Talk:Affiliated operator

Incorrect statements on wikipedia page
Unfortunately various statements on this page are incorrect. The easiest way to define affiliation is to say that T commutes with each unitary in the commutant and use the uniqueness of the polar decomposition. Only in the type I and type II case, in the presence of a finite trace, does the set of affiliated operators form a *-algebra. This was first proved by Murray and von Neumann. By defining measurable operators, Edward Nelson simplified their work and extended it to infinite traces; his approach is discussed in Takesaki's book. In the type III case the different non-commutative Lp spaces correspond to different powers of the reference modular operator and so are disjoint: this is Connes' spatial theory of non-commutative integration which is discussed in his book on non-commutative geometry. I will therefore shorten and completely rewrite this article. --Mathsci 11:05, 8 July 2007 (UTC)

What is M2(M)?
The page pleasantly defines almost all required notions. It would be nice to also say what is M2(M) in the definition. Unfortunately I cannot figure this out by myself. 109.172.129.12 (talk) 14:13, 4 March 2019 (UTC)