A family of singular oscillatory integral operators and failure of weak amenability (Q556467): Difference between revisions
From MaRDI portal
Latest revision as of 11:44, 10 June 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | A family of singular oscillatory integral operators and failure of weak amenability |
scientific article |
Statements
A family of singular oscillatory integral operators and failure of weak amenability (English)
0 references
21 June 2005
0 references
A locally compact group is called weakly amenable if its Fourier algebra \(A(G)\) admits an approximate unit which is bounded in the completely bounded operator norm on \(A(G)^*\). The infimum of all such norm bounds is denoted \(\Lambda(G)\), and \(\Lambda(G)\) is defined to be \(\infty\) if no such norm bound exists. Note that \(\Lambda(G)= 1\) when \(G\) is amenable. The problem of which groups are weakly amenable is very important and difficult, and many authors have contributed specific results. Most notably, some twenty years ago \textit{U. Haagerup} [Group \(C^*\)-algebras without the completely bounded approximation property, unpublished manuscript] has shown that \(\Lambda(G)=\infty\) for every noncompact simple Lie group with finite centre and real rank at least two. Significant progress in the subject is achieved by the present paper. Let \(G\) be a connected Lie group with Lie algebra \({\mathfrak g}\). Let \({\mathfrak g}= {\mathfrak s}\oplus{\mathfrak r}\) be its Levi decomposition and \({\mathfrak s}= {\mathfrak s}_1\oplus\cdots\oplus{\mathfrak s}_m\) the decomposition of \({\mathfrak s}\) as a direct sum of simple ideals. Let \(R\), \(S\) and \(S_i\) be the analytic subgroups of \(G\) corresponding to \({\mathfrak r}\), \({\mathfrak s}\) and \({\mathfrak s}_i\), \(i= 1,\dots,m\), respectively, and suppose that \(S\) has finite centre. In this situation, it is shown that \(G\) is weakly amenable if and only if for each \(i= 1,\dots, m\), either \(S_i\) is compact or \(S_i\) is noncompact and of real rank one and \({\mathfrak s}_i\) centralizes \({\mathfrak r}\). Moreover, if these conditions are satisfied, then \(\Lambda(G)=\prod^{m}_{i=1}\Lambda(S_i)\) and \(\Lambda(G)\) can be computed since \(\Lambda(H)\) is already known for all real rank one simple Lie groups with finite centre. The proof of this deep theorem is extremely involved. The first step is to show that after passing to a finite covering group of \(G\), one can assume that \(S\) and \(S_i\) are closed in \(G\) and that \(G\) is the semidirect product of \(S\) and \(R\), \(G:= S\ltimes R\). Next, supposing that some \(S_i\) is noncompact and does not centralize \(R\), it is shown that \(G\) contains closed subgroups \(K\) and \(L\) such that \(K\subseteq L\), \(K\) is normal in \(L\) and \(L/K\) (or a double cover of \(L/K\)) is isomorphic to a group belonging to one of the following two families: (1) \(G_n= \text{SL}(2,\mathbb{R})\ltimes H_n\) \((n\geq 1)\), where \(H_n\) is the \((2n+ 1)\)-dimensional Heisenberg group and \(\text{SL}(2,\mathbb{R})\) acts on \(H_n\) by fixing the centre and operating on \(\mathbb{R}^{2n}\) by the unique irreducible representation of dimension \(2n\). (2) \(F_n= \text{SL}(2,\mathbb{R})\ltimes\mathbb{R}^n\) \((n\geq 2)\), the semidirect product formed with the action of the unique irreducible representation of \(\text{SL}(2,\mathbb{R})\) on \(\mathbb{R}^n\). Since \(\Lambda(G)\geq\Lambda(L/K)\), it suffices to show that \(\Lambda(G_n)= \infty\) for all \(n\geq 1\) and \(\Lambda(F_n)=\infty\) for all \(n\geq 2\). In the manuscript mentioned above, Haagerup proved that \(\Lambda(G_1)= \Lambda(F_2)=\infty\). Later \textit{B. Dorofaeff} [Math. Ann. 297, No. 4, 707--724 (1993; Zbl 0791.43004)] showed that \(\Lambda(F_n)= \infty\) for all \(n\geq 2\). In the paper under review, using a modification of Haagerup's approach for \(G_1\), the authors show that matters can be reduced to the estimation of certain singular oscillatory integral operators on \(\mathbb{R}^2\). However, obtaining the required \(L^2\)-bounds for these operators is rather intricate and occupies most of the paper.
0 references
connected Lie group
0 references
semidirect product
0 references
Fourier algebra
0 references
weak amenability
0 references
singular oscillatory integral operators
0 references
\(L^2\)-bounds
0 references
0 references
0 references
0 references