Ulam stability for classes of nuclear C*-algebras

Vadim Alekseev, Ilijas Farah, Andreas Thom · 2026

Read Ulam stability for classes of nuclear C*-algebras by Vadim Alekseev, Ilijas Farah, Andreas Thom (2026) 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-2606-03757.txt.

Introduction

Alekseev, Farah, and Thom study Ulam stability for near maps between C-star algebras. The term asks when near-solutions of an equation sit close to exact ones. Here the maps are near homomorphisms, and the targets are von Neumann algebras. Nuclear cases are the focus.

Stability is proven for the abelian case in this setting. Large classes from the Elliott program also qualify. Stability survives standard constructions, as permanence results show. Cases where it fails are marked by counterexamples. Uses include rigidity results for corona algebras. Operator algebraists get a map of which nuclear classes are stable. The work links approximation to the classification program in a concrete way.

Catalog record

CREATORVadim Alekseev, Ilijas Farah, Andreas Thom
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE89 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2606.03757

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Ulam stability for classes of nuclear C*-algebras Vadim Alekseev, Ilijas Farah, Andreas Thom Abstract . We study Ulam stability for approximate ∗ * -homomorphisms of C ∗ C^{*} -algebras. We prove stability results for several classes of nuclear C ∗ C^{*} -algebras with respect to von Neumann algebra targets, including abelian C ∗ C^{*} -algebras and large classes arising in the Elliott classification program. We also discuss permanence properties, counterexamples, and related stability phenomena. As applications, we obtain rigidity and independence results for corona algebras. 1. Introduction In [ 40 ] , S. Ulam defined an \varepsilon -homomorphism from a group G into a metric group (H,d) as a map \varphi\colon G\to H that satisfies d(\varphi(gh),\varphi(g)\varphi(h))<\varepsilon for all g,h\in G , and asked under what conditions such \varepsilon -homomorphism can be uniformly approximated by a true homomorphism. This definition has been adapted to other algebraic structures equipped with a metric, giving rise to the very general and well-studied concept of Ulam stability. Early on, Hyers observed that for an \varepsilon -homomorphism \varphi\colon\mathfrak{X}\to\mathfrak{Y} between Banach spaces the sequence 2^{-n}\varphi(2^{n}x) converges to a homomorphism that \varepsilon -approximates \varphi ( [ 23 ] , see also [ 22 ] ). In case of Banach space-based structures such as C*-algebras one could meaningfully weaken the definition of an \varepsilon -homomorphism in more than one way. B. Johnson studied Ulam stability of linear maps \varphi between Banach algebras that are approximately multiplicative, in the sense that \|\varphi(ab)-\varphi(a)\varphi(b)\|<\varepsilon\|a\|\|b\| for all a and b in the domain ( [ 25 ] , [ 24 ] ), apparently unaware of Ulam’s question. Tangentially related is the well-studied notion of weakly stable relations on a C*-algebra ( [ 30 ] ); it could be considered as a local-to-local analog of Ulam stability of * -homomorphisms, or as the Ulam stability of equations. The study of \varepsilon -C*-algebras ( [ 28 ] ) is subtly different from Ulam stability. Our notion of \varepsilon - * -homomorphisms defined below is motivated by the study of rigidity of coronas of C*-algebras ( [ 16 ] , [ 31 ] , [ 41 ] , [ 17 , §17.2] , [ 12 ] ); see § 7.2 (also Theorem D and Theorem C ) for details and applications of our results. The following definition first appeared explicitly in [ 31 , Def. 1.1] . Definition 1.1 . Let A,B be unital C^{*} -algebras and \varepsilon\geq 0 . An \varepsilon - * -homomorphism is a map \varphi\colon A\to B such that for all a,b\in A_{\leq 1} , and all \lambda\in\mathbb{C}_{\leq 1} , we have \varphi(a)\in B_{\leq 1} and \displaystyle\|\varphi(a+b)-\varphi(a)-\varphi(b)\| \displaystyle\leq\varepsilon, \displaystyle\|\varphi(ab)-\varphi(a)\varphi(b)\| \displaystyle\leq\varepsilon, \displaystyle\|\varphi(a^{*})-\varphi(a)^{*}\| \displaystyle\leq\varepsilon, \displaystyle\|\varphi(\lambda a)-\lambda\varphi(a)\| \displaystyle\leq\varepsilon. Definition 1.2 . We will denote by \sfFD the class of finite-dimensional C^{*} -algebras, by \sfAF the class of approximately finite-dimensional C^{*} -algebras, by \sfUC the class of unital abelian C^{*} -algebras, by \mathsf{C^{*}} the class of all C^{*} -algebras, and by \sfvN the class of von Neumann algebras. Definition 1.3 . Suppose that \mathsf{C} and \mathsf{D} are classes of C*-algebra. Following the conventions, we say that the pair (\mathsf{C},\mathsf{D}) is Ulam stable (or that \mathsf{C} is Ulam stable with respect to \mathsf{D} ) if there is a function f_{\mathsf{C},\mathsf{D}}(\varepsilon)\to 0 as \varepsilon\to 0 such that for every A\in\mathsf{C} , B\in\mathsf{D} , and \varepsilon - * -homomorphism \varphi\colon A\to B there is a * -homomorphism \psi\colon A\to B such that \sup_{\|a\|\leq 1}\|\varphi(a)-\psi(a)\|\leq f_{\mathsf{C},\mathsf{D}}(\varepsilon) . We will write f_{\mathsf{C}} for f_{\mathsf{C},\sfvN} . The following summarizes existing results on Ulam stability. Theorem 1.4 . Each of the following pairs is Ulam stable. (1) (\sfFD,\sfFD) , by the second author [ 16 ] . (2) (\sfFD,\mathsf{C^{*}}) , by McKenney–Vignati [ 31 ] . (3) (\sfAF,\sfvN) , by McKenney–Vignati [ 31 ] . (4) (\sfUC,\sfUC) , by Šemrl [ 36 ] . ∎ The implication from ( 2 ) to ( 3 ) is a consequence of a result of B. Johnson (see Theorem 4.1 ( 5 )). The result of ( 4 ) is a consequence of a more general theorem. In [ 36 , Theorem 2.1] it was proved that approximate homomorphisms (not necessarily * -homomorphisms, and with a different definition of ‘approximate’) from an arbitrary Banach algebra into a Banach algebra of the form C(X) for a compact Hausdorff X are Ulam-stable. This implies Ulam stability for approximate * -homomorphisms between unital abelian C*-algebras (see [ 12 , Theorem 5.18] ). The aim of this note is to extend Theorem 1.4 ( 3 ) to a much larger class of nuclear C*-algebras and give applications to corona rigidity, but we first give an alternative proof of Theorem 1.4 ( 4 ) in Theorem 3.5 . The case when the range is \mathbb{C} (which is the key step in our proof) has been proved independently by Jennifer Pi, by using completely different methods. After proving various closure properties of classes of C*-algebras that are Ulam stable with respect to von Neumann algebras, using the recent breakthrough results in the classification of nuclear C*-algebras ( [ 20 ] , [ 21 ] , also [ 6 ] ), we prove the following (it is a consequence of Theorem 5.8 ). Theorem A . The class of all stably finite Elliott-classifiable unital C*-algebras is Ulam stable with respect to von Neumann algebras. The following is proved as Corollary 5.1 . Theorem B . The class of all inductive limits of C*-algebras of the form C(X)\otimes F for a compact Hausdorff space X and an \sfAF algebra F is Ulam stable with respect to the von Neumann algebras. While this does not cover all AH C*-algebras (approximately homogeneous C*-algebras), it covers many counterexamples to Elliott’s conjecture, such as all Villadsen algebras of the first type ( [ 42 ] , see also [ 38 , §3.1] ) including Toms’s counterexample to the Elliott conjecture ( [ 39 , Theorem 1.1] ; see the beginning of Section 3). While the most interesting instance of these theorems is the case of separable C*-algebras, it should be noted that separability is not required. This is a consequence of a general fact proved by using a standard reflection argument discussed in § 5.3 . Our initial motivation comes from the rigidity theory of coronas of separable C*-algebras under set-theoretic assumptions. More details and proof of the Theorem C below(stated as Theorem 7.6 ) will be given in § 7.2 . Forcing axioms \OCAT and \MA will not be used explicitly, instead we will rely on results from [ 32 ] and [ 41 ] . What matters for us is that they are axioms relatively consistent with \ZFC that provide an environment for strongest possibly rigidity of quotient structures in several categories; see [ 12 ] for more details. As usual, for a C*-algebra A we denote its multiplier algebra by \mathcal{M}(A) , its corona \mathcal{M}(A)/A by \mathcal{Q}(A) , and write A_{\infty}=\ell_{\infty}(A)/c_{0}(A). Theorem C . Assume \OCAT and \MA . If A_{n} , B_{n} , for n\in\mathbb{N} , are Elliott-classifiable unital C*-algebras, then the following are equivalent. (1) The coronas of \bigoplus_{n}A_{n} and \bigoplus_{n}B_{n} are isomorphic. (2) There is a bijection \pi between cofinite subsets of \mathbb{N} such that A_{\pi(n)}\cong B_{n} for all n\in\dom(\pi) . Some

Read the complete text