---
title: 'Lifting Homomorphism: Theory & Applications'
url: https://www.emergentmind.com/topics/lifting-homomorphism
type: topic
---

# Lifting Homomorphism: Theory & Applications

Searching arXiv for recent and foundational papers on lifting homomorphisms across algebra, operator algebras, and topology.
A lifting homomorphism is a homomorphism \(\widetilde f\) that sits above a given homomorphism \(f\) through a quotient, reduction, completion, or covering map, so that the relevant diagram commutes. In categorical language, this is the existence of a diagonal filler in a commutative square, written \(f\perp g\) for left orthogonality; in more concrete settings it appears as lifting a mod-\(p\) representation to Witt vectors, lifting a \(*\)-homomorphism through a quotient \(D\to D/I\), lifting a reduction map on Dade groups, or lifting morphisms between \(K\)-theoretic invariants to honest algebra maps [1707.06615]. The subject is therefore not a single theorem but a family of existence, uniqueness, and obstruction problems whose precise form depends on the ambient category.

## 1. Abstract formulation and categorical setting

The general categorical definition is as follows. For morphisms \(f:A\to B\) and \(g:X\to Y\), one says that \(f\) has the left lifting property with respect to \(g\), denoted \(f\perp g\), if every commutative square admits a diagonal filler \(\ell:B\to X\) such that \(\ell\circ f=u\) and \(g\circ \ell=v\) [1707.06615]. This packages a large class of familiar existence statements into a single orthogonality condition.

In the algebraic examples emphasized in the literature, split homomorphisms are precisely those admitting the appropriate lifting property with respect to initial or terminal maps, projective modules are characterized by the left lifting property against all epimorphisms, injective modules by the right lifting property against all monomorphisms, and free groups are exactly the projective objects in \(\mathbf{Grp}\) [1707.06615]. In this form, “lifting homomorphism” is not merely a constructive device; it is a structural criterion defining classes of objects and morphisms.

A second abstract pattern is the commutative lifting square attached to a surjection or quotient. For a continuous homomorphism \(\rho:\Gamma_F\to \mathrm{GL}_n(k)\), saying that \(\rho\) lifts to \(W_2(k)\) means that there exists \(\widetilde\rho:\Gamma_F\to \mathrm{GL}_n(W_2(k))\) making the diagram commute with the reduction map \(T:\mathrm{GL}_n(W_2(k))\to \mathrm{GL}_n(k)\) [2501.18906]. The same diagrammatic pattern governs \(*\)-homomorphisms \(B\to D/I\) lifted through the quotient \(q:D\to D/I\) in operator algebraic homotopy theory [2508.00125].

This suggests that the modern literature treats lifting homomorphisms in two complementary ways: as a universal orthogonality condition and as a concrete existence problem over a specified reduction or quotient morphism.

## 2. Sections of reduction maps in the Dade group

A particularly explicit lifting theorem occurs in modular representation theory for endo-permutation modules of a finite \(p\)-group \(P\). Let \(O\) be a complete discrete valuation ring of characteristic \(0\) with maximal ideal \(pO\), residue field \(k=O/p\) of characteristic \(p\), and containing all \(\exp(P)\)-th roots of unity. For \(R\in\{O,k\}\), strongly capped endo-permutation \(RP\)-lattices form the Dade group \(D_R(P)\), and reduction modulo \(p\) defines a surjective homomorphism
\[
T_p:D_O(P)\to D_k(P).
\]
Its kernel is isomorphic to \(X(P)\), the group of one-dimensional \(OP\)-lattices [1807.04487].

The lifting problem is whether \(T_p\) admits a section that is itself a group homomorphism. The paper proves that it always does. The construction begins with a strongly capped endo-permutation \(kP\)-module \(M\). Because \(\dim_k M\) is prime to \(p\), there is, among all \(OP\)-lifts of \(M\), a unique lift \(P_M\) whose representation \(P\to \mathrm{GL}_O(P_M)\) has trivial determinant. This lift satisfies \(P_M/pP_M\cong M\), \(\det(g,P_M)=1\) for all \(g\in P\), and is compatible with duals and tensor products:
\[
P_{M^*}\cong (P_M)^*,\qquad
P_{M\otimes_k N}\cong P_M\otimes_O P_N.
\]
These compatibilities are the basis for multiplicativity [1807.04487].

When \(p\) is odd, any permutation \(OP\)-lattice has determinant \(1\), so the assignment
\[
s([M])=[P_M]
\]
is well defined on classes and yields a group homomorphism \(s:D_k(P)\to D_O(P)\) splitting \(T_p\). When \(p=2\), the proof uses the structure theorem
\[
D_k(P)\cong (\mathbb Z/2)^a\times (\mathbb Z/4)^b\times \mathbb Z^c
\]
and explicit generators; if \([M]\) has order \(2\) or \(4\), then \([P_M]\) has the same order, so one can choose lifts of generators of exactly matching orders and extend multiplicatively. In both cases,
\[
T_p\circ s=\mathrm{id},
\qquad
D_O(P)\cong X(P)\times D_k(P)
\]
follows [1807.04487].

The significance of this result is that reduction modulo \(p\) is not only surjective but split by a coherent multiplicative choice of characteristic-\(0\) lifts. The paper further states that blocks of finite groups whose source algebras are described by Dade-group elements can be lifted in families, preserving tensor products and duals [1807.04487].

## 3. Homotopy and asymptotic lifting in \(C^*\)-algebras

In \(C^*\)-theory, lifting homomorphisms is closely tied to homotopy, asymptotic multiplicativity, and extension theory. If \(\phi,\psi:B\to D/I\) are homotopic \(*\)-homomorphisms and \(\psi\) lifts to a discrete asymptotic homomorphism \(\{\Psi_n:B\to D\}_n\), then \(\phi\) also lifts to a discrete asymptotic homomorphism \(\{\Phi_n:B\to D\}_n\), and the entire homotopy lifts as well [2508.00125]. The same paper proves a cp version and a version in which \(\phi\) itself is replaced by an asymptotic homomorphism.

The proof strategy is based on the mapping cylinder
\[
Z_\psi=\{(\eta,b)\in C([0,1],A)\oplus B:\eta(0)=\psi(b)\},
\]
together with quasicentral approximate units, homogeneous relations, and factorization of homotopies through \(Z_\psi\) [2508.00125]. The explicit point is that obstructions to lifting a single \(*\)-homomorphism can be bypassed by passing to discrete or continuous asymptotic lifts.

These lifting theorems have structural consequences. The paper derives that the MF-property is homotopy invariant; that if \(A\) is homotopy dominated by \(B\), one of \(A\) or \(B\) is exact, and every amenable trace on \(B\) is quasidiagonal, then every amenable trace on \(A\) is quasidiagonal; and that under nuclear domination, hyperlinear traces can be shown to be MF [2508.00125].

A parallel development establishes an asymptotic homotopy lifting property (AHLP). If \(A\) is a separable \(C^*\)-algebra that is a sequential inductive limit of semiprojective \(C^*\)-algebras, then for every surjection \(\pi:E\to B\), the pair \((A,\pi)\) satisfies AHLP [2311.06677]. A second theorem shows that AHLP also holds for any separable \(A\) when the extension \(0\to \ker(\pi)\to E\to B\to 0\) is approximately decomposable; every quasidiagonal extension is approximately decomposable, and in the unital case the two notions coincide [2311.06677].

A common misconception is that homotopy lifting in \(C^*\)-theory is controlled only by semiprojectivity and exact lifts. The later results show instead that asymptotic homomorphisms, approximate decomposition, and homotopy domination provide broader lifting mechanisms [2508.00125].

## 4. Lifting \(K\)-theoretic or graph-theoretic data to algebra homomorphisms

For Leavitt path algebras of finite graphs, the lifting problem takes the form of realizing morphisms between invariants by honest graded \(*\)-homomorphisms. Let \(E\) and \(F\) be finite graphs and \(\ell\) a commutative ring with involution. Any pointed, preordered module map
\[
\phi:BF_{\mathrm{gr}}(E)\to BF_{\mathrm{gr}}(F)
\]
between graded Bowen–Franks modules lifts to a unital, graded, diagonal-preserving \(*\)-homomorphism
\[
\Psi:L_\ell(E)\to L_\ell(F),
\]
and the induced map on graded Grothendieck groups agrees with \(\phi\) via the canonical comparison maps [2206.06759]. Over a field, the comparison maps are isomorphisms, so the result establishes the fullness part of Hazrat’s conjecture.

The proof is entirely combinatorial. After expressing \(\phi\) in a filtered-colimit presentation of \(BF_{\mathrm{gr}}(F)\), one produces nonnegative integer matrices satisfying intertwining relations with adjacency matrices, chooses set-theoretic partitions of path spaces, defines bijections encoding edge incidence, and then defines \(\Psi\) on vertices and edges so that the Leavitt relations are satisfied [2206.06759]. The paper further characterizes the maps arising from this construction over \(\mathbb Z\): they are exactly the scalar extensions of unital, graded, \(*\)-homomorphisms preserving a positive-cone \(*\)-semiring \(PC(E)\) [2206.06759].

A related lifting theorem connects graph \(C^*\)-algebras and Leavitt path algebras. If \(\xi:C^*(E)\to C^*(F)\) is a unital \(*\)-homomorphism between simple purely infinite Cuntz–Krieger algebras of finite graphs, then there exists a unital \(*\)-homomorphism \(\Phi:L(E)\to L(F)\) such that the completed map \(\hat\Phi:C^*(E)\to C^*(F)\) is \(C^*\)-homotopic to \(\xi\). Moreover, \(\hat\Phi\) is a \(C^*\)-homotopy equivalence if and only if \(\Phi\) is an algebraic polynomial homotopy equivalence up to \(M_2\)-homotopy [2110.03314].

These results show two distinct lift phenomena. In one direction, order-unit-preserving morphisms of \(K\)-theoretic invariants can be lifted to algebra maps. In the other, analytic \(*\)-homomorphisms between graph \(C^*\)-algebras can be lifted, up to homotopy, to algebraic maps of Leavitt path algebras.

## 5. Obstructions, non-liftability, and arithmetic or topological variants

Not every lifting problem has a positive answer, and much of the theory is organized around explicit obstruction classes. For mod-\(p\) Galois representations, the obstruction to lifting
\[
\rho:\Gamma_F\to \mathrm{GL}_n(k)
\]
to \(\mathrm{GL}_n(W_2(k))\) is the pullback \(\rho^*(\alpha)\in H^2(\Gamma_F,M_n(k))\) of the extension class of
\[
0\to M_n(k)\to \mathrm{GL}_n(W_2(k))\to \mathrm{GL}_n(k)\to 1.
\]
The lifting problem is completely classified: for a fixed field \(F\), a field \(k\) of characteristic \(p>0\), and \(n\ge 1\), every continuous \(n\)-dimensional representation over every extension \(K/F\) lifts if and only if one of the following holds: \(\mathrm{char}\,F=p\), or \(n\le 2\), or \(|k|=2\) and \(n\le 4\) [2501.18906].

In topological bundle theory, if \(P\to X\) is a principal \(K\)-bundle and \(\widehat K\) is a central extension of \(K\) by \(Z\), there is a natural obstruction class \(\delta_1(P)\in \check H^2(X,\underline Z)\) whose vanishing is equivalent to the existence of a \(\widehat K\)-bundle \(\widehat P\) with \(P\cong \widehat P/Z\). When \(Z\) is a quotient of a contractible group by a discrete group \(\Gamma\), the induced homomorphism \(\pi_3(X)\to \Gamma\) is \(\partial_2^{\widehat K}\circ \partial_3^P\); when \(Z\) is discrete, the induced homomorphism \(\pi_2(X)\to Z\) is \(-\partial_1^{\widehat K}\circ \partial_2^P\) [1108.5853]. Here the obstruction is not merely existential; it is made explicit in homotopy-theoretic terms.

A valuation-theoretic form of lifting appears in tropical geometry. For a field \(k\) with real valuation \(\nu\) and a \(k\)-algebra \(R\), there exists a \(k\)-algebra \(K\) with a valuation \(\mu\) extending \(\nu\) such that every real valuation of \(R\) extending \(\nu\) is induced by \(\mu\) via some homomorphism \(R\to K\); \(K\) may be taken to be an algebraically closed field [1304.7726]. When \(\nu\) is trivial and \(R\) is a complete Noetherian local \(k\)-algebra, every local valuation can be matched on finitely many chosen elements by a homomorphism into the Hahn series ring \(\bar k[[t^{\mathbb R}]]\), and every point of the local tropical variety lifts to a \(K\)-point [1304.7726].

The same theme of possible failure appears in \(KK\)-theory and low-dimensional topology. For generalized dimension-drop interval algebras, there exist \(KK\)-elements preserving the Dadarlat–Loring order structure on \(K\)-theory with coefficients but failing to lift to a homomorphism [1306.3623]. For finite regular branched covers of closed oriented \(3\)-manifolds, lifting homeomorphisms along the cover gives a virtual homomorphism between mapping class groups, but unlike the surface case the lifting map is generally not injective for most regular branched covers of \(3\)-manifolds; in the double cover of \(S^3\) branched over the unlink, the kernel has an explicit finite normal generating set [2408.07798].

## 6. Recurring mechanisms and conceptual patterns

Across these settings, several recurrent mechanisms govern lifting homomorphisms. One is the construction of a section to a reduction map by a canonical normalization condition, exemplified by determinant-\(1\) lifts in the Dade group [1807.04487]. Another is passage from exact liftability to asymptotic liftability, which weakens the target condition while preserving enough structure to prove homotopy invariance statements in \(C^*\)-theory [2508.00125].

A third mechanism is lifting from invariants. In graph-algebra contexts, module maps on Bowen–Franks modules or classes in bivariant \(K\)-theory can sometimes be realized by honest homomorphisms after one inserts appropriate combinatorial or homotopy-theoretic data [2206.06759]. A fourth is obstruction theory: extension classes in \(H^2\), Čech obstruction classes, and positivity criteria on \(K\)-theory with coefficients all determine when lifting fails [2501.18906].

The literature also shows that liftability and injectivity are distinct issues. A lift may exist but fail to be unique; a lifting homomorphism may be well defined but have nontrivial or infinite kernel; and positivity on one invariant may be insufficient for actual realizability by homomorphisms [1306.3623]. This suggests that “lifting homomorphism” is best understood not as a single property, but as a spectrum of existence, coherence, and obstruction questions whose answers depend sharply on the algebraic, analytic, or topological structure of the category in which the problem is posed.

Source: https://www.emergentmind.com/topics/lifting-homomorphism