Higher order hulls in \(H^\infty\). II (Q1589670)

Though the Banach algebra \(H^\infty\) of bounded analytic functions on the open unit disc of the complex plane has been widely studied for some time, there is much we still have to learn about its structure. This paper makes a significant contribution to our understanding by providing a thorough analysis of a class of closed ideals of \(H^\infty\). By way of background, recall that the space of maximal ideals of \(H^\infty\) may be identified with the space of non-zero complex multiplicative linear functionals on \(H^\infty\) and is denoted by \(M(H^\infty)\). Also, each element \(f\) in \(H^\infty\) may be identified with its so-called Gelfand transform, \(\widehat{f}\), defined on \(M(H^\infty)\) by \(\widehat{f}(m)=m(f)\). For an ideal \(I\) in \(H^\infty\), the zero set of \(I\) is defined by \(Z(f)=\{m\in M(H^\infty): f(m)=0\) \(\forall f\in I\}\). For two points \(x\) and \(m\) in \(M(H^\infty)\), the pseudohyperbolic distance between them is defined by \(\rho(x,m)=\sup\{|f(m)|: f\in H^\infty\), \(\|f\|_\infty \leq 1\), \(f(x)=0\}\). The relation \(x\sim m \Leftrightarrow \rho(x,m)<1\) is then an equivalence relation on \(M(H^\infty)\). The equivalence class containing the point \(m\) is called the Gleason part of \(m\), denoted \(P(m)\). If the part \(P(m)\) is not just a singleton, then the point \(m\) is called non-trivial. The set of all non-trivial points in \(M(H^\infty)\) is denoted by \(G\). For \(m\) in \(M(H^\infty)\) and \(f\) in \(H^\infty\) with \(f(m)=0\), the order of the zero of \(f\) at \(m\) is defined by \(\text{ord}(f,m)=\sup\{n\in{\mathbb N}: f=f_1\cdots f_n\), \(f_j\in H^\infty\), \(f_j(m)=0\) for \(j=1,2,\dots ,n\}\). If \(f(m)\neq 0\) we say that \(\text{ord}(f,m)=0\), while, for an ideal \(I\) in \(H^\infty\) we define \(\text{ord}(I,m)=\min\{\text{ord}(f,m):f\in I\}\). The main theorem of the paper (Theorem 2.2) provides a criterion for determining, for certain ideals, when a given function \(f\) is in the ideal. Specifically, if \(I\) is a closed ideal in \(H^\infty\) such that \(Z(I)\subseteq G\), then \(I=\{f\in H^\infty: \text{ord}(f,x)\geq \text{ord}(I,x)\) \(\forall x \in Z(I)\}\). This criterion is useful because it is generally fairly easy to determine the order of the zero of a function at a given point. This theorem leads to a characterization of certain ideals in terms of their higher order zero sets. For \(n\) a natural number or \(\infty\) and \(I\) an ideal in \(H^\infty\), define \(E_n(I) = \{x\in M(H^\infty): \text{ord}(I,x)\geq n \}\). For instance, \(E_1(I)\) is the zero set of \(I\). Then define \(I[E_1(I),\dots,E_N(I)]\) to be the set \(\{f\in H^\infty: \text{ord}(f,x)\geq n\) for every \(x\in E_n(I)\), \(n=1,2,\dots,N\}\). With this notation, Theorem 2.2 can be reformulated to say (Corollary 2.4) that, if \(I\) is an ideal in \(H^\infty\) such that \(Z(I)\subseteq G\), then there exists a natural number \(N\) such that \(E_n(I)=\emptyset\) for \(n\geq N+1\) and \(I=I[E_1(I),\dots,E_N(I)]\). The proof of Theorem 2.2 is fairly involved and relies heavily on the role of interpolating Blaschke products and Carleson-Newman Blaschke products in the ideal theory of \(H^\infty\). The theorem can also be restated to characterize the appropriate ideals in terms of pseudohyperbolic derivatives. The authors compare their results with some examples of Bourgain to show where the hypothesis that \(Z(I)\subseteq G\) is needed. Section 3 of the paper addresses the question of generators for ideals of \(H^\infty\). For an ideal \(I\) with the property that there exists a natural number \(N\) such that \(E_n(I)=\emptyset\) for \(n\geq N+1\), Theorem 3.4 asserts that \(I\) is generated algebraically by Carleson-Newman Blaschke products of order \(N\). Moreover, the higher order zero sets of these generators are close to the higher order zero sets of the ideal. Finally, section 4 of the paper presents a nice result (Theorem 4.1) on finitely generated ideals satisfying \(Z(I)\subseteq G\).