Non-embeddable \(\mathrm{II}_1\) factors resembling the hyperfinite \(\mathrm{II}_1\) factor (Q6039642)

This paper studies II\(_1\) factors, which are tracial semifinite von Neumann algebras of trivial center. Inarguably the most important II\(_1\) factor is the hyperfinite II\(_1\) factor \(\mathcal R\), which can be constructed using nested copies of matrix algebras. Connes showed that indeed \(\mathcal R\) is the unique separable II\(_1\) factor possessing any of the following properties: injectivity, semidiscreteness, or amenability. The study of II\(_1\) factors then focused on those objects which can at least be embedded in (tracial) ultrapowers of \(\mathcal R\). These are called \(\mathcal R\)-embeddable. The Connes Embedding Problem (CEP) asked whether all II\(_1\) factors are \(\mathcal R\)-embeddable. CEP was given a negative answer [\textit{Z. Ji} et al., Commun. ACM 64, No. 11, 131--138 (2021; Zbl 1503.68075)], and this paper departs from there. Among \(\mathcal R\)-embeddable II\(_1\) factors, \(\mathcal R\) can be characterised by many properties. For example, \(\mathcal R\) is the sole \(\mathcal R\)-embeddable separable II\(_1\) factor \(M\) satisfying any of these two properties: \begin{itemize} \item[(1)] it has the \textit{Jung property} (after [\textit{K.~Jung}, Math. Ann. 338, No.~1, 241--248 (2007; Zbl 1121.46052)] and [\textit{S.~Atkinson} et al., Adv. Math. 396, Article ID 108107, 53~p. (2022; Zbl 1487.46068)]), that is, any embedding of \(M\) into an ultrapower \(M^{\mathcal U}\) is unitarily conjugate to the diagonal embedding; \item[(2)] it is \textit{self-tracially stable} (after \textit{S.~Atkinson} [Indiana Univ. Math. J. 70, No.~3, 1167--1187 (2021; Zbl 1484.46050)]), that is, any embedding of \(M\) into \(M^\mathcal U\) has an `approximate lifting'. \end{itemize} Now, given that separable non-\(\mathcal R\)-embeddable II\(_1\) factors appear to exist, what do these properties mean in this setting? Is there even a single non \(\mathcal R\)-embeddable II\(_1\)-factor with the Jung property or which is self-tracially stable? Answering this question is the paper's main goal. Even though the author does not prove that such a factor exists, it is shown that certain model theoretic properties of a II\(_1\) factor imply conditions (1) and (2). In particular, the author shows that, if a II\(_1\) factor is existentially closed (this is a model theoretic concept on mutual embeddings between von Neumann algebras, and can be viewed as a genericity condition), then being self-tracially stable gives the Jung property. Further, it is shown that enforceable II\(_1\) factors, if such exist, have the Jung property, and the same is shown for certain prime models (which again, unfortunately, are not yet proven to exist).