A unified approach for classifying simple nuclear C ∗ C^{\ast} -algebras
Ben Bouwen, James Gabe · 2024
Read A unified approach for classifying simple nuclear C ∗ C^{\ast} -algebras by Ben Bouwen, James Gabe (2024) 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-2412-15968.txt.
Introduction
Bouwen and Gabe give a fresh proof of a landmark classification theorem for a broad family of operator algebras, in this 2024 paper. The Kirchberg-Phillips theorem sorts the wildest of these algebras, the purely infinite ones, by a short list of invariants. Older proofs leaned on heavy machinery built for that purpose alone.
The new proof borrows a framework first built for the tamer, finite side of the family. One approach now covers both worlds. The argument also avoids the classic Geneva theorems entirely. Instead, those theorems fall out as consequences for nuclear algebras. The authors present the unity itself as the advance. Specialists gain a single road map for the whole classification program.
Catalog record
| CREATOR | Ben Bouwen, James Gabe |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 154 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-15968
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A unified approach for classifying simple nuclear C ∗ C^{\ast} -algebras Thanks: This was supported by DFF grants 1054-00094B and 1026-00371B Ben Bouwen, James Gabe Abstract . We provide a new proof of the Kirchberg–Phillips theorem by adapting the framework laid out by Carrión–Gabe–Schafhauser–Tikuisis–White for classifying separable simple unital nuclear stably finite 𝒵 \mathcal{Z} -stable C ∗ C^{\ast} -algebras satisfying the UCT. Not only does this give a unified approach to classifying stably finite and purely infinite C ∗ C^{\ast} -algebras, in contrast to the other proofs of the Kirchberg–Phillips theorem, our proof does not rely on Kirchberg’s Geneva Theorems, but instead implies them as corollaries (for nuclear C ∗ C^{\ast} -algebras). 2020 Mathematics Subject Classification 46L35, 46L80 1. Introduction The Kirchberg–Phillips theorem, proved independently by Kirchberg ( [ 21 ] ) and Phillips ( [ 28 ] ) in the 90s, marked a major advancement in the Elliott classification program. It provided a complete classification by K -theory of separable, nuclear, simple, purely infinite C^{*} -algebras satisfying the universal coefficient theorem (UCT). In contrast, its stably finite counterpart took the joint effort of numerous researchers ( [ 18 ] , [ 19 ] , [ 13 ] , [ 34 ] and the vast body of work leading up to these papers) over the past three decades to complete. Combining the classification theorems on both sides gives a comprehensive classification of all separable, nuclear, unital, simple, {\mathcal{Z}} -stable C^{*} -algebras satisfying the UCT, a landmark result in classification of C^{*} -algebras; we will refer to C^{*} -algebras satisfying these conditions as classifiable. In recent years, a new, more abstract approach to classification was developed by Carrión, the author JG, Schafhauser, Tikuisis and White in [ 6 ] . This approach is carried out in the stably finite setting, while for the purely infinite case, the original Kirchberg–Phillips theorem is used. This paper adapts the new abstract framework and techniques to the purely infinite setting to prove the Kirchberg–Phillips theorem, paving the way to a unified approach for classifying C^{*} -algebras. One of the major upsides of the new approach is that Kirchberg’s Geneva theorems in the nuclear setting (Corollary C ) are obtained as straightforward consequences of classification. These were crucial and highly non-trivial ingredients in the original proof, Rørdam’s summary [ 30 ] and Gabe’s alternative proof [ 16 ] . With the new approach, we no longer require these results as prerequisites. Purely infinite C^{*} -algebras The distinction between stably finite and purely infinite C^{*} -algebras is in complete analogy with the dichotomy between semifinite and infinite von Neumann algebras. There exist simple, separable and nuclear C^{*} -algebras which fall between these two classes ( [ 31 ] ), but due to a result by Kirchberg ( [ 30 , Theorem 4.1.10] ), this behavior cannot occur for classifiable C^{*} -algebras. Moreover, the dichotomy is witnessed by the presence of traces: a classifiable C^{*} -algebra is stably finite if and only if it has a tracial state, and is purely infinite otherwise. Purely infinite, simple C^{*} -algebras were introduced by Cuntz as C^{*} -algebraic analogues of type III factors in [ 9 ] , where he defined the Cuntz algebras \mathcal{O}_{n} and showed that they are examples of such C^{*} -algebras. Simple and purely infinite C^{*} -algebras arise naturally as C^{*} -algebraic constructions from various dynamical systems, examples being Cuntz–Krieger algebras (derived from symbolic dynamics) and crossed products ( [ 1 ] ). Building on deep work in tensor products and exactness, Kirchberg announced breakthrough results at the 1994 ICM satellite conference in Geneva: the {\mathcal{O}_{2}} -embedding theorem, stating that every separable, exact C^{*} -algebra embeds into {\mathcal{O}_{2}} ( [ 23 ] ), and the {\mathcal{O}_{\infty}} -absorption theorem, which says that a separable, simple, nuclear C^{*} -algebra is purely infinite if and only if it absorbs {\mathcal{O}_{\infty}} tensorially ( [ 22 ] ). These results known commonly as Kirchberg’s Geneva Theorems were key ingredients in the proofs of the Kirchberg–Phillips theorem. Following this theorem, more general classification results in the purely infinite setting were obtained in the subsequent decades. Kirchberg outlined a strategy to classify all separable, nuclear, {\mathcal{O}_{\infty}} -stable C^{*} -algebras up to stable isomorphism by ideal-related KK -theory. This result was proved (in a very different way) in [ 16 ] by the author JG. An abstract approach to classification On the stably finite side, progress was slower and more involved. The prevailing strategy here consisted of a very careful analysis of the internal structure of these C^{*} -algebras. For a more in-depth survey, see for example [ 6 , §1.2] . More recently, a distinct, more abstract strategy to classify stably finite C^{*} -algebras was laid out by Carrión, Gabe, Schafhauser, Tikuisis and White ( [ 6 ] ). Ever since Elliott’s classification of A\mathbb{T} -algebras ( [ 12 ] ) and Rørdam’s classification of certain crossed products ( [ 29 ] ), one of the guiding principles in the classification program in general has been to first classify * -homomorphisms between C^{*} -algebras, and then deduce a classification of the C^{*} -algebras themselves by use of an intertwining argument. In the unital setting, we then aim to classify unital * -homomorphisms between unital C^{*} -algebras A and B . In practice, it turns out to be much more tractable to first classify approximate * -homomorphisms, sequences of maps between A and B which become * -homomorphisms in the limit. These can be encoded more elegantly as honest * -homomorphisms from A into the sequence algebra B_{\infty}:=\ell^{\infty}(B)/c_{0}(B) . While B is assumed to be simple, B_{\infty} is not, so we need to put a simplicity condition on the allowed maps. A * -homomorphism \varphi:A\to B is said to be full if \varphi(a) is full for each non-zero a\in A , i.e. \varphi(a) generates B as an ideal. The equivalence relation up to which one aims to classify maps is approximate unitary equivalence, as this is what is necessary in order to apply the intertwining argument. In the sequence algebra, this corresponds to unitary equivalence, so in other words, the main objective becomes classifying unital and full * -homomorphisms A\to B_{\infty} up to unitary equivalence. In the new approach, the main object of study is the so-called trace-kernel extension of B , a short exact sequence given by (1.1) {\lx@inpgf@ignorespaces 0} {\lx@inpgf@ignorespaces J_{B}} {\lx@inpgf@ignorespaces B_{\infty}} {\lx@inpgf@ignorespaces B^{\infty}} {\lx@inpgf@ignorespaces 0,} \scriptstyle{\lx@inpgf@ignorespaces j_{B}} \scriptstyle{\lx@inpgf@ignorespaces q_{B}} where (1.2) J_{B}:=\{(x_{n})_{n}\in B_{\infty}\mid\textstyle{\sup_{\tau\in T(B)}}\tau(x_{n}^{*}x_{n})\xrightarrow{n\to\infty}0\}, T(B) is the set of tracial states on B and B^{\infty}:=B_{\infty}/J_{B} . This construction becomes especially nice in (and was partially inspired by) the case where B has a unique trace \tau . In this setting (and after replacing the sequence algebra with the ultrapower B_{\omega} ), the quotient is isomorphic with the von Neumann algebra (\pi_{\tau}(B)^{\prime\prime},\tau)^{\omega} , the tracial von Neumann ultrapower of \pi_{\tau}(B)^{\prime\prime} , with \pi_{\tau} the GNS representation associated to \tau . The main idea is now to first use von Neumann algebraic classification to classify maps into B^{\infty} . In general, B^{\infty} will not be a von Neumann algebra, but it still retains enough of this flavor to allow for von Neumann algebraic classification techniques to be used; see [ 7 , Theorem