On definability of C ∗ \mathrm{C}^{*} -tensor norms
Isaac Goldbring, Thomas Sinclair · 2025
Read On definability of C ∗ \mathrm{C}^{*} -tensor norms by Isaac Goldbring, Thomas Sinclair (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-2509-15086.txt.
Introduction
Isaac Goldbring and Thomas Sinclair ask whether tensor product norms for C-star algebras can be defined inside the algebras' own logic, in a 2025 mathematics paper. C-star algebras are the abstract backbone of quantum theory, and combining two systems requires choosing a tensor norm. The minimal and maximal norms are the two extreme choices, and they can differ.
Working in continuous model theory, the authors prove neither norm is definable uniformly across all C-star algebras. The minimal case leans on a deep theorem of Kirchberg about exactness, and the maximal case on a recent characterization by Pisier of the lifting property. Further negative results cover particular algebras and stricter, computable forms of definability, drawing on quantum complexity breakthroughs. A question of Fritz, Netzer, and Thom falls along the way.
Catalog record
| CREATOR | Isaac Goldbring, Thomas Sinclair |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 72 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-15086
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
On definability of C ∗ \mathrm{C}^{*} -tensor norms Thanks: I. Goldbring was partially supported by NSF grant DMS-2054477. Thanks: T. Sinclair was partially supported by NSF grant DMS-2055155. Isaac Goldbring, Thomas Sinclair Abstract . We initiate the study of definability (in the model-theoretic sense) of C ∗ \mathrm{C}^{*} -tensor norms. We show that neither the minimal nor maximal tensor norms are definable uniformly over all C ∗ \mathrm{C}^{*} -algebras. The proof in the case of the minimal tensor product norm uses a deep theorem of Kirchberg characterizing exactness in terms of tensor products with matrix ultraproducts while the case of maximal tensor products uses Pisier’s recent characterization of the lifting property in terms of maximal tensor products and ultraproducts. We also study the question of when one of these tensor products can be definable in a particular C ∗ \mathrm{C}^{*} -algebra. We establish some negative results along these lines for particular C ∗ \mathrm{C}^{*} -algebras and when the definability condition is strengthened to be computable and of a restricted quantifier-complexity; these results use the quantum complexity results MIP ∗ =RE and MIP c o = {}^{co}= coRE. As a byproduct of our arguments, we answer a question of Fritz, Netzer, and Thom by showing that the norm on C ∗ ( 𝔽 n × 𝔽 n ) C^{*}(\mathbb{F}_{n}\times\mathbb{F}_{n}) is not computable for any n ∈ { 2 , 3 , … , ∞ } n\in\{2,3,\ldots,\infty\} . 1. Introduction One of the more interesting aspects of \mathrm{C}^{*} -algebra theory is the problem of defining the tensor product of \mathrm{C}^{*} -algebras. Given \mathrm{C}^{*} -algebras A and B , their so-called algebraic tensor product A\odot B is simply their tensor product as vector spaces. It is quite easy to see that A\odot B can be equipped with a * -algebra structure. The question becomes: how can one equip A\odot B with a norm such that its completion with respect to this norm is a \mathrm{C}^{*} -algebra? In general, there can be many ways of accomplishing this task, that is, there can be norms on A\odot B whose respective completions are nonisomorphic \mathrm{C}^{*} -algebras. For example, a result of Junge and Pisier [ 19 ] is that the algebraic tensor product B(H)\odot B(H) admits a continuum of \mathrm{C}^{*} -norms whose completions are nonisomorphic \mathrm{C}^{*} -algebras. That being said, there are always two canonical norms on A\odot B , the so-called minimal tensor norm \|\cdot\|_{\min} and maximal tensor norm \|\cdot\|_{\max} . As the names indicate, they are indeed the least and greatest norms on A\odot B whose completions are \mathrm{C}^{*} -algebras. Their precise definitions will be recalled in the next section, but for now let us remark that they are both defined in terms of representations of the algebras; in other words, their definitions are “extrinsic.” One might wonder if it possible to give a more “intrinsic” definition of these norms, that is, formulae for these norms that are “definable” in terms of the algebra structures on A and B themselves. To prevent overcomplication, we mainly study this question under the assumption that A=B . Presumably the answer to the previous inquiry, in general, is: no! Indeed, if there were such definitions, then they would be written down in every \mathrm{C}^{*} -algebra textbook. But the problem with verifying this seemingly obvious attitude is: what exactly does one mean by “definable”? One way of making this idea precise is to use the language of model theory and that is the approach taken in the current paper. Roughly speaking, to say that the tensor norm is definable is to say that it can be given by a formula in the natural first-order language for studying \mathrm{C}^{*} -algebras. A more general and precise definition of definability can be found in the next section. Before describing our main results, we first mention that, for \|\cdot\|_{\alpha} for \alpha\in\{\min,\max\} , there are actually two different questions one can ask concerning their definability: (1) Is the norm \|\cdot\|_{\alpha} on A\odot A definable uniformly across all \mathrm{C}^{*} -algebras A ? (2) Given a particular \mathrm{C}^{*} -algebra A , is the norm \|\cdot\|_{\alpha} on A\odot A definable? We refer to these questions as the “global” and ”local” questions respectively. We begin by discussing our results in the global case. For each of the two choices for \alpha , the answer is indeed negative: Theorem . For \alpha\in\{\min,\max\} , the norm \|\cdot\|_{\alpha} is not uniformly definable across all \mathrm{C}^{*} -algebras. To prove this theorem, we introduce the following notion: given a \mathrm{C}^{*} -algebra A , an ultrafilter \mathcal{U} , and \alpha\in\{\min,\max\} , we say that A has property \alpha - \mathcal{U} if there is a natural isometric embedding A^{\mathcal{U}}\otimes_{\alpha}A^{\mathcal{U}}\hookrightarrow(A\otimes_{\alpha}A)^{\mathcal{U}} . (A more precise version of this definition will be given in Section 3 below.) We show that if the norm \|\cdot\|_{\alpha} is uniformly definable across all \mathrm{C}^{*} -algebras, then each \mathrm{C}^{*} -algebra A has property \alpha - \mathcal{U} . On the other hand, we prove the following theorem: Theorem . For each \alpha\in\{\min,\max\} , there are \mathrm{C}^{*} -algebras without property \alpha - \mathcal{U} . In the case of the minimal tensor product, using a deep theorem of Kirchberg, we show that if A is not exact, then A does not have property \min - \mathcal{U} for any nonprincipal ultrafilter \mathcal{U} on \mathbb{N} . Meanwhile, the examples of \mathrm{C}^{*} -algebras without property \max - \mathcal{U} use Pisier’s recent ultraproduct characterization of the lifting property [ 25 ] . On the other hand, if one restricts to n -subhomogeneous \mathrm{C}^{*} -algebras for a fixed n , then the minimal tensor product is uniformly definable: Theorem . For a fixed positive integer n , \|\cdot\|_{\min} is uniformly definable over all n -subhomogeneous \mathrm{C}^{*} -algebras. The proof of this theorem uses, amongst other things, the Beth Definability Theorem from model theory together with the fact that all separable subhomogeneous \mathrm{C}^{*} -algebras are subalgebras of matrix algebras over commutative \mathrm{C}^{*} -algebras. We now turn to the “local” situation. Unfortunately, our results in this case are only partial as showing that \|\cdot\|_{\alpha} is not definable when restricted to a particular \mathrm{C}^{*} -algebra seems to be a much more delicate problem. In order to explain our partial results, say that \|\cdot\|_{\alpha} is explicitly definable for a given \mathrm{C}^{*} -algebra A if there is an “effective” mapping (n,\epsilon)\mapsto\varphi_{n,\epsilon}(\vec{x},\vec{y}) , where \varphi_{n,\epsilon} is a formula in 2n -variables, for which, given any \vec{a},\vec{b}\in A^{n} , one has \left|\left\|\sum_{k=1}^{n}a_{k}\otimes b_{k}\right\|_{A\otimes_{\alpha}A}-\varphi_{n,\epsilon}(\vec{a},\vec{b})^{A}\right|<\epsilon. Given this definition, we can now state our local results: Theorem . Suppose that A is a locally universal \mathrm{C}^{*} -algebra. Then \|\cdot\|_{\min} is not explicitly definable in A . If, in addition, A has Kirchberg’s QWEP property, then \|\cdot\|_{\max} is not explicitly definable in A . Here, a \mathrm{C}^{*} -algebra is locally universal if every \mathrm{C}^{*} -algebra embeds in an ultrapower of A . We note that every existentially closed \mathrm{C}^{*} -algebra (see [ 16 ] ) is both locally universal and has the QWEP (in fact, the WEP), whence satisfies the hypotheses of the above theorem. We note also that the hypothesis of the second part of the above theorem can be weakened to assuming that (A,A) is a Tsirelson pair in the sense introduced by the first author and Hart