On orthogonal systems, two-sided bases, and regular subfactors (Q2194521)

Vaughan Jones initiated the modern theory of subfactors in his seminar paper [\textit{V. F. R. Jones}, Invent. Math. 72, 1--25 (1983; Zbl 0508.46040)]. Pimsner and Popa introduced the notion of left Pimsner-Popa basis and right Pimsner-Popa basis in 1986 [\textit{M. Pimsner} and \textit{S. Popa}, Ann. Sci. Éc. Norm. Supér. (4) 19, No. 1, 57--106 (1986; Zbl 0646.46057)]. Let \(N \subset M\) be a unital inclusion of von Neumann algebras equipped with a faithful normal conditional expectation \(E\) from \(M\) onto \(N\). Then, a finite set \(B := \{\lambda_1, \dots, \lambda_n \}\subset M\) is called a left Pimsner-Popa basis for \(M\) over \( N\) via \(E\) if every \(x\in M\) can be expressed as \(x=\sum_{i=1}^{n} E(x \lambda_i)\lambda_i \). The definition of a right Pimsner-Popa basis is similar. \(B\) is said to be a two-sided basis if it is simultaneously a left and a right Pimsner-Popa basis. Pimsner and Popa [loc. cit.] showed that if \(N \subset M \) is a subfactor of type II\(_1\) with finite Jones index, then there always exists a left (equivalently, a right) Pimsner-Popa basis for \(M\) over \( N\) via the unique trace preserving conditional expectation \(E_N : M \rightarrow N\). But it is known that non-extremal subfactors do not admit two-sided bases. It is natural to ask whether there always exists a two-sided basis for every finite index extremal subfactor or not. In this article, the authors answer this question in the affirmative for all regular subfactors of type II\(_1\) with finite Jones index (without assuming extremality) as follows. {Theorem 3.10.} Let \(N \subset M\) be a regular subfactor of type II\(_1\) with finite Jones index. Then \( M\) admits a two-sided basis over \( N\). As a consequence, the authors show that every finite index regular subfactor of type II\(_1\) is extremal.