🔢 mathematics

Metrics on triangulated categories and restrictions of (co)- $t$ -structures

This paper establishes that the restriction of silting-induced $t$ -structures and co- $t$ -structures to completions of triangulated categories is equivalent to the contravariant finiteness of silting subcategories, providing a categorical characterization of right coherent rings and extending Koenig-Yang correspondences to the metric framework while preserving mutation and partial orders.

Original authors: Wei Hu, Ziheng Liu

Published 2026-04-30

📖 4 min read🧠 Deep dive

CC0 1.0

Original authors: Wei Hu, Ziheng Liu

Original paper dedicated to the public domain under CC0 1.0 (http://creativecommons.org/publicdomain/zero/1.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a vast, complex city called Triangulated Category. This city is built on a strange set of rules where objects (like buildings) can be shifted, rotated, and combined in ways that don't quite make sense in our normal world, but they follow a strict internal logic. Mathematicians use this city to study deep patterns in algebra and geometry.

Inside this city, there are two powerful tools used to organize the chaos: t-structures and co-t-structures. Think of these as different ways to draw a map or a grid over the city.

A t-structure is like a vertical grid that slices the city into "layers" (like floors in a skyscraper), allowing mathematicians to find a stable, "flat" ground (called an abelian heart) where normal arithmetic works.
A co-t-structure is like a horizontal grid, offering a different perspective on how these objects relate.

The Problem: The "Metric" Map

Recently, mathematicians discovered a new way to measure distances in this city called a metric. It's not a ruler with inches or centimeters, but a way to say, "These objects are getting closer and closer together as we move through a sequence."

The authors of this paper, Wei Hu and Ziheng Liu, asked a specific question: If we have a perfect map (a t-structure) for the whole city, does that map still work if we only look at a specific neighborhood?

In this city, there are "compact" objects (small, self-contained buildings) and "large" objects (massive skyscrapers made of infinite combinations of the small ones). The authors focused on a special neighborhood called the precompletion and the completion. You can think of these as:

Precompletion: A neighborhood that includes all the small buildings and some of the "almost-finished" large ones.
Completion: The fully finished neighborhood, including every possible limit of those sequences.

The big question was: When does the map (the t-structure) that works for the whole city also work perfectly inside this specific neighborhood?

The Discovery: The "Finite" Rule

The authors found a simple, categorical rule to answer this. They discovered that the map works in the neighborhood if and only if the neighborhood has a property called contravariant finiteness.

The Analogy:
Imagine the neighborhood is a club. The "silting subcategory" is a group of VIP members.

Contravariant finiteness means that for every person in the club, you can find a VIP member who is the "closest match" or the "best approximation" for them.
If every person in the neighborhood has a VIP "best friend" within the VIP group, then the city's map (the t-structure) works perfectly inside the neighborhood.
If there are people in the neighborhood who have no VIP "best friend" to approximate them, the map breaks down, and the structure falls apart.

The Big Result: Coherent Rings

One of the most exciting outcomes of this research is a new way to identify a specific type of ring (a mathematical structure used in algebra) called a right coherent ring.

Usually, identifying these rings requires checking complex algebraic conditions. This paper says:

"A ring is 'right coherent' if and only if the standard map (t-structure) on its derived category restricts perfectly to the neighborhood of bounded complexes of projective modules."

In plain English: You can tell if a ring is "well-behaved" (coherent) simply by checking if the city's map fits inside a specific, smaller neighborhood. If the map fits, the ring is coherent. If it doesn't, it's not.

The Extended Correspondence (The "Koenig-Yang" Connection)

Finally, the paper takes a famous set of connections discovered by Koenig and Yang and extends them into this new "metric" world.

Previously, mathematicians knew that four different things were secretly the same thing, just dressed differently:

Silting Objects: Special "key" buildings.
Co-t-structures: Horizontal grids.
t-structures: Vertical grids.
Simple-minded Collections: A specific set of "simple" buildings.

The authors proved that even when you add the new "metric" rules (the precompletion and completion neighborhoods), these four things are still perfectly linked.

If you change one (like mutating a key building), the others change in a synchronized way.
The order of these structures is preserved.

Summary

In short, this paper builds a bridge between the "big city" of triangulated categories and its "smaller neighborhoods" using a new measuring tool (metrics). They found a simple rule (contravariant finiteness) that tells us exactly when the city's organizational maps work in these neighborhoods. This not only solves a theoretical puzzle but also gives a brand-new, purely structural way to identify "coherent rings," and it confirms that the deep connections between different mathematical structures hold true even in these new, metric-based settings.

1. Problem Statement

The paper addresses two interconnected problems in the theory of triangulated categories, specifically within the context of compactly generated triangulated categories ( $\mathcal{T}$ ) equipped with metrics (a theory developed by Neeman to study completions of categories of compact objects).

Restriction of Silting-Induced t-structures:
In a compactly generated category $\mathcal{T}$ , a silting subcategory $\mathcal{G}$ of compact objects ( $\mathcal{T}^c$ ) induces a canonical t-structure $(\mathcal{T}^{\leq 0}_{\mathcal{G}}, \mathcal{T}^{\geq 0}_{\mathcal{G}})$ . While this t-structure is well-defined on $\mathcal{T}$ , it is not guaranteed to restrict to the precompletion ( $\mathcal{T}^-_c$ ) or the completion ( $\mathcal{T}^b_c$ ) of $\mathcal{T}^c$ with respect to the metric induced by $\mathcal{G}$ . The authors seek a precise characterization of the conditions under which this restriction occurs.
Extension of Koenig-Yang Correspondences:
Koenig and Yang established bijections between silting objects, bounded co-t-structures, bounded t-structures, and simple-minded collections for finite-dimensional algebras. The paper asks whether these correspondences extend to the general metric framework of triangulated categories and whether they commute with mutation operations and preserve partial orders.

2. Methodology

The authors employ a combination of approximation theory, metric completion techniques, and homological algebra within the framework of Neeman's theory of metrics on triangulated categories.

Metric Framework: They utilize the definitions of "good metrics," Cauchy sequences, and the resulting precompletion ( $\mathcal{T}^-_c$ ) and completion ( $\mathcal{T}^b_c$ ) as subcategories of the functor category $\text{Mod-}\mathcal{T}$ .
Approximation Theory: A central tool is the study of contravariant finiteness. The authors analyze when objects in the larger categories admit right approximations by objects in the silting subcategory $\mathcal{G}$ .
Inductive Constructions: They construct specific sequences of triangles (referred to as "slices" and "stacks") to approximate objects in $\mathcal{T}^-_c$ and $\mathcal{T}^b_c$ using objects from $\mathcal{G}$ and its extensions.
Duality and Representability: To establish the Koenig-Yang correspondences, they utilize Neeman's representability theorems and duality arguments (specifically involving the dual of Hom-spaces) to lift simple-minded collections from the completion back to silting objects in the compact subcategory.

3. Key Contributions and Results

A. Characterization of Restriction (Theorem 1.1)

The paper provides a complete characterization of when a silting-induced t-structure restricts to metric completions. Let $\mathcal{G}$ be a silting subcategory of $\mathcal{T}^c$ . The following conditions are equivalent (with implications for the completion):

$\mathcal{G}$ is contravariantly finite in $\mathcal{T}^c$ .
$\mathcal{G}$ is contravariantly finite in the precompletion $\mathcal{T}^-_c$ .
The induced t-structure restricts to a t-structure on $\mathcal{T}^-_c$ .
The induced t-structure restricts to a t-structure on the completion $\mathcal{T}^b_c$ .
Every object in $\mathcal{T}^b_c$ admits a right $\text{add}(\mathcal{G})$ -approximation.

Consequences:

If the restriction holds, the restricted t-structure on $\mathcal{T}^-_c$ is above-bounded, and on $\mathcal{T}^b_c$ it is bounded.
The heart of the restricted t-structure is equivalent to $\text{mod-}\mathcal{G}$ and $\text{mod-}H^0(\mathcal{G})$ .
The intersection of the restricted t-structure defines a co-t-structure on $\mathcal{T}^-_c$ (and $\mathcal{T}^b_c$ if $\mathcal{T}^c \subseteq \mathcal{T}^b_c$ ) with co-heart $\mathcal{G}$ .

B. Categorical Characterization of Right Coherent Rings (Corollary 1.2)

A major application of the restriction theorem is a purely categorical characterization of right coherent rings.

Result: A ring $R$ is right coherent if and only if the standard t-structure on the derived category $D(\text{Mod-}R)$ restricts to the homotopy category of bounded complexes of finitely generated projective modules, $K^{-,b}(\text{proj-}R)$ .
Significance: This generalizes previous results that were limited to tilting complexes over coherent rings, showing that the coherence of the ring is intrinsically linked to the restriction behavior of the standard t-structure.

C. Extended Koenig-Yang Correspondences (Theorem 1.3)

Under the assumption that $\mathcal{T}$ is a compactly generated $k$ -linear category with a silting object $G$ having finite-dimensional Hom-spaces, and that $\mathcal{T}$ is generated by an object in $\mathcal{T}^b_c$ , the authors establish bijections between:

Isomorphism classes of basic silting objects in $\mathcal{T}^c$ .
Bounded co-t-structures on $\mathcal{T}^c$ .
Bounded t-structures on $\mathcal{T}^b_c$ with length hearts.
Equivalence classes of simple-minded collections in $\mathcal{T}^b_c$ .

Key Properties:

These bijections commute with mutation operations (both left and right).
They preserve partial orders defined on these structures (e.g., $M \geq N$ if $\text{Hom}(M, N[m]) = 0$ for $m > 0$ ).
The paper constructs the inverse maps explicitly, particularly showing how to construct a silting object from a simple-minded collection using homotopy colimits and duality, even in the absence of a Nakayama functor.

4. Significance

Unification of Frameworks: The paper successfully bridges the gap between classical ring-theoretic results (like coherence) and the abstract metric theory of triangulated categories. It demonstrates that metric completions are not just technical tools but natural settings for studying structural properties like t-structure restrictions.
Generalization of Classical Results: By extending the Koenig-Yang correspondences to the metric framework, the authors generalize the theory of silting objects and simple-minded collections beyond finite-dimensional algebras to a much broader class of triangulated categories.
New Characterization of Coherence: The result characterizing right coherent rings via the restriction of the standard t-structure offers a new, purely homological perspective on ring theory, independent of module-theoretic definitions of coherence.
Structural Stability: The proof that these correspondences commute with mutations and preserve orders confirms the robustness of the silting theory in metric settings, suggesting that the combinatorial structure of silting objects is preserved even when moving between compact objects and their metric completions.

In summary, this paper provides a rigorous foundation for understanding how silting theory behaves under metric completion, offering powerful new tools for analyzing derived categories and establishing deep connections between ring properties, t-structures, and combinatorial mutations.