Graphs that are not complete pluripolar (Q4804101)

A subset \(E\) of \(\mathbb C^n\) is said to be pluripolar if there exists \(u\in \text{PSH}({\mathbb C}^n)\), \(u\not \equiv -\infty\), such that \(E\subseteq u^{-1}(-\infty)\) and it is said to be complete pluripolar if \(u\) can be chosen such that \(E=u^{-1}(-\infty)\). The authors are dealing with the question whether graphs \(\Gamma_f=\{(z,f(z)\); \(z\in D\}\) of holomorphic functions defined in domains \(D\) of \(\mathbb C\) are pluripolar. In [Indiana Univ. Math. J. 41, 515-532 (1992; Zbl 0763.32010)], \textit{N. Levenberg, G. Martin}, and \textit{E. A. Poletsky} state the following conjecture: If \(f\) is not analytically continuable across any boundary point of the unit disc \(\mathbb D\), then \(\Gamma_f\) is complete pluripolar. In the present paper the authors show that this conjecture is false by proving the following theorem: If \(D_1\subset D_2\) are domains in \(\mathbb C\) and \(D_2\setminus D_1\) has a density point in \(D_2\), then there exists a holomorphic function \(f\) with domain of existence \(D_1\) such that the graph \(\Gamma_f\) of \(f\) over \(D_1\) is not complete pluripolar in \(D_2\times \mathbb C\). The case \(D_1=\mathbb D\) and \(D_2=\mathbb C\) refutes the conjecture.