On the facial structure of the unit ball in the dual space of a JB\(^*\)-triple (Q711578)

Some twenty-five years ago, \textit{C. M. Edwards} and \textit{G. T. Rüttimann} [J. Lond. Math. Soc., II. Ser. 38, No.~2, 317--332 (1988; Zbl 0621.46043)] made a study of the facial structure of the unit ball in various complex Banach spaces. They discovered that the problem became tractable when the open unit ball in the Banach space in question was symmetric, or, equivalently, when the Banach space \(A\) in question was a JB\(^*\)-triple. They discovered that, when \(A\) was a Banach dual space, that is to say, when \(A\) is a JBW\(^*\)-triple, there is an order isomorphism \[ u \mapsto \{u\}_{\prime} = \{x \in A_{*1}: x(a) = 1\} \] between the complete lattice \(\tilde \mathcal U(A)\) of tripotents in \(A\) and the complete lattice of norm-closed faces of the unit ball \(A_{*1}\) in the necessarily unique predual \(A_*\) of \(A\) and an order anti-isomorphism \[ u \mapsto \{u\}_{\prime}^{\prime} = u + A_0(u)_1 \] from \(\tilde \mathcal U(A)\) onto the complete lattice of weak\(^*\)-closed faces of the unit ball \(A_1\) in \(A\). In this case, \(A_0(u)\) is the algebraic Peirce-zero space corresponding to the tripotent \(u\). The so-called `face-problem', consisting of giving an algebraic description of the geometric facial property, was therefore solved for norm-closed faces of the unit ball in the pre-dual of a JBW\(^*\)-triple \(A\) and for weak\(^*\)-closed faces of the unit ball in \(A\) itself. In 1992, \textit{C. A. Akemann} and \textit{G. K. Pedersen} [Proc. Lond. Math. Soc., III. Ser. 64, No. 2, 418--448 (1992; Zbl 0759.46050)], unaware of the Edwards-Rüttimann results, re-proved these for the special case of a W\(^*\)-algebra, which is itself a JBW\(^*\)-triple with respect to the triple product \[ \{a\,\,b\,\,c\} = 1/2(ab^*c + cb^*a). \] However, they went further by solving the `face problem' for the norm-closed faces of the unit ball in a C\(^*\)-algebra \(A\) and the weak\(^*\)-closed faces of the unit ball in its dual \(A^*\). They showed that there was an order isomorphism \(u \mapsto \{u\}_{\prime}\) from the complete atomic lattice \(\tilde{\mathcal U}_c(A)\) of what were later to become known as compact tripotents in the W\(^*\)-algebra \(A^{**}\) onto the complete lattice of weak\(^*\)-closed faces of the unit ball \(A^*_1\) in \(A^*\) and an order anti-isomorphism from \(u \mapsto \{u\}_{\prime \prime}\) from \(\tilde{\mathcal U}_c(A)\) onto the complete lattice of norm-closed faces of \(A_1\). Since the second dual \(A^{**}\) of a JB\(^*\)-triple is a JBW\(^*\)-triple, it was conjectured that these results also held for JB\(^*\)-triples. However, the methods used by Akemann and Pedersen depended crucially upon the notion of global positivity for elements of a C\(^*\)-algebra and the fact that the \(2 \times 2\) matrices over a C\(^*\)-algebra is itself a C\(^*\)-algebra, neither of which hold for JB\(^*\)-triples. Various attacks were made on this and related problems in the following years, including one by \textit{C. M. Edwards} and \textit{G. T. Rüttimann} [Math. Proc. Camb. Philos. Soc. 120, No.~1, 155--173 (1996; Zbl 0853.46070)] that led to the definition of compact tripotents and another giving an alternative definition of compactness by the authors [J. Lond. Math. Soc., II. Ser. 74, No.~1, 75--92 (2006; Zbl 1108.46049)]. In [J. Reine Angew. Math. 641, 123--144 (2010; Zbl 1204.46039)], the authors of this paper in collaboration with \textit{C. M. Edwards} and \textit{C. S. Hoskin} were able to use the techniques developed by the authors themselves to show that the second conjecture was true for JB\(^*\)-triples. The current paper that uses completely different and highly inventive analytic techniques solves the `face problem' for the complete lattice of weak\(^*\)-closed faces of the unit ball in the dual of a JB\(^*\)-triple and resolves a well-known problem the answer to which has been open for many years, and, therefore, represents a major contribution to the field.