Tangential boundary limits and exceptional sets for holomorphic functions in Dirichlet-type spaces (Q1114821)

We consider classes of holomorphic functions in the unit ball B of \({\mathbb{C}}^ n\) whose boundary values are, in a sense, potentials of \(L^ p\) functions. Specifically, such functions are given by ``fractional Cauchy integrals'' \[ C_{\alpha}f(z)=\int_{S}[f(\zeta)/(1-<z,\zeta >)^{n-\alpha}]d\sigma (\zeta)\quad (z\in B), \] where S is the unit sphere of \({\mathbb{C}}^ n\), \(f\in L^ p(S)\), \(1\leq p\leq \infty\), and \(0<\alpha <n\). It was proved by \textit{M. S. Gowda} [Ph. D. Thesis, Univ. Wisconsin (1982)] that \(C_{\alpha}f\) has boundary limits a.e. within certain tangential approach regions that are wider than A. Korányi's admissible regions. This is in analogy to the real-variable situation treated by \textit{A. Nagel}, \textit{W. Rudin} and \textit{J. Shapiro} [Ann. Math., II. Ser. 116, 331-360 (1982; Zbl 0531.31007)], where they obtained the a.e. existence of tangential boundary limits for harmonic extensions of potentials of \(L^ p\) functions. However, Gowda only proved weak type (p,p) of the maximal operator associated to the approach regions, whereas in the real-variable case, Nagel, Rudin and Shapiro showed the maximal operator to be of strong type (p,p). In this paper it is proved that the maximal operator studied by Gowda is in fact of strong type (p,p). A second related question considered in the paper has to do with the size, in the sense of Hausdorff dimension, of the exceptional sets for the existence of tangential boundary limits of fractional Cauchy integrals \(C_{\alpha}f\). A relationship is obtained between the (nonisotropic) order of tangency of the approach regions and the (nonisotropic) Hausdorff dimension of the corresponding exceptional sets. This is analogous to the real-variable situation studied by \textit{P. Ahern} and \textit{A. Nagel} [Duke Math. J. 53, 359-393 (1986; Zbl 0601.32007)].