Bochner-Schoenberg-Eberlein property for abstract Segal algebras (Q2448581)

Let \({\mathcal A}\) be a commutative Banach algebra without order (i.e., \({\mathcal A}\) has no non-zero annihilators). A bounded complex-valued continuous function \(\sigma\) on the spectrum \(\Delta({\mathcal A})\) of \({\mathcal A}\) is called a BSE-function if there is a constant \(C>0\) such that, for any finite collection \(\phi_1\), \(\phi_2\), \dots, \(\phi_n\) in \(\Delta({\mathcal A})\) and complex numbers \(c_1\), \(c_2\), \dots, \(c_n\), we have \[ \biggl|\sum_{j=1}^n c_j \sigma(\phi_j)\biggr| \leq C\biggl \|\sum_{j=1}^n c_j \phi_j\biggr\|. \] Every element \(T\) in the multiplier algebra \({\mathcal M}({\mathcal A})\) defines, in a canonical way, a continuous function \(\widehat T\) on \(\Delta({\mathcal A})\), and \({\mathcal A}\) is called a BSE-algebra if every BSE-function is of this form. The group algebra \(L^1(G)\) of a locally compact abelian group \(G\) is a BSE-algebra by a result due to S. Bochner, I. J. Schoenberg and W. F. Eberlein, which explains the nomenclature. \textit{J. Inoue} and \textit{S.-E. Takahasi} proved in [Acta Sci. Math. 66, No. 1--2, 257--271 (2000; Zbl 0980.46038)] that any Segal algebra \(S(G)\subset L^1(G)\) is a BSE-algebra if and only if it admits a bounded \(\Delta\)-weak approximate identity (a bounded net that is an approximate identity with respect to the weak topology induced by \(\Delta({\mathcal A})\)). In the paper under review, this result is extended to abstract Segal algebras in BSE-algebras in the following way: an abstract Segal algebra which is a so-called essential ideal in a BSE-algebra is itself a BSE-algebra if and only if it admits a bounded \(\Delta\)-weak approximate identity. The ``only if'' part follows from a result of \textit{S.-E. Takahasi} and \textit{O. Hatori} [Proc. Am. Math. Soc. 110, No. 1, 149--158 (1990; Zbl 0722.46025)]. Finally, the authors apply this result to some concrete examples of abstract Segal algebras.