---
title: Global Glimm Property in C*-Algebras
url: https://www.emergentmind.com/topics/global-glimm-property
type: topic
---

# Global Glimm Property in C*-Algebras

The Global Glimm Property is a regularity property of \(C^*\)-algebras introduced in the non-simple setting by Kirchberg–Rørdam and formulated in current work as a hereditary square-zero condition: every hereditary subalgebra contains an almost full square-zero element. It is tightly linked to nowhere scatteredness, to divisibility properties of the Cuntz semigroup, and to several major open problems in the structure theory of non-simple \(C^*\)-algebras, including the weakly purely infinite problem and non-simple variants of the Toms–Winter program [2204.13059] [2512.13334].

## 1. Definition and basic formulations

In the formulation used throughout recent work, a \(C^*\)-algebra \(A\) has the Global Glimm Property if for every \(a\in A_+\) and every \(\varepsilon>0\) there exists \(r\in \overline{aAa}\) such that
\[
r^2=0,
\qquad
(a-\varepsilon)_+ \in \operatorname{Ideal}(r)=\overline{ArA}.
\]
The 2025 paper on topological dimension zero notes that the abstract says “almost full nilpotent element,” while the body uses the more precise phrase “almost full square-zero element” [2507.16261]. The 2022 paper gives the equivalent Kirchberg–Rørdam formulation: for every \(a\in A_+\), every \(\varepsilon>0\), and every integer \(k\ge 2\), there exists a \(^*\)-homomorphism
\[
\varphi: C_0((0,1],M_k)\longrightarrow aAa
\]
such that \((a-\varepsilon)_+\) belongs to the closed ideal generated by \(\varphi(C_0((0,1],M_k))\) [2204.13059].

This property is a global analogue of Glimm’s square-zero phenomenon for simple non-type-I algebras. The survey emphasizes that the Global Glimm Problem asks whether there is a single correct non-simple analogue of non-elementariness, or whether nowhere scatteredness and the Global Glimm Property are genuinely distinct [2512.13334].

A decisive reformulation is via the Cuntz semigroup. For a \(C^*\)-algebra \(A\),
\[
A \text{ has the Global Glimm Property}
\iff
\Cu(A) \text{ is } (2,\omega)\text{-divisible}.
\]
Here \((2,\omega)\)-divisibility means that for every \(x',x\in \Cu(A)\) with \(x'\ll x\), there exist \(n\in\mathbb N\) and \(d\in \Cu(A)\) such that
\[
2d\le x,
\qquad
x' \le nd.
\]
This equivalence is stated in both the 2022 structural paper and the 2025 topological-dimension-zero paper [2204.13059] [2507.16261].

## 2. Nowhere scatteredness and the Global Glimm Problem

The minimal regularity condition paired with the Global Glimm Property is nowhere scatteredness. One formulation is that \(A\) has no nonzero elementary ideal-quotients [2204.13059]. Another, used in the 2025 paper, is that no hereditary sub-\(C^*\)-algebra of \(A\) admits a finite-dimensional irreducible representation [2507.16261]. The survey records these as equivalent to several other conditions, including that no hereditary subalgebra has a one-dimensional irreducible representation and, for separable \(A\), that \(\widehat A\) is nowhere scattered as a topological space [2512.13334].

The Global Glimm Property always implies nowhere scatteredness. This implication is treated as known in the 2022 paper and is reiterated in the 2025 papers [2204.13059] [2507.16261]. The Global Glimm Problem is the converse question:
\[
A \text{ nowhere scattered}
\ \stackrel{?}{\Longrightarrow}\
A \text{ has the Global Glimm Property}.
\]
The survey describes this problem as open for over two decades and as central to several other regularity questions [2512.13334].

The Cuntz-semigroup translation makes the contrast precise. Nowhere scatteredness is equivalent to weak \((2,\omega)\)-divisibility of \(\Cu(A)\), meaning that for every \(x',x\in \Cu(A)\) with \(x'\ll x\), there exist \(n\in\mathbb N\) and \(d_1,\dots,d_n\in \Cu(A)\) such that
\[
2d_1,\dots,2d_n\le x,
\qquad
x' \le d_1+\cdots+d_n.
\]
Thus the Global Glimm Problem becomes the question whether weak \((2,\omega)\)-divisibility upgrades to full \((2,\omega)\)-divisibility for Cuntz semigroups arising from \(C^*\)-algebras [2204.13059] [2507.16261].

The 2025 survey records several positive cases known before the most recent work: real rank zero, stable rank one, Hausdorff finite-dimensional primitive ideal space, and finite decomposition rank [2512.13334]. The 2025 topological-dimension-zero paper adds a new broad class by proving the equivalence for all \(C^*\)-algebras of topological dimension zero [2507.16261].

## 3. Cuntz-semigroup mechanism: ideal-filteredness and property (V)

The 2022 paper turns the Global Glimm Problem into a purely order-theoretic question about \(\Cu(A)\). Its main theorem states that for a \(C^*\)-algebra \(A\),
\[
A \text{ has the Global Glimm Property}
\iff
A \text{ is nowhere scattered and } \Cu(A) \text{ is ideal-filtered and has property (V)}.
\]
Equivalently,
\[
\Cu(A) \text{ is } (2,\omega)\text{-divisible}
\iff
\Cu(A) \text{ is weakly }(2,\omega)\text{-divisible, ideal-filtered, and has property (V)}.
\]
The abstract Cu-semigroup theorem is proved under axioms (O5)–(O8), which are automatic for Cuntz semigroups of \(C^*\)-algebras [2204.13059].

Ideal-filteredness is a downward-directedness condition for generators of ideals. One equivalent formulation given in the paper is: whenever
\[
v'\ll v\le \omega x,\omega y,
\]
there exists \(z\) with
\[
v'\ll z\le \omega x,\omega y.
\]
Property (V) is a weak two-variable replacement for a sup-semilattice structure. In the formulation used there, if
\[
d'_j \ll d_j \quad (j=1,2), \qquad d_1,d_2\le c, \qquad c+d_1,\ c+d_2 \ll x,
\]
then there exist \(y,z\in \Cu(A)\) such that
\[
y+z\le x,
\qquad
d'_1+d'_2 \le \omega y,\ \omega z.
\]
These two properties identify the precise gap between weak divisibility and full divisibility [2204.13059].

The same paper shows that ideal-filteredness and property (V) are automatic in several important classes. In particular, they hold for \(C^*\)-algebras of stable rank one and for \(C^*\)-algebras of real rank zero, thereby recovering earlier solutions of the Global Glimm Problem in those settings [2204.13059].

## 4. Solved cases and the topological-dimension-zero theorem

A \(C^*\)-algebra \(A\) has topological dimension zero if \(\operatorname{Prim}(A)\) has a basis consisting of compact, open sets [2507.16261]. Brown–Pedersen introduced this notion as a generalization of real rank zero, and the 2025 paper proves a sharp theorem in this class:
\[
A \text{ has topological dimension zero}
\quad\Longrightarrow\quad
\bigl(A \text{ has GGP}\iff A \text{ is nowhere scattered}\bigr).
\]
This is stated as solving the Global Glimm Problem in that setting [2507.16261].

The proof passes through the Cuntz semigroup. The paper proves
\[
A \text{ has topological dimension zero}
\iff
\Cu(A)\otimes \{0,\infty\} \text{ is algebraic}.
\]
Using earlier work, nowhere scatteredness gives weak \((2,\omega)\)-divisibility of \(\Cu(A)\). A new Cu-semigroup lemma then shows that weak \((2,\omega)\)-divisibility upgrades to full \((2,\omega)\)-divisibility when \(S\otimes\{0,\infty\}\) is algebraic. Combining this with the Cuntz-semigroup characterization of the Global Glimm Property yields the theorem [2507.16261].

The survey situates this result among several other solved cases. It records that nowhere scattered \(C^*\)-algebras with Hausdorff finite-dimensional primitive ideal space have the Global Glimm Property, and that stable rank one algebras satisfy nowhere scatteredness \(\iff\) Global Glimm Property [2512.13334]. The 2022 paper likewise states the stable-rank-one and real-rank-zero results in the Cu-semigroup language [2204.13059].

The topological-dimension-zero theorem has immediate consequences. One is that if \(A\) is nowhere scattered, has finite nuclear dimension, and has topological dimension zero, then \(A\) is pure [2507.16261]. Another is that for \(C^*\)-algebras of topological dimension zero,
\[
A \text{ is purely infinite } \iff A \text{ is weakly purely infinite},
\]
thereby resolving the weakly purely infinite problem in that regime [2507.16261].

## 5. Soft elements, purity, and related regularity phenomena

The 2023 paper on soft operators introduces a structural refinement between the Global Glimm Property and nowhere scatteredness. A \(C^*\)-algebra is soft if it has no nonzero unital quotients, and an element \(x\in A\) is soft if the hereditary subalgebra \(x^*Ax\) is soft [2304.11644]. For positive elements, softness admits a spectral characterization: a positive element \(a\in A_+\) is soft if and only if for every closed ideal \(I\subset A\), either \(a\in I\) or \(0\) is a limit point of the spectrum of \(a+I\) [2304.11644].

The same paper defines an abundance of soft elements: for every \(a\in A_+\) and \(\varepsilon>0\), there exists a positive soft element \(b\in aAa\) such that
\[
(a-\varepsilon)_+ \triangleleft b,
\]
meaning that \((a-\varepsilon)_+\) lies in the ideal generated by \(b\) [2304.11644]. This is formally parallel to the square-zero definition of the Global Glimm Property, but with “soft” in place of “square-zero”.

The structural implications are one-sided in general:
\[
\text{Global Glimm Property} \Rightarrow \text{abundance of soft elements} \Rightarrow \text{nowhere scattered}.
\]
The first implication is proved via Cuntz-semigroup divisibility, and the second via weak \((2,\omega)\)-divisibility [2304.11644]. The paper further shows
\[
A \text{ has the Global Glimm Property}
\iff
A \text{ has an abundance of soft elements and } \Cu(A) \text{ is ideal-filtered}.
\]
This identifies abundance of soft elements as a new intermediate regularity property and isolates ideal-filteredness as the remaining obstruction [2304.11644].

The survey places these results in a broader regularity framework. It emphasizes that a positive solution of the Global Glimm Problem would imply a positive solution to the weakly purely infinite problem and would settle large parts of the non-simple Toms–Winter program, because finite nuclear dimension together with the Global Glimm Property yields pureness, hence strict comparison, and in many settings \(\mathcal Z\)-stability [2512.13334]. The survey also states that for finite nuclear dimension the Global Glimm Property implies purity, and that for stable algebras pureness is equivalent to strict comparison together with divisibility of ranks [2512.13334].

## 6. Scope of the term and related Glimm-type literature

Recent \(C^*\)-algebra papers use “Global Glimm Property” for the hereditary square-zero condition and the associated Cuntz-semigroup divisibility problem [2204.13059] [2507.16261] [2512.13334]. Related literature uses “global Glimm” in other, older senses tied to Glimm spaces and orbit-space regularity.

In tensor-product theory, McConnell studies how the global topology and ideal structure encoded in the Glimm space behaves under minimal tensor products, showing that when \(A\) has “good” global Glimm behavior, \(A\otimes_\alpha B\) inherits a product-type Glimm space for every \(B\) [1212.3971]. In work on compact subsets of \(\mathrm{Glimm}(A)\), Lazar describes a global control phenomenon for the entire Glimm space in terms of norm superlevel sets associated to strictly positive elements [1203.3350]. In the separable case, Lazar–Somerset characterize the topology of Glimm spaces of separable \(C^*\)-algebras and obtain a global decomposition into locally compact metrizable and non-locally-compact parts [1509.06140]. Ungermann, in a different direction, generalizes Glimm’s theorem for orbit spaces of hereditary Lindelöf locally compact groups and describes a “global Glimm property” for \(G\)-spaces in terms of almost Hausdorff orbit spaces and homogeneous orbits [1202.5018].

These uses are conceptually related through Glimm’s original concerns with non-type-I structure, quotient topology, and orbit-space regularity. A plausible implication is that the contemporary Global Glimm Property isolates the non-simple square-zero/divisibility aspect of that legacy, while the earlier literature treats global topological behavior of Glimm spaces and orbit spaces. Within current regularity theory for \(C^*\)-algebras, however, the decisive technical content is the equivalence between hereditary square-zero generation and \((2,\omega)\)-divisibility of \(\Cu(A)\), together with the open question whether nowhere scatteredness already forces that divisibility [2204.13059] [2512.13334].

The present state of the subject is therefore sharply stratified. The implication
\[
\text{Global Glimm Property} \Rightarrow \text{nowhere scattered}
\]
is settled, the converse is known in several major classes, and topological dimension zero provides the broadest recent positive solution [2507.16261]. The general problem remains open, but its reformulation through weak versus full \((2,\omega)\)-divisibility, ideal-filteredness, property (V), and abundance of soft elements has made the obstruction highly explicit [2204.13059] [2304.11644].

Source: https://www.emergentmind.com/topics/global-glimm-property