Real operator algebras and real completely isometric theory (Q743469)

In operator algebras theory it is usually assumed that the Hilbert spaces be complex. This implies a series of handy results, e.g., a positive operator is also selfadjoint. The case of real \(C^*\)-algebras and real von Neumann algebras has been considered in the past see, e.g. \textit{K. R. Goodearl} [Notes on real and complex \(C^*\)-algebras. Cheshire, England: Shiva Publishing Limited (1982; Zbl 0495.46039)], \textit{B. Li} [Real operator algebras. Book manuscript (2002)], and \textit{E. Størmer} [Duke Math. J. 47, 145--153 (1980; Zbl 0462.46044)]. Recently, \textit{Z. J. Ruan} moved forward to the development of real operator spaces theory [Acta Math. Sin., Engl. Ser. 19, No. 3, 485--496 (2003; Zbl 1058.46037); Ill. J. Math. 47, No. 4, 1047--1062 (2003; Zbl 1055.46040)]. The paper under review continues with, and builds on this program. The author shows that several fundamental results on complex operator spaces still hold for real operator spaces. To the previous work of Ruan [Zbl 1058.46037; Zbl 1055.46040] on the characterisation of real operator spaces, Stinespring's theorem and Arveson's extension theorem, the author adds the Blecher-Ruan-Sinclair characterisation theorem, Meyer's theorem, and the existence of the injective envelope, of the \(C^*\)-envelope, and of the Shilov boundary for a real operator space. Further results towards the development of the theory of one-sided \(M\)-ideals are introduced. We remark here several difficulties when developing the theory of real operator spaces. One is that positivity of an element does not imply that the element is selfadjoint. Therefore the role of ``positive'' mappings is now played by those that are both selfajoint and positive. Secondly, several facts from Banach space theory do not pass over. For example, \((X^*)_r\) is not completely isometrically isomorphic to \((X_r)^*\), where \(X_r\) denotes \(X\) considered as a real operator space. The key element in the paper under review is to use the complexification of a space. In particular, Ruan [Zbl 1055.46040] showed that, given a real Hilbert space \(H\), then the complexification \(B(H)_c\) of \(B(H)\) is completely isometrically isomorphic to \(B(H_c)\), for \(H_c = H + i H\) (the sum taken as a direct sum of linear spaces). Moreover, the complexification \(B(H)_c = B(H) + i B(H)\) (as a direct sum of linear spaces) can be identified (completely isometrically) with a subspace of \(M_2(B(H))\) in the following way: \[ x + iy \mapsto \begin{bmatrix} x & -y \\ y & x \end{bmatrix}. \] Then a complexification \(V_c\) of a real operator space \(V\) can be seen to be \(V + iV\) inside \(B(H)_c\) with the inherited operator space norm structure. The arguments in the paper under review build on reducing the problem from the real case to the complex case, and then on retrieving back the desired result.