We present an introductory survey to first order logic for metric structures and its applications to C*-algebras.
IntroductionThis survey is designed as an introduction to the study of C*-algebras from the perspective of model theory for metric structures. The intended readership consists of anyone interested in learning about this subject, and naturally includes both logicians and operator algebraists. Considering this, we will not assume in these notes any previous knowledge of model theory, nor any in-depth knowledge of functional analysis, beyond a standard graduate-level course. A familiarity with C*-algebra theory and the classification programme [29] might be useful as a source of motivation and examples. Several facts from C*-algebra theory will be used, and detailed references provided. Most of the references will be to the comprehensive monographs [13,70].Logic for metric structures is a generalization of classical (or discrete) logic, suitable for applications to metric objects such as C*-algebras. The monograph [11] presents a quick but complete introduction to this subject, and explains how the fundamental results from classical model theory can be recast in the metric setting. The model-theoretic study of C*-algebras has been initiated in [38][39][40] where, in particular, it shown how C*-algebras fit into the framework of model theory for metric structures. The motivations behind this study are manifold. With no pretense of exhaustiveness, we attempt to illustrate some of them here.The most apparent contribution of first-order logic is to provide a syntactic counterpart to the semantic construction of ultraproducts and ultrapowers: the notion of formulas. Formulas allow one to express the fundamental properties of ultrapowers of C*-algebras (saturation) and diagonal embeddings into the ultrapower (elementarity). These general principles underpin most of the applications of the ultraproduct construction in C*-algebra theory, as they have appeared in various places in the literature under various names-Kirchberg's ε-test, reindexing arguments, etc. Isolating such general principles provides a valuable service of clarification and uniformization in the development of C*-algebra theory. In particular, this allows one to distinguish between, on one hand, what is just an instance of "general nonsense" and, on the other, what is a salient point where C*-algebras theory is crucially used.This abstract model-theoretic point of view also makes it easier to recognize analogies between different contexts. Furthermore, it provides a language to formalize such analogies as precise mathematical statements, rather than just intuitive ideas. For instance, this paradigm can be applied to some aspects of the equivariant theory of C*-algebras, which studies C*-algebras endowed with a group action (C*-dynamical systems). At least when the acting group is compact or discrete, C*-dynamical systems fit in the setting of first-order logic [47]. Adopting this perspective, one can naturally and effortlessly transfer ideas an...