A Galois theory for finite factors
A Galois theory for finite factors
复制标题
有限因素的伽罗瓦理论
DOI:
10.3792/pja/1195524026
复制
发表时间:
1960
期刊:
影响因子:
--
通讯作者:
Z. Takeda
中科院分区:
文献类型:
--
作者:
M. Nakamura;Z. Takeda
According to a closed analogy between theories of classical simple algebras and continuous finite factors, it is natural to ask that a continuous finite factor obeys a kind of Galois theory. Literally, it is known that I. M. Singer [4 gave an attempt in this direction. This note will present a trial towards it in the following THEOREM. If A is a continuous finite factor acting standardly on a separable Hilbert space H, if G is a finite group of outer automorphisms of A, if B is the subfactor of A consisting of all elements invariant under G, and if moreover the commutor B’ of B is finite. Then, the lattices of all subgroups of G and of all intermediate subfactors between B to A are dually isomorphic under the Galois correspondence which carries a subgroup F to an intermediate subfactor C invariant under F in element-wise. It is expected that the assumption on B’ is provable from the finiteness of G for which the authors hope to discuss in the next occasion. It is also to be remarked that the continuity assumption on A in the theorem is superfluous since a discrete finite factor has no non-trivial group of outer automorphisms. 1o Since A acts standardly on H, there is a unitary u for any g such as ( 1 ) =uxu*, where x means the action of g on x eA. Throughout the remainder, for the sake of convenience, it is to be assumed that the correspondence g -->u satisfies (2) u_=u*. It is to be noticed that u belongs to B’, since x--x--uxu* by the assumption. LEMMA 1. By (1), g gives an outer automorphism on A’. If xeA’, then for any aeA, (1) and (2) imply ax auxu uaqxu uxa-lUg UgXU ( X a which shows that g conserves A’. Hence (1) gives an automorphism on A’. If it is inner, then there is a unitary wA’ such that xq--w*xw or uxu =w*xw for any xeA’, whence wuqx=xwuq for any xeA’, that is, wu commutes with every element of A’. Hence the unitary operator w’=wu belongs to A. Therefore, by we A’,