---
title: Reduced Endomorphism K-Theory
url: https://www.emergentmind.com/topics/reduced-k-theory-of-endomorphisms
type: topic
---

# Reduced Endomorphism K-Theory

Searching arXiv for recent and foundational papers on reduced K-theory of endomorphisms.
Reduced K-theory of endomorphisms is a family of constructions that isolate the genuinely endomorphic contribution inside algebraic, topological, or operator-algebraic \(K\)-theory by splitting off a canonical “base” summand. In the algebraic setting, this appears through quotienting endomorphism \(K\)-theory by the image of objects equipped with zero or identity endomorphism, or through categorical reduction by killing endomorphisms factoring through a chosen object [1302.1214], [2507.05956], [1208.1599]. In stable \(\infty\)-categorical formulations, reduced endomorphism \(K\)-theory is naturally expressed as a homotopy fiber or cofiber attached to the forgetful or section functor \(K(\mathrm{End}(\mathcal{C})) \leftrightarrows K(\mathcal{C})\) [2311.13836]. In \(A\)-theory it is identified with the loop of the fiber of a localization sequence for \(\mathbb{N}\)-spaces [1511.08376]. In operator-algebraic contexts, “reduced” instead refers to reduced crossed products or reduced \(C^*\)-algebras, with associated \(K\)-theory computed via Morita equivalence and groupoid methods [2507.20319]. Across these settings, the subject is organized by splittings, localization sequences, and explicit reduction procedures that extract invariants such as Witt vectors, nil-terms, characteristic polynomials, zeta functions, or direct-sum decompositions over simpler rings [1302.1214], [2005.04334], [1604.05673].

## 1. Algebraic reduction and the Chen–Xi framework

A central algebraic model is the reduction of endomorphism rings developed for additive categories and idempotent ideals. Let \(\mathcal{C}\) be an additive category and \(f:Y \to X\) a covariant morphism. Writing \(\operatorname{End}_{\mathcal{C}}(X)\) for the endomorphism ring, one forms the ideal \(I_Y\) generated by endomorphisms of \(X\) that factor through objects in \(\operatorname{add}(Y)\), and defines the quotient ring
\[
\operatorname{End}_{\mathcal{C},Y}(X):=\operatorname{End}_{\mathcal{C}}(X)/I_Y.
\]
This quotient is the reduced endomorphism ring of \(X\) relative to \(Y\), obtained by killing precisely the endomorphisms seen through \(Y\) [1208.1599].

The categorical splitting theorem states that if \(f:Y\to X\) is covariant, then
\[
K_n(\operatorname{End}_{\mathcal{C}}(X\oplus Y))
\cong
K_n(\operatorname{End}_{\mathcal{C},Y}(X))
\oplus
K_n(\operatorname{End}_{\mathcal{C}}(Y))
\]
for all \(n\in \mathbb{N}\), and the body of the paper establishes a homotopy equivalence of \(K\)-theory spaces, yielding the splitting on \(K_0\) as well [1208.1599]. The reduced summand \(K_n(\operatorname{End}_{\mathcal{C},Y}(X))\) is thus a direct factor of the \(K\)-theory of the larger endomorphism ring.

The ring-theoretic counterpart uses an idempotent \(e\in R\) and the ideal \(J=ReR\). If \(J\) is homological and \({}_RJ\) has a finite projective resolution by finitely generated projective \(R\)-modules, then
\[
K_n(R)\cong K_n(R/J)\oplus K_n(eRe)
\]
for all \(n\in\mathbb{N}\) [1208.1599]. This converts computation of \(K\)-groups of \(R\) into computations for a quotient ring and a corner ring. The two constructions are complementary: one reduces endomorphisms by factorization through an object, the other reduces a ring by an idempotent ideal under homological hypotheses [1208.1599].

These decompositions support explicit calculations for standardly stratified rings, hereditary orders, affine cellular algebras, extended affine Hecke algebras of type \(\tilde{A}\), triangular matrix rings, matrix subrings, and Auslander–Reiten situations [1208.1599]. A key limitation is that the homological and finite-type hypotheses are essential: without them, excision may fail and Nil-terms may appear, so the product decomposition need not hold [1208.1599].

## 2. Exact categories, stable \(\infty\)-categories, and reduced spectra

A second major formulation treats endomorphism \(K\)-theory as an invariant of exact or stable \(\infty\)-categories. For a small stable \(\infty\)-category \(\mathcal{C}\), the endomorphism category is \(\operatorname{End}(\mathcal{C})\), concretely the \(\infty\)-category of pairs \((X,f:X\to X)\), and its \(K\)-theory is \(K(\operatorname{End}(\mathcal{C}))\) [2311.13836]. More generally, if \(M:\mathcal{C}\to \operatorname{Ind}(\mathcal{C})\) is a suitable bimodule, one forms the laced category \(\operatorname{Lace}(\mathcal{C},M)\); endomorphisms occur as the Yoneda case \(M=j\) [2311.13836].

In this framework, reduced endomorphism \(K\)-theory is defined at spectrum level as the homotopy fiber of the forgetful map
\[
\widetilde{K}(\operatorname{End}(\mathcal{C}))
:=
\operatorname{fib}\bigl(K(\operatorname{End}(\mathcal{C}))\to K(\mathcal{C})\bigr),
\]
or dually as the cofiber of the section induced by \(X\mapsto (X,\operatorname{id}_X)\) [2311.13836]. The paper explicitly adopts the fiber convention by analogy with cyclic \(K\)-theory for laced categories, while observing that at \(\pi_0\) the cofiber recovers the classical reduced \(K_0\) of endomorphisms modulo summands generated by identity endomorphisms [2311.13836].

The theorem of the heart is decisive here. If \(\mathcal{C}\) carries a bounded weight structure, then
\[
K(\operatorname{End}(\mathcal{C}^{\heartsuit})) \to K(\operatorname{End}(\mathcal{C}))
\]
is an equivalence, and similarly \(K(\mathcal{C}^{\heartsuit})\to K(\mathcal{C})\) is an equivalence [2311.13836]. Consequently, reduced endomorphism \(K\)-theory is invariant under passage to the heart:
\[
\widetilde{K}(\operatorname{End}(\mathcal{C}))
\simeq
\widetilde{K}(\operatorname{End}(\mathcal{C}^{\heartsuit})).
\]
The same statement extends to laced categories with weighted bimodules [2311.13836].

This result places reduced endomorphism \(K\)-theory within a resolution-based formalism. Quillen’s resolution theorem is extended to exact \(\infty\)-categories, resolving and op-resolving inclusions become \(K\)-equivalences, and stable envelopes allow the passage from exact \(\infty\)-categories to stable ones without changing \(K\)-theory in the bounded-heart setting [2311.13836]. A plausible implication is that reduced endomorphism \(K\)-theory is not merely computable from a heart in favorable cases, but structurally controlled by the exact data encoded there.

## 3. Motives, canonical splittings, and Witt vectors

The noncommutative-motivic approach extends \(K\)-theory of endomorphisms from ordinary rings to stable \(\infty\)-categories and identifies it as an additive invariant co-represented by the noncommutative motive of \(\mathbb{S}[t]\), the tensor algebra on the sphere spectrum [1302.1214]. For a small idempotent-complete stable \(\infty\)-category \(\mathcal{C}\),
\[
K\operatorname{End}(\mathcal{C}) \simeq
\operatorname{Map}_{\mathrm{Mot}_{\mathrm{add}}}
\bigl(U_{\mathrm{add}}(\operatorname{Perf}\mathbb{S}[t]),U_{\mathrm{add}}(\mathcal{C})\bigr)
\]
[1302.1214]. This yields a conceptual universal property for endomorphism \(K\)-theory.

The reduced spectrum is defined in this setting as the cofiber of the zero-endomorphism section
\[
\widetilde{K}\operatorname{End}(\mathcal{C})
:=
\operatorname{cofib}\bigl(K(\mathcal{C})\to K\operatorname{End}(\mathcal{C})\bigr),
\]
and on \(K_0\) one obtains a direct sum decomposition
\[
K_0\operatorname{End}(\mathcal{C})\cong K_0(\mathcal{C})\oplus \widetilde{K}_0\operatorname{End}(\mathcal{C})
\]
[1302.1214]. The key structural fact is a coalgebra splitting
\[
U_{\mathrm{add}}(\operatorname{Perf}\mathbb{S}[t])
\simeq
U_{\mathrm{add}}(\operatorname{Perf}\mathbb{S}) \vee W_0,
\]
which induces
\[
K\operatorname{End}(\mathcal{C})
\simeq
K(\mathcal{C}) \times \operatorname{Map}(W_0,U_{\mathrm{add}}(\mathcal{C}))
\]
[1302.1214]. The reduced summand is therefore the Witt-spectrum piece.

For a commutative ring \(A\), this reduced part recovers rational Witt vectors:
\[
\pi_0\operatorname{Map}(W_0,U_{\mathrm{add}}(\operatorname{Perf}HA))
\cong W_0(A),
\]
and hence
\[
K_0\operatorname{End}(A)\cong K_0(A)\oplus W_0(A)
\]
[1302.1214]. The reduced class of an endomorphism is encoded by a determinant-type characteristic series, and the paper emphasizes that rational Witt vectors arise canonically from the symmetric monoidal structure on noncommutative motives [1302.1214].

This framework also classifies natural operations on \(K\operatorname{End}(-)\): \(\pi_0\) of the natural endotransformations is identified with \(\mathbb{Z}\oplus W_0(\mathbb{Z}[t])\), solving a problem raised by Almkvist [1302.1214]. Frobenius and Verschiebung classes appear as specific elements in this algebra of natural operations [1302.1214]. This suggests that reduced endomorphism \(K\)-theory is naturally the receptacle for Witt-type operations even before one imposes commutativity hypotheses.

## 4. Frobenius, Verschiebung, and the twisted noncommutative setting

The 2025 work on Frobenius and Verschiebung develops these operations directly on reduced \(K_0\) of twisted endomorphisms of modules over noncommutative rings [2507.05956]. Here one considers bimodule data \(M,N\) and the exact category \(\operatorname{End}(R,S;M,N)\) of objects
\[
f:M\otimes_R P \to P\otimes_S N
\]
with \(P\) finitely generated projective as an \(S\)-module [2507.05956]. The reduced group is defined by quotienting \(K_0\) by the image of objects with zero endomorphism:
\[
\widetilde{K}_0(R,S;M,N)
=
\operatorname{coker}\bigl(K_0(\operatorname{Mod}^c_{R,S})\to K_0(\operatorname{End}(R,S;M,N))\bigr)
\]
[2507.05956].

In the \(\alpha\)-semilinear case, taking \(M={}_\alpha R\) and \(N=R\) recovers twisted endomorphisms \(f(rp)=\alpha(r)f(p)\) [2507.05956]. Frobenius is defined by iterating the endomorphism and composing through the canonical \(\Gamma\)-functor:
\[
F^n:\widetilde{K}_0(R,S;M,N)\to \widetilde{K}_0(R,S;M^{\otimes_R n},N^{\otimes_S n}),
\]
with \(F^mF^n=F^{mn}\) [2507.05956]. Under the hypothesis \(M=R\) or \(N=S\), Verschiebung is defined in the opposite direction,
\[
V^n:\widetilde{K}_0(R,S;M^{\otimes_R n},N^{\otimes_S n})\to \widetilde{K}_0(R,S;M,N),
\]
and satisfies \(V^mV^n=V^{mn}\) [2507.05956].

The fundamental identity is that \(F^nV^n\) is the transfer for a cyclic \(\mathbb{Z}/n\)-action; in the commutative case this becomes \(n\cdot \mathrm{id}\) [2507.05956]. The iterated trace map
\[
\mathrm{tr}_\bullet:
\widetilde{K}_0(R,S;M,N)\to
\prod_{n\ge 1}\operatorname{Map}_{\mathbb{Z}/n}\bigl(\{M^{\otimes n}\},\{N^{\otimes n}\}\bigr)
\]
produces ghost-component formulas exactly paralleling Witt vector identities [2507.05956]. In the commutative case,
\[
g_r(F_nx)=g_{nr}(x),\qquad
g_r(V_nx)=
\begin{cases}
n\,g_{r/n}(x), & n\mid r,\\
0, & \text{otherwise.}
\end{cases}
\]
[2507.05956].

This work is strictly \(K_0\)-level, and the paper explicitly notes that higher \(K\)-groups are not developed there [2507.05956]. Even so, it gives a precise noncommutative lift of Frobenius and Verschiebung from Witt vectors to reduced endomorphism \(K\)-theory itself. That construction sharpens the earlier motivic picture by making the operations explicit in twisted bimodule categories rather than only in natural-operation form [1302.1214], [2507.05956].

## 5. Trace methods, characteristic polynomials, zeta functions, and \(A\)-theory

Trace-theoretic methods connect reduced \(K\)-theory of endomorphisms to characteristic polynomials, topological restriction homology, zeta functions, and nil-terms. In the Waldhausen setting, the endomorphism category \(\operatorname{End}(\mathcal{C})\) has exact functors including objects with zero endomorphism \(\iota_0\), including objects with identity endomorphism \(\iota_1\), and forgetting the endomorphism [2005.04334]. The paper distinguishes two reductions. For THH and TR, the relevant reduced theory is the cyclic quotient
\[
\widetilde{K}\operatorname{End}(\mathcal{C})
=
K(\operatorname{End}(\mathcal{C}))/\iota_0K(\mathcal{C}),
\]
since the Dennis trace and the TR-trace kill the image of \(\iota_0\) [2005.04334]. For classical reduced \(K_0\) of endomorphisms over a commutative ring \(A\), one instead quotients by \(\iota_1K_0(A)\), the image of identity endomorphisms [2005.04334].

The TR-trace
\[
\mathrm{tr}_{TR}: K\operatorname{End}(A;M)\to TR(A;M)
\]
extends the Lindenstrauss–McCarthy construction from discrete rings to ring spectra and general spectral Waldhausen categories [2005.04334]. On \(\pi_0\), its ghost coordinates are traces of iterates. For a commutative ring \(A\), the identification \(\pi_0TR(A)\cong W(A)\cong (1+tA[[t]])^\times\) implies that the reduced class of \([f:P\to P]\) maps to \(\det(1-tf)\), and the ghost coordinates are \(\operatorname{Tr}(f^n)\) [2005.04334]. The paper states that Almkvist’s characteristic polynomial map is injective on classical reduced \(K_0\), with image the quotients of polynomials, and that the TR-trace recovers this invariant [2005.04334].

In the \(A\)-theoretic setting of spaces, reduced \(K\)-theory of endomorphisms is modeled as the homotopy fiber of the forgetful functor from an endomorphism category over a space \(X\) to the Waldhausen category of retractive spaces [1511.08376]. The main theorem identifies it as
\[
\widetilde{K}(\operatorname{End}_X^S)\simeq
\Omega\,
\operatorname{hofib}\bigl(K(\mathcal{R}_?^S(M))\to K(\mathcal{C}_?(M))\bigr),
\]
a non-linear analogue of Grayson’s theorem [1511.08376]. In the nilpotent case \(S=\{x^n\}\), this recovers the identification of reduced nil-endomorphism \(K\)-theory with the loop of the positive nil-term in \(A\)-theory [1511.08376].

These trace and localization results clarify a common misconception: “reduced” does not have a single universal meaning. Depending on the context, it may mean quotienting by zero endomorphisms, quotienting by identity endomorphisms, taking a homotopy fiber over the forgetful map, or extracting the reduced part of a localized or crossed-product \(K\)-theory spectrum [2005.04334], [2311.13836], [1511.08376].

## 6. Explicit computations and operator-algebraic variants

Several papers provide fully explicit computations. For a field \(k\), let \(E_n\) be the exact category of finite-dimensional \(k[t_1,\dots,t_n]\)-modules, equivalently finite-dimensional \(k\)-vector spaces with \(n\) commuting endomorphisms. Then
\[
K_i(E_n)\cong \bigoplus_{M\in \operatorname{Max}(k[t_1,\dots,t_n])} K_i(k[t_1,\dots,t_n]/M)
\]
for every \(i\ge 0\), and the augmentation at the origin \(M_0=(t_1,\dots,t_n)\) splits off the base \(K_i(k)\)-summand [1604.05673]. The reduced group is therefore
\[
\widetilde{K}_i(E_n)\cong
\bigoplus_{M\neq (t_1,\dots,t_n)} K_i(k[t_1,\dots,t_n]/M)
\]
[1604.05673]. For \(n=1\), the reduced \(K_0\) identifies with the big Witt group, compatibly with the characteristic polynomial description [1604.05673].

The algebraic Chen–Xi framework yields similarly explicit decompositions for endomorphism rings with weak trace, for instance
\[
K_n(\operatorname{End}_R(R\oplus \operatorname{soc}(R)))
\cong
K_n(\operatorname{End}_R(\operatorname{soc}(R)))\oplus K_n(R/\operatorname{soc}(R)),
\]
as well as formulas for triangular matrix rings and standardly stratified rings [1208.1599]. The scope here is broad, but the decompositions depend sharply on covariance and homological ideal hypotheses [1208.1599].

In \(C^*\)-algebraic settings, the phrase “reduced \(K\)-theory of endomorphisms” requires care. For reduced \(C^*\)-algebras of product systems associated to compact-type \(E_0\)-semigroups, the reduced algebra \(C^*_{\mathrm{red}}(X)\) is Morita equivalent to a reduced semigroup crossed product \(K_B(E)\rtimes_{\mathrm{red}} P\), and under suitable functoriality one obtains \(K_*(C^*_{\mathrm{red}}(X))\cong K_*(B\rtimes_{\mathrm{red}}P)\) [2507.20319]. Groupoid models then provide nuclearity, exactness, and homotopy invariance criteria [2507.20319]. The paper explicitly notes that “reduced” there refers to reduced crossed products or reduced \(C^*\)-algebras, not to reduced topological \(K\)-theory \(\widetilde{K}_*\) [2507.20319].

A related but distinct operator-algebraic computation appears for \(C^*\)-algebras generated by endomorphisms and polymorphisms of compact abelian groups. There one computes \(K\)-groups through exact sequences involving the endomorphism \(b(\varphi)\) on \(K^*(H)\), and reduced \(K_0\) is obtained by quotienting by the unit class \([1]\) [1202.5960]. For \(H=\mathbb{T}\) and \(\alpha(z)=z^n\), the paper derives
\[
K_0(A[\varphi])\cong \mathbb{Z}\oplus \mathbb{Z}/(n-1),\qquad K_1(A[\varphi])\cong \mathbb{Z},
\]
and then
\[
\widetilde{K}_0(A[\varphi])\cong \mathbb{Z},\qquad \widetilde{K}_1(A[\varphi])\cong \mathbb{Z}
\]
after modding out the unit class [1202.5960]. This is another instance where “reduced” means quotient by the trivial base contribution, but now in operator \(K\)-theory.

A plausible synthesis is that reduced endomorphism \(K\)-theory is best understood as a strategy rather than a single invariant: one identifies a canonical non-endomorphic contribution—coming from zero endomorphisms, identity endomorphisms, factorization through a chosen object, a unit class, or an ambient category—and then removes or splits it off. The resulting reduced piece is what supports the extra structure: Witt operations, heart invariance, trace formulas, nil-term identifications, direct-sum decompositions, or groupoid-crossed-product computations [2311.13836], [1302.1214], [2507.05956], [1511.08376], [2507.20319].

Source: https://www.emergentmind.com/topics/reduced-k-theory-of-endomorphisms