j=0}^{\infty}\xi_j\frac{b_j}{j!}z^j. \tag*{(6)}

These are genuine random complex coefficients; the probabilistic argument extracts one deterministic sample at the end.

We will use only the following elementary facts about ww:

Wn≥1,w is increasing,W2nWn⟶1,Wnlog⁡n⟶∞,Wnn⟶0.(7)W_n\ge1,\qquad w\text{ is increasing},\qquad\frac{W_{2n}}{W_n}\longrightarrow1,\qquad\frac{W_n}{\log n}\longrightarrow\infty,\qquad\frac{W_n}{\sqrt n}\longrightarrow0. \tag*{(7)}

Indeed, for t≥0t\ge0, w′(t)=(log⁡log⁡(t+A)+1)/(t+A)>0w'(t)=(\log\log(t+A)+1)/(t+A)>0, and the remaining limits follow from w(t)∼log⁡tlog⁡log⁡tw(t)\sim\log t\log\log t.

Lemma 4 (Normal convergence and deterministic growth). Every sample in (6) is entire. For all sufficiently large rr, uniformly over all samples,

Mf(r)≤2exp⁡(rw(⌈r2⌉)).(8)M_f(r)\le2\exp\bigl(rw(\lceil r^2\rceil)\bigr). \tag*{(8)}

Almost surely all Taylor coefficients are nonzero, so ff is transcendental almost surely.

Proof. Monotonicity gives bj≤Wjjb_j\le W_j^j. The elementary factorial bound j!≥(j/e)jj!\ge(j/e)^j gives

(bjj!)1/j≤eWjj⟶0.\left(\frac{b_j}{j!}\right)^{1/j}\le\frac{eW_j}{j}\longrightarrow0.

Thus the series converges normally on compact sets, uniformly in the bounded coefficients. The same is true after any fixed number of termwise differentiations.

For the growth bound put J=⌈r2⌉J=\lceil r^2\rceil. For 0≤j≤J0\le j\le J we have bj≤WJjb_j\le W_J^j, so the sum of these absolute terms is at most exp⁡(rWJ)\exp(rW_J). For j>Jj>J, r≤jr\le\sqrt j, and therefore

bjrjj!≤(erWjj)j≤(eWjj)j.\frac{b_jr^j}{j!}\le\left(\frac{e rW_j}{j}\right)^j\le\left(\frac{eW_j}{\sqrt j}\right)^j.

By (7), for all sufficiently large rr the last quantity is at most 2−j2^{-j} for every j>Jj>J. The tail is at most 2−J≤12^{-J}\le1, which is at most the preceding exponential bound. This proves (8). Finally each coefficient law is nonatomic, so P(ξj=0)=0\mathbb{P}(\xi_j=0)=0 for each jj. A countable union of null events still has probability zero. □\square

Normalize the derivatives by

Gn(z)=f(n)(z)bn=∑k=0∞ξn+kBn,kzkk!,Bn,k=bn+kbn=∏j=n+1n+kWj.(9)G_n(z)=\frac{f^{(n)}(z)}{b_n}=\sum_{k=0}^{\infty}\xi_{n+k}B_{n,k}\frac{z^k}{k!},\qquad B_{n,k}=\frac{b_{n+k}}{b_n}=\prod_{j=n+1}^{n+k}W_j. \tag*{(9)}

The empty product is one. Normalization does not change any zero set.

Lemma 5 (A uniform upper bound for a derivative). For every fixed H>0H>0, for all sufficiently large nn and all 0≤r≤H0\le r\le H,

Sn(r):=∑k=0∞Bn,krkk!≤erW2n+2−n≤2erW2n.(10)S_n(r):=\sum_{k=0}^{\infty}B_{n,k}\frac{r^k}{k!}\le e^{rW_{2n}}+2^{-n}\le2e^{rW_{2n}}. \tag*{(10)}

In particular ∣Gn(z)∣≤2e∣z∣W2n|G_n(z)|\le2e^{|z|W_{2n}} for ∣z∣≤H|z|\le H.

Proof. For k≤nk \le n, all factors in Bn,kB_{n,k} are at most W2nW_{2n}. These terms sum to at most erW2ne^{rW_{2n}}. For k>nk>n, we have n+k≤2kn+k \le2k, hence

Bn,krkk!≤(eHw(2k)k)k.\frac{B_{n,k}r^k}{k!} \le\left(\frac{eHw(2k)}{k}\right)^k.

Since w(2k)/k→0w(2k)/k \to0, for all sufficiently large nn the last quantity is at most 2−k2^{-k} simultaneously for every k>nk>n. The sum of these terms is 2−n2^{-n}. The last assertion uses ∣ξj∣≤1|\xi_j| \le1.

One-coordinate anti-concentration

Lemma 6 (An inverse-amplitude moment). If ξ\xi is uniform on the unit disk, a≠0a \ne0 and v∈Cv \in\mathbb{C}, then

P(∣aξ+v∣<s)≤min⁡{1,s2/∣a∣2}(s>0),E1∣aξ+v∣≤2∣a∣.(11)\mathbb{P}(|a\xi+v|<s) \le\min\{1,s^2/|a|^2\} \quad(s>0), \qquad\mathbb{E}\frac{1}{|a\xi+v|} \le\frac{2}{|a|}. \tag*{(11)}

Proof. The first inequality follows by comparing areas: the preimage of the disk of radius ss has area at most πs2/∣a∣2\pi s^2/|a|^2 in the unit disk. The layer-cake formula and this inequality give

E1∣aξ+v∣≤∫01/∣a∣1 dt+∫1/∣a∣∞1∣a∣2t2 dt=2∣a∣.\mathbb{E}\frac{1}{|a\xi+v|} \le\int_0^{1/|a|}1\,dt+\int_{1/|a|}^{\infty}\frac{1}{|a|^2t^2}\,dt=\frac{2}{|a|}.

Lemma 7 (A near-peak coefficient). For t≥1t \ge1 and m=⌊t⌋m=\lfloor t\rfloor,

tmm!≥ete2t.(12)\frac{t^m}{m!}\ge\frac{e^t}{e^2t}. \tag*{(12)}

Consequently, if ∣z∣Wn≥1|z|W_n \ge1, the term with m=⌊∣z∣Wn⌋m=\lfloor|z|W_n\rfloor in (9) has coefficient a=Bn,mzm/m!a=B_{n,m}z^m/m! satisfying

∣a∣≥e∣z∣Wne2∣z∣Wn.|a|\ge\frac{e^{|z|W_n}}{e^2|z|W_n}.

Proof. For m≥1m \ge1, comparison with an integral gives

log⁡(m!)≤mlog⁡m−m+1+log⁡m.\log(m!)\le m\log m-m+1+\log m.

Thus

log⁡tmm!≥mlog⁡(t/m)+m−1−log⁡m≥t−2−log⁡t,\log\frac{t^m}{m!}\ge m\log(t/m)+m-1-\log m\ge t-2-\log t,

because m≤tm\le t, m≥t−1m\ge t-1, and log⁡m≤log⁡t\log m\le\log t. Monotonicity of ww gives Bn,m≥WnmB_{n,m}\ge W_n^m, proving the second assertion.

Lemma 8 (Uniform inverse moments on a compact set). For any H>0H>0, put H+=max⁡{1,H}H_+=\max\{1,H\}. For every n≥1n\ge1 and ∣z∣≤H|z|\le H,

E(eWn∣z∣∣Gn(z)∣)≤2e2H++Wn.(13)\mathbb{E}\left(\frac{e^{W_n|z|}}{|G_n(z)|}\right)\le2e^2H_++W_n. \tag*{(13)}

The value at Gn(z)=0G_n(z)=0 is interpreted as +∞+\infty.

Proof. If ∣z∣Wn≥1|z|W_n \ge1, choose mm as in Lemma 7 and condition on all coefficients except ξn+m\xi_{n+m}. The normally convergent series becomes aξn+m+va\xi_{n+m}+v, with vv fixed under this conditioning. Independence, Lemma 6, and Lemma 7 give

E(eWn∣z∣∣Gn(z)∣∣all other coefficients)≤2e2∣z∣Wn≤2e2H+Wn.\mathbb{E}\left(\left.\frac{e^{W_n|z|}}{|G_n(z)|}\right| \text{all other coefficients}\right) \le2e^2|z|W_n \le2e^{2H_+W_n}.

Integration removes the conditioning. If ∣z∣Wn<1|z|W_n<1, condition instead on ξn\xi_n, whose coefficient in Gn(z)G_n(z) is one. The left side is then at most 2e2e, which is no larger than the right side since H+,Wn≥1H_+,W_n\ge1. This includes z=0z=0.

Summable probabilities of zero-free disks

Lemma 9 (Strict circular gain). For c∈Cc\in\mathbb{C} and η>0\eta>0,

γ(c,η)=12π∫02π∣c+ηeiθ∣ dθ−∣c∣>0.\gamma(c,\eta)=\frac{1}{2\pi}\int_{0}^{2\pi}|c+\eta e^{i\theta}|\,d\theta-|c|>0.

Proof. Pair the angles θ\theta and θ+π\theta+\pi. The triangle inequality gives ∣c+ηeiθ∣+∣c−ηeiθ∣≥2∣c∣|c+\eta e^{i\theta}|+|c-\eta e^{i\theta}|\ge2|c|. For c≠0c\ne0 the inequality is strict on an interval of angles near a direction perpendicular to cc. For c=0c=0 the assertion is immediate. Integrating proves strict positivity.

Fix c≠0c\ne0 and 0<R<∣c∣0<R<|c|, put η=R/2\eta=R/2 and H=∣c∣+ηH=|c|+\eta, and let Hn(c,R)\mathcal{H}_n(c,R) be the event that GnG_n has no zero in B(c,R)B(c,R). Define the nonnegative random variable

Yn=12π∫02πexp⁡(Wn∣c+ηeiθ∣)∣Gn(c+ηeiθ)∣ dθ.(14)Y_n=\frac{1}{2\pi}\int_{0}^{2\pi}\frac{\exp(W_n|c+\eta e^{i\theta}|)}{|G_n(c+\eta e^{i\theta})|}\,d\theta. \tag*{(14)}

Tonelli’s theorem and (13) imply

EYn≤2e2H+Wn.(15)\mathbb{E}Y_n\le2e^{2H_+W_n}. \tag*{(15)}

On Hn(c,R)\mathcal{H}_n(c,R), log⁡∣Gn∣\log|G_n| is harmonic throughout B(c,R)B(c,R), and therefore

12π∫02πlog⁡∣Gn(c+ηeiθ)∣ dθ=log⁡∣Gn(c)∣.(16)\frac{1}{2\pi}\int_{0}^{2\pi}\log|G_n(c+\eta e^{i\theta})|\,d\theta=\log|G_n(c)|. \tag*{(16)}

Jensen’s inequality for the exponential, applied to (14), and (16) give

log⁡Yn≥Wn(∣c∣+γ(c,η))−log⁡∣Gn(c)∣.\log Y_n\ge W_n(|c|+\gamma(c,\eta))-\log|G_n(c)|.

For all sufficiently large nn, Lemma 5 bounds the last term. Thus, on Hn(c,R)\mathcal{H}_n(c,R),

Yn≥12exp⁡(γ(c,η)Wn−∣c∣(W2n−Wn)).(17)Y_n\ge\frac{1}{2}\exp\bigl(\gamma(c,\eta)W_n-|c|(W_{2n}-W_n)\bigr). \tag*{(17)}

Markov’s inequality, (15), and (17) now yield

P(Hn(c,R))≤4e2H+Wnexp⁡(−γ(c,η)Wn+∣c∣(W2n−Wn))(18)\mathbb{P}(\mathcal{H}_n(c,R))\le4e^{2H_+W_n}\exp\bigl(-\gamma(c,\eta)W_n+|c|(W_{2n}-W_n)\bigr) \tag*{(18)}
≤4e2H+Wnexp⁡(−12γ(c,η)Wn)(19)\le4e^{2H_+W_n}\exp\bigl(-\tfrac{1}{2}\gamma(c,\eta)W_n\bigr) \tag*{(19)}

for all sufficiently large nn. The second inequality uses W2n/Wn→1W_{2n}/W_n\to1 and γ(c,η)>0\gamma(c,\eta)>0.

For every a>0a>0, (7) implies

∑n=1∞Wne−aWn<∞.(20)\sum_{n=1}^{\infty} W_n e^{-aW_n}<\infty. \tag*{(20)}

For example, eventually Wn≤nW_n\le\sqrt{n} and aWn≥4log⁡naW_n\ge4\log n, so the summand is at most n−7/2n^{-7/2}. Consequently the probabilities in (19) are summable. The first Borel–Cantelli lemma proves that, almost surely, only finitely many GnG_n are zero-free in this fixed disk. No independence between different derivative orders is used.

For completeness the events above are measurable. Normal convergence makes (ω,z)↦Gn(ω,z)(\omega,z)\mapsto G_n(\omega,z) jointly continuous when the sample space is the countable product of closed unit disks with its product topology. Zero-freeness on an open disk equals positivity of the minimum modulus on each of a countable increasing sequence of closed subdisks. Each minimum is a measurable function; it can also be computed as an infimum over a countable dense set. The nonnegative integrand in (14) is measurable, so Tonelli applies without any integrability assumption in advance.

Completion of Theorem 2. Take the countable family

B={B(c,R):c∈Q+iQ, R∈Q>0, R<∣c∣}.\mathcal{B}=\{B(c,R):c\in\mathbb{Q}+i\mathbb{Q},\ R\in\mathbb{Q}_{>0},\ R<|c|\}.

Every nonempty open subset of C\mathbb{C} contains a member of B\mathcal{B}: choose a nonzero point in it, then a nearby rational complex center and a sufficiently small rational radius. The preceding Borel–Cantelli conclusion holds simultaneously for all disks in B\mathcal{B} outside a countable union of null events. Intersect with the probability-one event on which all coefficients are nonzero. Choose a sample in this intersection. The resulting deterministic transcendental entire ff has (C); the quantifier lemma gives the original subsequence formulation as well.

Lemma 4 implies (1), since w(⌈r2⌉)=O(log⁡rlog⁡log⁡r)w(\lceil r^2\rceil)=O(\log r\log\log r). Thus ρ(f)≤1\rho(f)\le1. Every entire function of order less than one has finite exponential type: choose ρ(f)<q<1\rho(f)<q<1, then eventually log⁡Mf(r)≤rq≤r\log M_f(r)\le r^q\le r, and adjust a constant on the remaining compact disk. Proposition 3 therefore forces ρ(f)≥1\rho(f)\ge1. It also rules out finite exponential type for our chosen function. This proves all assertions of the theorem.

A further check: limiting zero density

The same construction provides a quantitative version. This section is not needed for the existence proof above. Let νn\nu_n be the locally finite measure that counts the zeros of f(n)f^{(n)}, with multiplicities.

Theorem 10. For almost every sample in (6),

νnWn⟶12π∣z∣ dA(z)\frac{\nu_n}{W_n}\longrightarrow\frac{1}{2\pi|z|}\,dA(z)

locally weakly as n→∞n\to\infty. The limiting measure has no atom at zero; its displayed density is locally integrable. In particular, for every fixed r>0r>0,

νn(B(0,r))∼rWn.\nu_n(B(0,r))\sim rW_n.

Proof. For a fixed H>0H>0, write

Vn(z)=(Wn∣z∣−log⁡∣Gn(z)∣)+,Qn=1πH2∫∣z∣<HVn(z) dA(z).(21)V_n(z)=(W_n|z|-\log|G_n(z)|)_+,\qquad Q_n=\frac{1}{\pi H^2}\int_{|z|<H}V_n(z)\,dA(z). \tag*{(21)}

The value +∞+\infty at a zero causes no difficulty. Nonzero analytic functions have locally integrable logarithms. Moreover ex+≤1+exe^{x_+} \le1 + e^x and Lemma 8 give

EeVn(z)≤1+2e2H+Wn(∣z∣≤H).\mathbb{E}e^{V_n(z)} \le1 + 2e^{2H_+W_n}\qquad(|z| \le H).

Jensen’s inequality for normalized area measure and Tonelli’s theorem imply

EeQn≤1+2e2H+Wn.\mathbb{E}e^{Q_n} \le1 + 2e^{2H_+W_n}.

Hence, for every ε>0\varepsilon> 0,

P(Qn≥εWn)≤(1+2e2H+Wn)e−εWn.\mathbb{P}(Q_n \ge\varepsilon W_n) \le(1 + 2e^{2H_+W_n})e^{-\varepsilon W_n}.

These probabilities are summable by (20). Apply Borel–Cantelli for integer H≥1H \ge1 and positive rational ε\varepsilon to obtain, almost surely, Qn/Wn→0Q_n/W_n \to0 on every compact disk. Set un(z)=Wn−1log⁡∣Gn(z)∣u_n(z) = W_n^{-1}\log|G_n(z)|. Lemma 5 gives

(un(z)−∣z∣)+≤H(W2nWn−1)+log⁡2Wn⟶0(u_n(z) - |z|)_+ \le H\left(\frac{W_{2n}}{W_n} - 1\right) + \frac{\log2}{W_n} \longrightarrow0

uniformly for ∣z∣≤H|z| \le H. The opposite part is exactly (∣z∣−un(z))+=Vn(z)/Wn(|z| - u_n(z))_+ = V_n(z)/W_n. Thus

un⟶∣z∣in Lloc1(C)almost surely.(22)u_n \longrightarrow|z| \quad\text{in } L^1_{\mathrm{loc}}(\mathbb{C}) \quad\text{almost surely}. \tag*{(22)}

For a nonzero analytic function hh, the elementary distributional identity

Δlog⁡∣h∣=2π∑h(a)=0mult⁡a(h)δa\Delta\log|h| = 2\pi\sum_{h(a)=0}\operatorname{mult}_a(h)\delta_a

follows from local factorization at each zero and Δlog⁡∣z−a∣=2πδa\Delta\log|z-a| = 2\pi\delta_a; the nonvanishing local factor has harmonic log modulus. Therefore νn/Wn=(2π)−1Δun\nu_n/W_n = (2\pi)^{-1}\Delta u_n. Taking distributional Laplacians in (20) gives

νnWn⟶12πΔ∣z∣=12π∣z∣ dA(z).\frac{\nu_n}{W_n} \longrightarrow\frac{1}{2\pi}\Delta|z| = \frac{1}{2\pi|z|}\,dA(z).

For the last equality, away from zero the radial Laplacian is 1/∣z∣1/|z|. Its flux through a circle of radius ss is 2πs2\pi s, which tends to zero, so there is no delta mass at the origin.

Distributional convergence of these positive measures implies local weak convergence: a smooth compactly supported upper cutoff bounds their masses on any fixed compact set, and continuous test functions can then be uniformly approximated by smooth ones. Each disk boundary has zero limiting measure, so the masses of any fixed disk converge to the mass prescribed by the limit. Polar integration gives ∫B(0,r)(2π∣z∣)−1 dA=r\int_{B(0,r)} (2\pi|z|)^{-1}\,dA = r.

General weights and verification notes

The proof uses no special identity for (5). It works for any nondecreasing sequence Wn≥1W_n \ge1 with

W2nWn⟶1,Wnlog⁡n⟶∞,Wnn⟶0,\frac{W_{2n}}{W_n} \longrightarrow1,\qquad\frac{W_n}{\log n} \longrightarrow\infty,\qquad\frac{W_n}{\sqrt{n}} \longrightarrow0,

using bn=∏j=1nWjb_n = \prod_{j=1}^{n} W_j. The deterministic bound becomes log⁡Mf(r)≤rW⌈r2⌉+log⁡2\log M_f(r) \le rW_{\lceil r^2\rceil} + \log2 eventually, and the limiting zero measure has normalization WnW_n. For example, slowly varying factors growing more slowly than log⁡log⁡n\log\log n can replace that factor in (5), as long as the three displayed conditions hold. The exact least possible growth rate inside the order-one infinite-type class is not determined by this argument.

The main proof is analytic and has no numerical assumptions. An accompanying Python script checks the coefficient peak inequality and the derivative majorant at selected parameter values using high precision arithmetic. Those checks are sanity checks, not substitutes for the proof and not a Lean formalization.

AI assistance and authorship disclosure

Jensen Kohlmeyer (GitHub: JENW1N) and Liam Kruer (GitHub: lkruer) are the designated authors of this AI-assisted research manuscript. OpenAI Codex assisted with the mathematical construction, proof development, coefficient checks, literature comparison, and manuscript preparation. The public submission is made from the JENW1N account. No independent human verification or distinct individual author roles are asserted by this disclosure; attribution and correctness are submitted for review.

The principal points for independent review are the conditional inverse moment, the uniform tail estimate for all k>nk > n, and the summability for every positive disk-dependent constant. The full proof above addresses the original infinite-subsequence quantifier. It has no unproved mathematical premise. First-solution credit is expressly not claimed; novelty of the sharper growth statements remains unverified.

References

  1. [1]The Justin Sun Prize, problem JSP-000752, catalogue snapshot read October 4, 2026. https://github.com/TheJustinSunPrize/awards/blob/main/problems/catalog-0701–0800.md
  2. [2]The Justin Sun Prize, contribution guidelines, snapshot read October 4, 2026. https://github.com/TheJustinSunPrize/awards/blob/main/CONTRIBUTING.md
  3. [3]T. F. Bloom, Erdős Problem 906 and its discussion. https://www.erdosproblems.com/906 https://www.erdosproblems.com/forum/thread/906?embed=1
  4. [4]E. Hou, Cofinite Zeros of High Derivatives, arXiv:2607.20816, 2026. Version 3 and the author’s current repository manuscript were inspected for this draft. https://arxiv.org/html/2607.20816v3 https://github.com/erichou1/cofinite-derivative-zeros

Paper details

vlnc:W_gy4wzm35
Version
v1 · latest
Published
Fresh — published within the last 30 days
License
CC BY 4.0
Usage

views · 10

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Codex

View statement
Looking for related papers…
Contents