---
title: Global Glimm Problem in C*-algebras
url: https://www.emergentmind.com/topics/global-glimm-problem
type: topic
---

# Global Glimm Problem in C*-algebras

Searching arXiv for recent papers on the Global Glimm Problem in C*-algebras.
The Global Glimm Problem is an open problem in the structure theory of non-simple \(C^*\)-algebras asking whether every nowhere scattered \(C^*\)-algebra has the Global Glimm Property. In a standard formulation, a \(C^*\)-algebra \(A\) has the Global Glimm Property if for every positive element \(a\in A_+\) and every \(\varepsilon>0\) there exists an element \(r\in\overline{aAa}\) such that \(r^2=0\) and \((a-\varepsilon)_+\in \mathrm{Ideal}(r)\); equivalently, for each integer \(k\ge 2\) one may require a nonzero \(*\)-homomorphism
\[
\varphi\colon M_k\bigl(C_0(0,1]\bigr)\longrightarrow \overline{aAa}
\]
whose image generates a hereditary subalgebra containing \((a-\varepsilon)_+\) [2512.13334]. It is known that the Global Glimm Property implies nowhere scatteredness, and the central question is whether the converse holds in general [2204.13059].

## 1. Historical origin and conceptual setting

The name of the problem goes back to Glimm’s work on Type I \(C^*\)-algebras. In the historical account given in the recent survey literature, Glimm’s original theorem shows that a simple non-Type I algebra contains, in each nonzero hereditary subalgebra, a square-zero element whose hereditary subalgebra is full; this “Glimm lemma” is the source of the later global formulation [2512.13334]. The modern problem emerged in the study of non-simple pure infiniteness, especially after Kirchberg and Rørdam introduced several notions of pure infiniteness for non-simple algebras and identified the Global Glimm Property as a missing regularity input [2512.13334].

This operator-algebraic formulation is now tied to several major regularity questions. The survey literature states that the problem has been open for over two decades and has strong ties to the weakly purely infinite problem and the non-simple Toms–Winter conjecture [2512.13334]. A complete positive solution would therefore be more than an isolated structural theorem: it would supply a mechanism for producing square-zero or nilpotent behavior uniformly across hereditary subalgebras and ideal-quotients.

## 2. Formulations of the property and the obstruction

The Global Glimm Property is most often stated as an “almost full nilpotent” condition. For each \(a\in A_+\) and \(\varepsilon>0\), one asks for an element \(r\in \overline{aAa}\) with
\[
r^2=0
\quad\text{and}\quad
(a-\varepsilon)_+ \in \overline{ArA},
\]
or, equivalently, for a \(*\)-homomorphism from \(C_0((0,1],M_k)\) into the hereditary subalgebra generated by \(a\) whose image generates the relevant ideal [2507.16261, 2204.13059].

The obstruction is encoded by the notion of nowhere scatteredness. In one formulation used by Thiel and Vilalta, \(A\) is nowhere scattered if no hereditary subalgebra of \(A\) admits a nonzero finite-dimensional irreducible representation [2204.13059]. In the survey formulation, \(A\) is nowhere scattered if no nonzero ideal-quotient \(I/J\) is elementary [2512.13334]. The basic implication
\[
\text{Global Glimm Property}\Longrightarrow \text{nowhere scattered}
\]
is immediate in the modern treatments; the unresolved direction is the converse [2204.13059].

A decisive advance was the reformulation in the Cuntz semigroup. If \(S=\mathrm{Cu}(A)\), then the Global Glimm Property is equivalent to \((2,\omega)\)-divisibility: for every \(x',x\in S\) with \(x'\ll x\), there exist \(n\in\mathbb N\) and \(d\in S\) such that
\[
2d\le x
\quad\text{and}\quad
x'\le nd.
\]
By contrast, nowhere scatteredness corresponds to weak \((2,\omega)\)-divisibility: for every \(x'\ll x\) there exist \(d_1,\dots,d_n\in S\) such that
\[
2d_i\le x,\quad i=1,\dots,n,
\qquad\text{and}\qquad
x'\le d_1+\dots+d_n.
\]
Thus the Global Glimm Problem becomes the problem of upgrading weak \((2,\omega)\)-divisibility to genuine \((2,\omega)\)-divisibility in \(\mathrm{Cu}(A)\) [2204.13059, 2507.16261].

## 3. Order-theoretic resolution in favorable classes

The paper "The Global Glimm Property" [2204.13059] isolates two additional order-theoretic conditions on \(\mathrm{Cu}(A)\): ideal-filteredness and property (V). In the formulation stated there, a Cu-semigroup \(S\) is ideal-filtered if for every \(v'\ll v\ll x,y\) in \(S\) there exists \(z\in S\) such that
\[
v'\ll z,\qquad z\ll x,\qquad z\ll y.
\]
Property (V) is a weak sup-semilattice-type condition: whenever
\[
d_1'\ll d_1,\; d_2'\ll d_2,\qquad d_1,d_2<c,\qquad c+d_1,\; c+d_2<x,
\]
there exist \(y,z\in S\) such that
\[
y+z\le x,\qquad d_1'+d_2'\ll y,\qquad d_1'+d_2'\ll z.
\]

With these notions in place, Thiel and Vilalta prove the central equivalence
\[
A\text{ has the Global Glimm Property}
\]
if and only if
\[
\mathrm{Cu}(A)\text{ is weakly }(2,\omega)\text{-divisible, ideal-filtered, and has property (V),}
\]
and hence if and only if
\[
A\text{ is nowhere scattered and }\mathrm{Cu}(A)\text{ is ideal-filtered and has property (V)}.
\]
This reframes the problem as the search for general mechanisms guaranteeing ideal-filteredness and property (V) [2204.13059].

The importance of this reformulation is methodological. Earlier proofs in special cases were often based on explicit square-zero constructions. The Cuntz-semigroup approach replaces those constructions by two abstract order properties, making it possible to recover old results and to identify genuinely new classes where the problem can be solved [2204.13059].

## 4. Classes where the problem is solved

Several major classes are now known to satisfy the Global Glimm Property exactly when they are nowhere scattered.

| Class of \(C^*\)-algebras | Result | Source |
|---|---|---|
| Stable rank one | \(A\) has GGP iff \(A\) is nowhere scattered | [2204.13059] |
| Real rank zero | \(A\) has GGP iff \(A\) is nowhere scattered | [2204.13059] |
| Hausdorff \(\mathrm{Prim}(A)\) of finite covering dimension | Nowhere scattered implies GGP | [2512.13334] |
| Topological dimension zero | \(A\) has GGP iff \(A\) is nowhere scattered | [2507.16261] |

For stable rank one and real rank zero, the point is that ideal-filteredness and property (V) become automatic. In the stable-rank-one case, \(\mathrm{Cu}(A)\) satisfies Riesz interpolation and is ideal-filtered; in the real-rank-zero case, \(\mathrm{Cu}(A)\) is zero-dimensional and compact elements come from projection classes in a refinement monoid, which yields property (V) [2204.13059].

The most complete recent solution concerns topological dimension zero. Here
\[
A\text{ has topological dimension zero}
\quad\Longleftrightarrow\quad
\mathrm{Prim}(A)\text{ admits a basis of compact-open sets,}
\]
equivalently, \(A\) has the ideal property in the sense that projections in \(A\) separate ideals [2507.16261]. Ng, Thiel, and Vilalta show that for such algebras,
\[
A\text{ has the Global Glimm Property}
\quad\Longleftrightarrow\quad
A\text{ is nowhere scattered,}
\]
thereby solving the Global Glimm Problem in this setting [2507.16261].

Their proof passes through the algebraicity of
\[
\mathrm{Cu}(A)\otimes\{0,\infty\},
\]
which is equivalent to topological dimension zero, and then uses a semigroup-theoretic lifting argument to upgrade weak \((2,\omega)\)-divisibility to full \((2,\omega)\)-divisibility under axioms \((O5)\)–\((O8)\) [2507.16261]. One corollary is that if \(A\) has finite nuclear dimension, topological dimension zero, and is nowhere scattered, then \(A\) is pure, where purity means that \(\mathrm{Cu}(A)\) is almost unperforated and almost divisible [2507.16261].

## 5. Soft operators and the current frontier

A distinct recent approach is developed in "Soft operators in \(C^*\)-algebras" [2304.11644]. There, a \(C^*\)-algebra is called soft if no proper closed ideal has a unital quotient, and an element \(x\in A\) is soft if the hereditary algebra \(x^*Ax\) is soft. For \(a\in A^+\), softness admits a spectral characterization:
\[
a\text{ soft}
\quad\Longleftrightarrow\quad
\forall\, I\lhd A,\;
a\notin I\Rightarrow 0\text{ is a limit point of }\sigma(a+I)
\]
[2304.11644].

The key intermediate notion is an abundance of soft elements: for every \(a\in A^+\) and \(\varepsilon>0\), there is a soft \(b\in aAa\) with
\[
(a-\varepsilon)_+\le b.
\]
Thiel and Vilalta prove that the Global Glimm Property implies an abundance of soft elements, and that an abundance of soft elements implies nowhere scatteredness [2304.11644]. Thus one obtains the chain
\[
\text{Global Glimm}
\Longrightarrow
\text{Abundance of soft elements}
\Longrightarrow
\text{Nowhere scattered.}
\]

They also show that abundance of soft elements is equivalent to a hereditary 2-splitting property and to an abundance of strongly soft elements in \(\mathrm{Cu}(A)\) [2304.11644]. Moreover, abundance of soft elements together with ideal-filteredness implies the Global Glimm Property. This splits the original problem into two subquestions: whether nowhere scatteredness forces abundance of soft elements, and whether abundance automatically brings the ideal-filtered behavior needed to recover full \((2,\omega)\)-divisibility [2304.11644]. This suggests that softness is not merely an auxiliary notion, but a candidate bridge between local hereditary structure and global divisibility.

## 6. Consequences, remaining open directions, and terminological scope

The broader significance of the Global Glimm Problem is twofold. First, in the non-simple pure infiniteness program, pure infiniteness is equivalent to weak pure infiniteness plus the Global Glimm Property, so a positive solution for weakly purely infinite algebras would settle the weakly purely infinite problem [2512.13334]. Second, in the nuclear setting, Robert–Tikuisis show that if \(A\) is separable of finite nuclear dimension then \(A\) nowhere scattered implies the central sequence algebra \(F(A)\) is nowhere scattered, while \(F(A)\) having the Global Glimm Property implies that \(A\) is \(\mathcal Z\)-stable [2512.13334]. The survey therefore identifies the non-simple Toms–Winter conjecture with a Global Glimm statement for central sequence algebras.

Several open questions remain explicit in the recent literature. It is not known whether every \(\mathrm{Cu}(A)\) automatically has property (V), nor whether ideal-filteredness holds for all nowhere scattered algebras [2512.13334]. The soft-operator program adds further questions: whether every strongly soft Cuntz class can be realized by a soft operator, whether abundance of soft elements passes to stabilizations, and whether nowhere scatteredness itself forces abundance of soft elements [2304.11644].

The phrase “Global Glimm Problem” also has a wider terminological range in the arXiv literature. In hyperbolic PDE and gas dynamics it is used for global existence, stability, or uniqueness questions resolved by modified Glimm or Glimm–Lax schemes, including weakly nonlinear gas dynamics [1606.00555], inhomogeneous balance laws [1603.08126], hydrodynamic escape and nozzle flows [1511.00804, 1611.10083], steady Euler multi-wave and conical-shock configurations [1807.06646, 2008.02409], and the uniqueness of Glimm–Lax limits for Sobolev data [2601.01349]. In operator algebra, however, the term has a more specific meaning: the unresolved implication
\[
\text{nowhere scattered}
\Longrightarrow
\text{Global Glimm Property}.
\]
Within that meaning, the current state of the subject is sharply defined: the problem is solved in several regularity classes, especially stable rank one, real rank zero, and topological dimension zero, while the general case remains open and is increasingly organized around Cuntz-semigroup divisibility, ideal-filteredness, property (V), and softness [2204.13059, 2507.16261, 2304.11644, 2512.13334].

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