A generalisation of the Bank-Laine property (Q2378578)

In this paper, the authors determine those entire functions~\(f\) such that \(f(z) \in A\) implies \(f^{(k)}(z) \in B\) for given \(k \in \mathbb{N}\) and two finite sets \(A, B \subset\mathbb{C}\) satisfying certain appropriate conditions. When \(k=1\), \(A=\{0\}\), and \(B=\{\pm 1\}\), such a function is called a \textit{Bank-Laine function} associated with a celebrated conjecture given by~\textit{S. B.~Bank} and \textit{I.~Laine} [Trans. Am. Math. Soc. 273, 351--363 (1982; Zbl 0505.34026)] about solutions of second order homogeneous linear differential equations. Although there are known various such functions like \(\sin z\), the classification of all Bank-Laine functions is still in the process of research and development. For the purpose to determine composite Bank-Laine functions, the first author of the present paper studied in his previous paper [Trans. Am. Math. Soc. 354, No. 3, 1177--1191 (2002; Zbl 0992.30019)] the ``double Bank-Laine property'', where two pairs of the sets \(A\) and \(B\) with \(\#A=1\) and \(\#B=2\) are considered when \(k=1\). Then he succeeded in detecting the entire functions with this property. Motivated by the spirit of this result, the authors of the paper under review study a more general question: \textit{What can be said if a transcendental meromorphic function \(f\) in the plane satisfies \[ f(z) \in A := \{ a_1, \ldots , a_m \} \quad \text{ implies that } \quad f^{(k)}(z) \in B := \{ b_1, \ldots , b_n \}, \tag{1} \] for some \(k, m, n \in \mathbb{N}\), and distinct \(a_j\), \(b_j\) in~\(\mathbb{C}\)?} It is remarked in Lemma 2.1 that Nevanlinna's second fundamental theorem implies that \(m \leq 2 + n(k + 1)\), and \(m \leq n + 1\) if \(f\) is entire. The authors give examples of suitable finite sets \(A\) and \(B\) with an elliptic function~\(f\) for which \(m = 2 + n(k + 1)\) holds when \((k,m,n)=(1,4,1), (3,6,1)\). Further, it is noted that the equality \(m=n+1\) in the latter is always possible for the entire function \(f(z)=a_1+e^{az+b}\). Then, by Theorem~1.2, they confirm that this is the case if \(n=m-1\geq 2\), \(k=1\), and the \(b_j\) are non-zero. Finally, Theorem~1.3 is given to show that there are \textit{no transcendental} entire functions of finite order having the property~(1) for \((k,m,n)=(2,2,1)\): \textit{Let \(a_1, a_2\in \mathbb{C}\) be distinct, and let \(f\) be an entire function of finite order with the property that \[ f(z) \in \{ a_1, a_2 \} \quad \text{ implies that } \quad f^{\prime\prime}(z) =0. \] Then \(f(z)=az+b\) with \(a,b\in\mathbb{C}\).} The proof of Theorem~1.3 requires the Pang-Zalcman rescaling lemma on a family of functions meromorphic in the plane domain, which is due to \textit{H.-H.~Chen} and \textit{Y.-X.~Gu} [Sci. China, Ser. A 36, No. 6, 674--681 (1993; Zbl 0777.30018)] together with the so-called Wiman-Valiron theory and a result due to \textit{J. Clunie} and \textit{W. K.~Hayman} [Comment. Math. Helv. 40, 117--148 (1966; Zbl 0142.04303)] on spherical derivatives. It is left open in the paper whether the hypothesis that~\(f\) has finite order is really needed in Theorem~1.3. To show Theorem~1.2, the authors skillfully use some generalizations of estimates by the first author in the paper cited above, standard estimates for logarithmic derivatives, and a result due to \textit{P.~Lappan} [Comment. Math. Helv. 49, 492--495 (1974; Zbl 0292.30029)], which is also proved by using the Pang-Zalcman rescaling lemma.