On permanence of regularity properties II

This paper demonstrates that under the condition of a tracially sequentially-split $*$-homomorphism of order zero, key regularity properties including tracial $m$-comparison, tracial $m$-almost divisibility, and tracial nuclear dimension less than $m$ are preserved from a target unital $C^*$-algebra $B$ to a source algebra $A$.

The Gist

Here is an explanation of the paper "On Permanence of Regularity Properties II" by Hyun Ho Lee, translated into everyday language using analogies.

The Big Picture: The "Family Resemblance" of Math Structures

Imagine you are studying a family of complex, magical structures called C-algebras*. These aren't physical buildings, but abstract mathematical objects used to describe quantum mechanics and other advanced physics.

Mathematicians have been trying to "classify" these structures for decades (like sorting books in a library). To do this, they look for specific "good behavior" traits, which they call regularity properties. Think of these properties as the "personality" of the structure:

1. Comparison: Can you tell which pieces are "bigger" or "smaller" based on how they behave?
2. Divisibility: Can you chop the structure into tiny, equal pieces without breaking it?
3. Nuclear Dimension: How "complicated" or "high-dimensional" is the structure? (Is it a flat sheet of paper, or a 3D cube?)

The Main Question: If you have a complex structure B that has all these "good" traits, and you have a smaller structure A that is connected to B in a specific way, does A automatically inherit those good traits?

This paper says: Yes. If the connection between them is strong enough, the "good behavior" flows from the big structure to the small one.

---

The Bridge: The "Tracially Sequentially-Split" Connection

The paper focuses on a specific type of connection (a mathematical map) between structure A and structure B. The author calls this connection "tracially sequentially-split by order zero." That's a mouthful, so let's break it down with an analogy.

Imagine B is a massive, high-resolution digital image, and A is a smaller, lower-resolution sketch of that image.

• Usually, if you try to reconstruct the big image from the small sketch, you lose details.
• However, this paper describes a special "magic lens" (the map $\phi$) that connects them.
• This lens has a special trick: It can take a tiny piece of the big image (B) and project it back onto the small sketch (A) with almost perfect accuracy, except for a tiny, negligible amount of "noise" or "static" that doesn't matter in the grand scheme of things.

The "Order Zero" part of the name is crucial. In math, "order zero" means the connection preserves orthogonality.

• Analogy: Imagine two people standing back-to-back. They are "orthogonal" (they don't overlap). If you shine a light through the "magic lens" from the big world to the small world, those two people must still be standing back-to-back. They cannot suddenly overlap. This preservation of "non-overlapping" is the key that makes the math work.

The Three "Good Traits" That Pass Through

The paper proves that three specific "good traits" pass from B to A through this magic lens:

1. Tracial m-Comparison (The "Ruler" Analogy)

• The Concept: This is about measuring size. If you have a pile of blocks (positive elements) and you want to know if one pile is smaller than a combination of other piles, you use a "ruler" (called a trace).
• The Result: If the big structure B has a ruler that works perfectly to compare sizes, then the small structure A also gets a working ruler. You can compare sizes in A just as easily as in B.

2. Tracial m-Almost Divisibility (The "Pizza" Analogy)

• The Concept: This asks: "Can I cut this structure into $m+1$ almost-equal slices?"
• The Result: If the big structure B is flexible enough to be sliced into many tiny, equal pieces, then the small structure A is also flexible enough to be sliced the same way. It proves that A is just as "cuttable" as B.

3. Tracial Nuclear Dimension (The "Blueprint" Analogy)

• The Concept: This is the hardest one. It measures how complex the structure is.
  • Imagine you want to build a house. If you can build a perfect replica of the house using only flat, 2D blueprints (finite-dimensional algebras), the house has a low "dimension."
  • If the house is a complex 3D sculpture, it has a high dimension.
• The Result: The paper proves that if the big structure B can be approximated by simple, flat blueprints (low dimension), then the small structure A can also be built using those same simple blueprints.
• The Technical Hurdle: The author had to solve a tricky problem: How do you take the "flat blueprint" from the big world and paste it onto the small world without the pieces overlapping (violating the orthogonality rule)? The paper introduces a sophisticated "lifting" technique (using a mathematical tool called an ultrapower) to ensure the pieces fit perfectly without crashing into each other.

Why Does This Matter?

In the world of mathematics, there is a famous guess called the Toms-Winter Conjecture. It suggests that for these special structures, having a good "ruler" (comparison), being "cuttable" (divisibility), and being "simple" (low dimension) are all the same thing.

This paper is a major step in proving that guess.

• Part 1 of the study (previous work) proved that the "ruler" and "cutability" pass through this connection.
• This paper (Part 2) proves that the "simplicity/low dimension" also passes through.

The Takeaway:
The author has built a unified framework. Whether you are looking at a group of symmetries (like rotating a shape) or a sub-algebra inside a bigger one, if they are connected by this "magic lens," they share the same fundamental "personality." If the big one is well-behaved, the small one is guaranteed to be well-behaved too.

This completes a major chapter in the "classification program," helping mathematicians organize the vast library of these complex structures with much greater confidence.

Technical Summary

Here is a detailed technical summary of the paper "On Permanence of Regularity Properties II" by Hyun Ho Lee.

1. Problem Statement and Context

The paper addresses a critical gap in the classification program for simple nuclear $C^*$-algebras, specifically concerning the Toms-Winter conjecture. This conjecture posits the equivalence of three regularity properties for simple, separable, unital, and nuclear $C^*$-algebras:

1. $\mathcal{Z}$-stability (absorption of the Jiang-Su algebra).
2. Strict comparison of positive elements.
3. Finite nuclear dimension.

While the permanence of $\mathcal{Z}$-stability and strict comparison under structural operations (like crossed products or inclusions) was established using tracially sequentially-split $*$-homomorphisms in previous work (specifically in [8] and [9]), the permanence of nuclear dimension remained an open problem.

The core difficulty lies in the nature of "tracial" splitting. In standard permanence proofs (e.g., Barlak-Szabó), a splitting map allows properties to transfer directly. However, in "tracial" settings (involving the weak tracial Rokhlin property or large subalgebras), the splitting is only approximate in the trace sense. The technical obstacle is transferring the "tracially small" remainder (which is negligible in the trace) to the "norm" part without inflating the topological dimension.

2. Methodology and Framework

The author refines the framework introduced in [9] by imposing a stricter condition on the splitting map: tracial sequential splitting by order zero maps.

• Tracial Sequential Splitting by Order Zero: A $*$-homomorphism $\phi: A \to B$ is tracially sequentially-split by order zero if, for every positive element $z \in A_\omega$ (ultrapower), there exists a completely positive contractive (c.p.c.) order zero map $\psi: B \to A_\omega$ and a nonzero contractive positive element $g \in A_\omega \cap A'$ such that:

  1. $\psi(\phi(a)) = ag$ for all $a \in A$.
  2. $1_{A_\omega} - g \lesssim z$(subequivalent to$z$ in the Cuntz semigroup).

• Role of Order Zero Maps: Order zero maps (introduced by Winter and Zacharias) preserve orthogonality. They are essentially $*$-homomorphisms from the cone $C_0((0,1]) \otimes A$. This structural rigidity is crucial because it allows the preservation of Cuntz semigroup data and orthogonality relations, which are necessary for lifting finite-dimensional approximations without dimension inflation.

• Key Technical Tools:

  • Cuntz Subequivalence ($\lesssim$): Used to compare positive elements and define the Cuntz semigroup $W(A)$.
  • Ultrapowers ($A_\omega$): Used to handle the "tracially small" remainders via the element $g$.
  • Kirchberg $\epsilon$-test (Lemma 3.8): A re-indexing argument allowing the transition from approximate satisfaction of conditions in the ultrapower to exact satisfaction in the sequence level.
  • Orthogonal Lifting (Lemma 3.9): A sophisticated technique to lift a c.p.c. map from the ultrapower $A_\omega$ back to $A$ while maintaining orthogonality with a finite-dimensional approximation.

3. Key Contributions and Results

The paper proves that if $B$ possesses certain regularity properties and $\phi: A \to B$ is tracially sequentially-split by order zero, then $A$ inherits these properties. The main results are:

A. Permanence of Tracial $m$-Comparison (Theorem 3.3)

• Definition: A property where the order of positive elements is determined by their tracial values up to a finite sum of $m$ elements.
• Result: If $B$ has tracial $m$-comparison, then $A$ has tracial $m$-comparison.
• Significance: This establishes the algebraic stability of the comparison property under the new framework, serving as a prototype for the more complex topological results.

B. Permanence of Tracial $m$-Almost Divisibility (Theorem 3.6)

• Definition: A property related to the ability to divide positive elements into $m+1$ almost orthogonal pieces with controlled trace.
• Result: If $B$ is tracially $m$-almost divisible, then $A$ is tracially $m$-almost divisible.
• Significance: This confirms the stability of the algebraic "divisibility" aspect of regularity, which is dual to comparison.

C. Permanence of Tracial Nuclear Dimension (Theorem 3.10)

• Definition: A topological invariant measuring the complexity of approximating the identity map by finite-dimensional algebras, allowing for a "tracially small" error term.
• Result: If $\text{Trdim}_{\text{nuc}}(B) \leq m$, then $\text{Trdim}_{\text{nuc}}(A) \leq m$.
• Technical Achievement: This is the paper's primary contribution. The proof involves:
  1. Constructing approximations in $B$ and pulling them back to $A_\omega$ via the order zero map $\psi$.
  2. Handling the "remainder" term ($1-g$) by showing it is subequivalent to a small positive element$a_0$.
  3. Using orthogonal lifting (Lemma 3.9) to lift the finite-dimensional approximation from the ultrapower to $A$ while ensuring the error term remains orthogonal to the approximation.
  4. Deforming the lifted map to ensure the error term is strictly subequivalent to the target element $a_0$ in the norm topology.

4. Significance and Impact

• Completion of the Tracial Regularity Program: This paper completes the program initiated in [8] and [9]. It demonstrates that all three pillars of the Toms-Winter conjecture ($\mathcal{Z}$-stability, strict comparison, and finite nuclear dimension) are permanent under the unified framework of tracial sequential splitting by order zero.
• Unification of Dynamical Systems and Inclusions: The framework encompasses two major classes of $C^*$-algebraic constructions:
  1. Actions of finite groups with the weak tracial Rokhlin property.
  2. Inclusions of $C^*$-algebras $P \subset A$ with a conditional expectation having the weak tracial Rokhlin property.
• Bridging Algebra and Topology: The work successfully bridges the gap between algebraic invariants (comparison/divisibility) and topological invariants (nuclear dimension). It proves that the "tracially negligible" remainder can be managed rigorously without destroying the topological structure, a feat that was previously elusive.
• Methodological Advancement: The development of the orthogonal lifting technique for order zero maps provides a powerful new tool for future research in the classification of $C^*$-algebras, particularly in scenarios where exact splitting is unavailable.

In summary, Hyun Ho Lee's paper resolves a long-standing technical hurdle in $C^*$-algebra theory, confirming that the regularity properties essential for the classification of nuclear $C^*$-algebras are robust even in the presence of tracially approximate structural operations.