Quasi-invariant states for the action of compact groups
Maria Elena Griseta · 2025
Read Quasi-invariant states for the action of compact groups by Maria Elena Griseta (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-20931.txt.
Introduction
Maria Elena Griseta studies quasi-invariant states for compact group actions in a 2025 note. A state in her frame is a weight rule on a C-star algebra, bent by a group act in a set way. Central support marks the case she treats. When the group act and the state's own time flow commute, a clean result lands. The built form of her state then matches that of a fully invariant state. The note is short and proof-led, aimed at operator algebra readers.
Catalog record
| CREATOR | Maria Elena Griseta |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 23 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-20931
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Quasi-invariant states for the action of compact groups Maria Elena Griseta Abstract . We analyze a natural C ∗ C^{*} -algebraic definition of G G -quasi-invariant states for the automorphic action of a compact group G G . We prove that, given a G G -quasi-invariant state with central support, when the action of the group G G commutes with the modular group, its GNS representation is equivalent to that of a G G -invariant state. Mathematics Subject Classification : 46L05, 46L30, 37N20. Key words : C ∗ C^{*} -dynamical systems, quasi-invariant states, invariant states, modular group. 1. Introduction States invariant under the action of a group G through * -automorphism on a C^{*} -algebra {\mathfrak{A}} are very often considered in various fields of mathematics and physics [ 2 , 7 ] . A larger class of states is associated with quasi-invariant measures, which are an important topic in ergodic theory and for example the natural setting for the construction of type III factors [ 10 , 9 ] . More in details, if (X,{\mathcal{B}}) is a measurable space and G denotes a group of automorphisms acting on it by x\in X\mapsto\alpha_{g}(x)\in X\,,\quad g\in G\,, a measure \mu on (X,{\mathcal{B}}) is said to be G -quasi-invariant if for any g\in G the transformed measure \mu\circ\alpha_{g} is equivalent to the measure \mu , that is, these measures are absolutely continuous with respect to each other. Recall also that, given \mu and \nu two regular Borel measures on the compact Hausdorff space X , the multiplication representations associated to them are equivalent if and only if \mu and \nu are absolutely continuous with respect to each other. Therefore, in this paper, given G a compact group, by G -quasi-invariant state we shall mean a state \omega on {\mathfrak{A}} such that, for every g\in G , one has that \pi_{\omega\circ\alpha_{g}} and \pi_{\omega} are unitarily equivalent representations. Our main goal is to prove that, under suitable assumptions, any quasi-invariant state is ”absolutely continuous” with respect to an invariant state. We should note that there does not seem to be a generally agreed-upon definition of what a quasi-invariant state should be in noncommutative settings. For example, in [ 6 ] quasi-invariance was defined in terms of quasi-equivalence of the representations \pi_{\omega\circ\alpha_{g}} and \pi_{\omega} . Recently, a different approach was given in [ 1 ] . Here, given G a group of * -automorphisms of a * -algebra {\mathcal{A}} , and assuming that for every g\in G there exists the Radon-Nikodym derivative x_{g}\in{\mathcal{A}} , the authors define a state \varphi on {\mathcal{A}} to be G -quasi-invariant if (1.1) \varphi(g(a))=\varphi(x_{g}a)\,,\quad a\in{\mathcal{A}}\,. In the paper mentioned above, the authors assume the existence of the Radon-Nikodym derivative. In these notes, however, we deduce the existence of this positive operator after proving some results. Successively, in [ 4 ] , several properties of G -quasi-invariant states defined in ( 1.1 ) were studied in depth. Finally, Dharhi and Ricard, in [ 5 ] , characterize G -quasi-invariant states with uniformly bounded cocycle x_{g} , in terms of invariant normal faithful states on the C^{*} -algebra {\mathcal{A}} . In these note, we show that, given a G -quasi-invariant state \omega with central support, when the action of the group G commutes with the modular group, its GNS representation is equivalent to that associated with the composition of \omega with the canonical conditional expectation, obtained by averaging the action of the group. To reach this result, we first prove some technical lemmas. 2. G -quasi-invariant states We consider the action of a compact group G through automorphisms of a C^{*} -algebra {\mathfrak{A}} . In other words, we are given a group homomorphism \alpha:G\rightarrow\textrm{Aut}({\mathfrak{A}}) , i.e. \alpha_{gh}=\alpha_{g}\circ\alpha_{h} for all g,h\in G . We also need to make some assumptions on the continuity of \alpha . As usually done in the literature, we will work under the hypothesis that \alpha is pointwise norm continuous, namely that for every a\in{\mathfrak{A}} the vector-valued function G\ni g\mapsto\alpha_{g}(a)\in{\mathfrak{A}} is continuous, where {\mathfrak{A}} is thought of as being given with its norm topology. We recall that a state \omega on {\mathfrak{A}} is said G -invariant if for every g\in G one has \omega\circ\alpha_{g}=\omega . Throughout this paper by G -quasi-invariant state we shall mean a state \omega on {\mathfrak{A}} such that, for every g\in G , one has that \pi_{\omega\circ\alpha_{g}} and \pi_{\omega} are unitarily equivalent representation. As is known, the hypotheses we are working under allow us to define a (faithful) condition expectation, E_{G} , from {\mathfrak{A}} onto {\mathfrak{A}}^{G}:=\{a\in{\mathfrak{A}}:\alpha_{g}(a)=a\,\,\textrm{for all}\,g\in G\} . This is obtained by averaging the action of the group, that is E_{G}(a):=\int_{G}\alpha_{g}(a)\textrm{d}g\,,a\in{\mathfrak{A}}\,. Let us \omega be a G -quasi-invariant state. We will always assume that \omega has central support. As is well known, this amounts to asking that the GNS vector \xi_{\omega} is separating for the von Neumann algebra \mathcal{R}_{\omega}:=\pi_{\omega}({\mathfrak{A}})^{\prime\prime} . Therefore, the von Neumann algebra \pi_{\omega}({\mathfrak{A}})^{\prime\prime}\subset{\mathcal{B}}({\mathcal{H}}_{\omega}) is in standard form, and we can consider the modular flow associated with the cyclic and separating vector \xi_{\omega} , which we will denote by \{\sigma_{\omega}^{t}:t\in\mathbb{R}\} . Lemma 2.1 . Let (G,{\mathfrak{A}},\alpha) be a C^{*} -dynamical system, where G is a compact group and {\mathfrak{A}} a C^{*} -algebra. For any state \omega on {\mathfrak{A}} , consider the following statements: (i) \pi_{\omega\circ E_{G}}\cong\pi_{\omega} (ii) For every g\in G , \pi_{\omega}\circ\alpha_{g}\cong\pi_{\omega} (iii) \omega is a G-quasi-invariant state Then (i)\Rightarrow(ii)\Rightarrow(iii) . Proof. For the implication (i)\Rightarrow(ii) note that from \pi_{\omega\circ E_{G}}\cong\pi_{\omega} one finds \pi_{\omega\circ E_{G}}\circ\alpha_{g}\cong\pi_{\omega}\circ\alpha_{g} for all g\in G . Since \pi_{\omega\circ E_{G}}\circ\alpha_{g}\cong\pi_{\omega\circ E_{G}\circ\alpha_{g}}=\pi_{\omega\circ E_{G}} and \pi_{\omega}\circ\alpha_{g}\cong\pi_{\omega\circ\alpha_{g}} , we have \pi_{\omega\circ E_{G}}\cong\pi_{\omega\circ\alpha_{g}} for all g\in G . The conclusion now follows by transitivity. The implication (ii)\Rightarrow(iii) is obvious. ∎ Lemma 2.2 . Let (G,{\mathfrak{A}},\alpha) be a C^{*} -dynamical system, where G is a compact group and {\mathfrak{A}} a C^{*} -algebra. If \omega is a G -quasi-invariant state, then the action of G can be lifted to \pi_{\omega}({\mathfrak{A}})^{\prime\prime} through automorphisms \{\widetilde{\alpha_{g}}:g\in G\}\subset{\rm Aut}(\pi_{\omega}({\mathfrak{A}})^{\prime\prime}) such that \widetilde{\alpha_{g}}(\pi_{\omega}(a))=\pi_{\omega}(\alpha_{g}(a))\,,\,\,\,a\in{\mathfrak{A}}. Proof. It is a straightforward consequence of the existence of the interpolating isomorphism between \pi_{\omega}({\mathfrak{A}})^{\prime\prime} and \pi_{\omega\circ\alpha_{g}}({\mathfrak{A}})^{\prime\prime} and the fact that \pi_{\omega\circ\alpha_{g}}\cong\pi_{\omega}\circ\alpha_{g} for every g\in G . ∎ Lemma 2.3 . Under the same hypotheses as the previous lemma and assuming that {\mathfrak{A}} is separable, for any T\in\mathcal{R}_{\omega} and x\in{\mathcal{H}}_{\omega} the vector-valued function G\ni g\mapsto\widetilde{\alpha_{g}}(T)x\in{\mathcal{H}}_{\omega} is Bochner integrable. For T\in\mathcal{R}_{\omega} , define \widetilde{E_{G}}[T]x:=\int_{G}\widetilde{\alpha_{g}}(T)x\,{\rm d}g\,,\,\,x\in{\mathcal{H}}_{\omega}\,, then \widetilde{E_{G}}[T] lies in \mathcal{R}_{\omega} . Furthermore, the map \widetilde{E_{G}}:\mathcal{R}_{\omega}\rightarrow\mathcal{R}_{\omega} is faithful and normal: T\in\mathcal{R}_{\omega} with \widetilde{E_{G}}[T^{*}T]=0 implies T=0 , and \varphi\circ\widetilde{E_{G}} is a normal state for every normal state \varphi on \mathcal{R}_{\omega} . Proof. We start by noting that {\mathcal{H}}_{\omega} is a separable Hilbert space. In particular, if T\in\mathcal{R}_{\omega} , there exist sequence \{a_{n}:n\in\mathbb{N}\}\subset{\mathfrak{A}} such that \pi_{\omega}(a_{n}) strongly converges to T , and \|\pi_{\omega}(a_{n})\|\leq M for every n\in\mathbb{N} . We next define a sequence \{f_{n}:n\in\mathbb{N}\} of vector-valued functions on G by f_{n}(g):=\widetilde{\alpha_{g}}(\pi_{\omega}(a_{n}))x . Note that each f_{n} is continuous (when {\mathcal{H}}_{\omega} is endowed with the norm topology). Further, if we set f(g):=\widetilde{\alpha_{g}}(T)x\in{\mathcal{H}}_{\omega} , we have that f is the pointwise norm limit of the sequence f_{n} , and thus f is measurable. At this point, to ascertain that f is Bochner-integrable, it is enough to check \int_{G}\|f(g)\|\textrm{d}g<\infty , which is obvious as \|f(g)\|\leq\|T\|\|x\| for all g\in G . That \widetilde{E_{G}}[T] sits in