(1) The positive quotient and polar decomposition. How does polar decomposition generalize to all unital $C^*$-algebras, and not just the supported ones? This question may be answered in terms of a canonical correspondence between wide étale subcategories of $\mathcal{M}$ and the quotients in $\mathscr{M}$ of the representable presheaf associated with the unit of $\mathcal{A}$.
(2) The Gelfand spectrum of a commutative $C^*$-algebra: 1-dimensional representations. How this is related to $\mathcal{M}$ follows a pattern similar to the Zariski spectrum of a commutative ring.
(3) GNS-representation theory (in any dimension), and the spectrum of an arbitrary
$C^*$-algebra. One of our tools is what we shall call a seminormed $\mathcal{A}$-module,
by which we mean a right $\mathcal{A}$-module $\mathcal{V}$ that carries a seminorm such that
\[
\forall v \in \mathcal{V}, \; \forall U \in \mathcal{A}, \quad \|vU\| \leq \|v\| \, \|U\| .
\]
The functional dual $B(\mathcal{A})$, consisting of bounded $\mathbb{C}$-linear maps
$\tau : \mathcal{A} \to \mathbb{C}$, is an important $\mathcal{A}$-module especially
for GNS-theory. Its right action is defined by
\[
\tau U(X) = \tau(XU^*), \quad \tau \in B(\mathcal{A}).
\]
Moreover, this action satisfies $\|\tau U\| \leq \|\tau\| \, \|U\|$, so $B(\mathcal{A})$
is a normed $\mathcal{A}$-module. $B(\mathcal{A})$ classifies functionals in the sense that
for any seminormed $\mathcal{A}$-module $\mathcal{V}$ there is a natural isomorphism
\[
B(\mathcal{V}) \cong \mathcal{A}\text{-Bdd}(\mathcal{V}, B(\mathcal{A})).
\]
Therefore, the presheaf associated with $B(\mathcal{A})$ is given by
\[
\widehat{B(\mathcal{A})}(R) = \mathcal{A}\text{-Bdd}(R\mathcal{A}, B(\mathcal{A}))
\;\cong\; B(R\mathcal{A}).
\]
From this point of view, $\widehat{B(\mathcal{A})}$ is the presheaf of functional germs
on $\mathcal{A}$, suggesting that $\widehat{B(\mathcal{A})}$ may be interpreted as the
complex numbers object of the topos $\mathscr{M}$. In any case, $\widehat{B(\mathcal{A})}$
is a ring object with a conjugation operation (such that
$\overline{\tau}(SX) = \overline{\tau(SX)}$) internal to $\mathscr{M}$.
The presentation of a commutative ring by generators and relations is (at least superficially) similar to the presentation of a first order theory in terms of sorts, function/relation symbols and axioms. More concretely, we can associate categories to rings and to theories:
Of course, there are many other categories we could have chosen on each side, but these particular constructions admit a mutual generalization which we call ultrarings. In this talk we explain what ultrarings are, how their presentations generalize those of commutative rings and first-order theories, and how they connect to the logical approach to complexity theory (descriptive complexity). We will also sketch how we hope to exploit this connection in the future to transport tools from algebraic geometry.
This talk is based on work with Baptiste Chanus and Damiano Mazza, https://doi.org/10.4230/LIPIcs.FSCD.2025.13, with slightly updated definitions.