---
title: Liftable Braid Groups
url: https://www.emergentmind.com/topics/liftable-braid-group
type: topic
---

# Liftable Braid Groups

Searching arXiv for recent papers on liftable braid groups and closely related braid lifting constructions.
A **liftable braid group** is, in the most common sense, the subgroup of a braid or mapping class group consisting of braids that admit a lift through a prescribed geometric structure, typically a branched covering or a Lefschetz fibration. In the setting of a genus-\(g\) Lefschetz fibration \(\pi:M\to S^2\) with critical values \(\Delta\), the liftable braid group is the image
\[
\Br(\pi)=\Im\bigl(\Mod(\pi)\to \Mod(S^2,\Delta)\bigr),
\]
also called the braid monodromy subgroup [2510.04389]. In the setting of a branched cover \(\pi:\widetilde\Sigma\to D^2\), a braid \(\beta\in B_n\) is liftable when there exists \(\widetilde\beta\in\Mod(\widetilde\Sigma)\) with \(\beta\circ\pi=\pi\circ\widetilde\beta\) [2508.05146]. These formulations are equivalent in spirit but differ technically: one is an image subgroup of a fiber-preserving mapping class group, while the other is a stabilizer-type subgroup defined by a covering datum.

## 1. Basic definitions and principal frameworks

For a Lefschetz fibration \(\pi:M\to S\) over a connected oriented surface \(S\) with singular values \(\Delta\subset \mathrm{int}(S)\), one considers
\[
\Diff^+(\pi)=\{F\in \Diff^+(M)\mid F \text{ preserves fibers of }\pi,\ F|_{\pi^{-1}(\partial S)}=\Id\},
\]
and its mapping class group
\[
\Mod(\pi)=\pi_0(\Diff^+(\pi)).
\]
Tracking how a fiber-preserving diffeomorphism permutes the fibers yields a natural homomorphism
\[
\Mod(\pi)\longrightarrow \Mod(S,\Delta).
\]
When \(S=S^2\) and \(|\Delta|=n\), one identifies \(\Mod(S^2,\Delta)\) with the spherical braid group \(\Br_n(S^2)\), and the image is denoted \(\Br(\pi)\) [2510.04389].

For a simple \(d\)-fold branched cover \(\pi:\widetilde\Sigma\to D^2\) with branch values \(\{p_1,\dots,p_n\}\), “simple” means that each branch point has local monodromy a single transposition in \(S_d\). The mapping class group of the \(n\)-marked disc is the Artin braid group \(B_n\). A braid \(\beta\in B_n\) is liftable if there exists \(\widetilde\beta\in\Mod(\widetilde\Sigma)\) such that
\[
\beta\circ\pi=\pi\circ\widetilde\beta.
\]
In general, the lifting homomorphism is defined only on a proper subgroup of \(B_n\) [2508.05146].

A third formulation appears for branched coverings \(\chi:\Sigma\to S\) of closed orientable surfaces. If \(\mu:\pi_1(S\setminus B)\to S_n\) is the monodromy, then \(\chi\) is liftable to a braided surface in \(S\times\mathbb R^2\) precisely when \(\mu\) admits a lift \(\widetilde\mu:\pi_1(S\setminus B)\to B_n\) such that \(p_n\circ\widetilde\mu=\mu\) and each meridian is sent to a non-trivial completely splittable braid [2004.09174].

The terminology is therefore uniform only at a high level: “liftable” always means compatibility with a covering or fibration, but the ambient group and the precise criterion depend on the geometric category.

## 2. Monodromy, Hurwitz action, and stabilizer descriptions

A fundamental structural theorem identifies liftable braid groups with stabilizers of monodromy data. For a Lefschetz fibration \(\pi:M\to S\), let
\[
\phi_\pi:\pi_1(S\setminus \Delta)\longrightarrow \Mod(\Sigma_g)
\]
be the monodromy representation. Under the Hurwitz action of \(\Mod(S,\Delta)\) on such representations, Moishezon and Matsumoto show
\[
\Br(\pi)=\Stab(\phi_\pi)\subset \Mod(S,\Delta).
\]
When \(S=S^2\), a braid lies in \(\Br(\pi)\) if and only if it preserves the conjugacy class of the factorization
\[
T_{\alpha_1}\cdots T_{\alpha_n}=\Id\subset \Mod(\Sigma_g)
\]
up to Hurwitz moves [2510.04389].

For simple branched covers of the disc, Licata and Vértési formulate the analogous structure in terms of a coloured braid groupoid. An object is a labeling
\[
\tau=(t_1,\dots,t_n)\in (S_d)^n
\]
by transpositions whose product equals a fixed total monodromy \(\mu\in S_d\). Morphisms are coloured braids acting on labelings by the Hurwitz rule
\[
\sigma_i:(t_1,\dots,t_i,t_{i+1},\dots,t_n)\longmapsto
(t_1,\dots,t_i t_{i+1} t_i,\ t_i,\dots,t_n).
\]
The classical liftable braid group attached to a fixed cover is then the endomorphism group
\[
\Hom_{\ColB_{n,d}}(\tau,\tau)=\{\beta\in B_n\mid \beta\cdot\tau=\tau\},
\]
so classical liftability is exactly invariance of the branch-label tuple under Hurwitz action [2508.05146].

These two descriptions—stabilizer of Lefschetz monodromy and stabilizer of branch-label data—are formally parallel. In both cases, liftability is not primarily a local condition on a braid word; it is a global invariance condition on monodromy data.

## 3. The lifting homomorphism and its extensions

For simple branched covers of the disc, the classical lifting homomorphism sends a liftable braid to the induced mapping class on the covering surface. Licata and Vértési extend this construction from a subgroup to a groupoid-valued functor. They first define a graphical groupoid \(\G(\mu)\) built from canonical systems of lifted arcs and then construct
\[
\phi:\ColB_{n,d}\longrightarrow \G(\mu),\qquad \phi(\sigma_i)=\widetilde\sigma_i.
\]
Using an Alexander-method argument, they define a unique mapping class
\[
\Phi(\beta):\widetilde\Sigma_\tau\to \widetilde\Sigma_{\tau'}
\]
for each coloured braid \(\beta:\tau\to\tau'\), satisfying
\[
\pi_{\tau'}\circ \Phi(\beta)=\beta\circ \pi_\tau.
\]
Thus every coloured braid is liftable as a morphism between possibly distinct covering surfaces, and the classical case is recovered precisely on endomorphisms of a fixed object [2508.05146].

This groupoid extension clarifies a recurrent source of confusion. A braid may fail to be classically liftable not because no lift exists in any sense, but because it changes the branch-labeling object. The obstruction is therefore often object preservation, not existence of a geometric lift per se.

The broader braided-surface criterion of Funar and Pagotto places this in an algebraic framework. For a degree-\(n\) branched covering \(\chi:\Sigma\to S\) with monodromy \(\mu:\pi_1(S\setminus B)\to S_n\), liftability is equivalent to a braid-group lift of \(\mu\) compatible with local branching. In that setting, the exact sequence
\[
1\to P_n\to B_n\to S_n\to 1
\]
provides the basic algebraic interface between permutation monodromy and braid lifts [2004.09174].

## 4. Size, index, and low-genus behavior

Recent work shows that liftable braid groups are frequently of infinite index inside the ambient braid group. Jackson proves that for a genus-\(g\) Lefschetz fibration \(\pi:M\to S^2\), the index
\[
[\Mod(S^2,\Delta):\Br(\pi)]
\]
is infinite in three major cases: when \(g=1\); when \(g\ge 2\) and \(\pi\) is expressible as a self-fiber sum; and when \(\pi\) is a holomorphic genus-\(2\) Lefschetz fibration whose vanishing cycles are nonseparating [2510.04389].

In genus \(1\), the argument can be expressed via the \(\SL_2\)-character variety
\[
Y_\Delta(\mathbb C)=\Hom(\pi_1(S^2\setminus \Delta),\SL_2(\mathbb C))\sslash \SL_2(\mathbb C).
\]
Because each local monodromy is an infinite-order Dehn twist, Lam–Landesman–Litt’s classification implies that the \(\Mod(S^2,\Delta)\)-orbit is infinite as soon as \(|\Delta|>6\), hence the liftable subgroup has infinite index [2510.04389].

The self-fiber-sum case relies on Auroux’s fibre-sum criterion. For \(\pi\#\pi\), one obtains
\[
\bigl[\Mod(S^2,\Delta_{\pi\#\pi}):\Br(\pi\#\pi)\bigr]=\infty,
\]
and in particular the index is infinite whenever \(\pi\) is nontrivial and \(g\ge 2\) [2510.04389].

There are also finite-index exceptions in low complexity. For the disk fibrations
\[
q_n:M_n\to D^2
\]
with alternating monodromy, Jackson computes
\[
[B_n:\Br(q_n)]<\infty \Longleftrightarrow n\le 4,
\]
with explicit values
\[
[B_2:\Br(q_2)]=3,\qquad [B_3:\Br(q_3)]=8,\qquad [B_4:\Br(q_4)]=27.
\]
For \(n\ge 5\), the index is infinite [2510.04389].

These results indicate that “most” braids are typically not liftable for a fixed geometric structure. A plausible implication is that liftable braid groups should usually be regarded as rigid monodromy stabilizers rather than as large ambient subgroups.

## 5. Hyperelliptic lifts, differentiable obstructions, and cohomology

The standard hyperelliptic embedding
\[
\varepsilon:\Br_{2g+2}\hookrightarrow \Mod_{g,2}
\]
provides a closely related but distinct lifting problem. It is obtained from the double-branched-cover construction
\[
\Sigma_{\mathbf z}=\{(w,t)\in \mathbb C\times D^2\mid w^2=\prod_{i=1}^{2g+2}(t-z_i)\},
\]
or equivalently by sending braid generators to Dehn twists along a standard \(A_{2g+1}\)-chain of curves [1511.09369].

Nariman shows that for \(g>1\) there is no group homomorphism
\[
\widetilde\varepsilon:\Br_{2g+2}\to \Diff(S_{g,2})
\]
lifting \(\varepsilon\). The obstruction uses the fact that the commutator subgroup of \(\Br_{2g+2}\) is finitely generated and perfect, while Thurston stability forces a nontrivial finitely generated subgroup of the \(C^1\)-stabilizer of a boundary point to surject onto \(\mathbb Z\) [1511.09369].

At the same time, there is no primary cohomological obstruction of the expected kind. Nariman proves that for every \(n\) and every abelian group \(A\),
\[
H^*(\Br_n;A)\longrightarrow H^*(\Diff^\delta(D^2_n,\partial D^2);A)
\]
is split-injective, and hence the cohomology of \(\Br_n\) appears as a direct summand in that of the discrete diffeomorphism group of the punctured disk [1511.09369].

This is one of the main conceptual distinctions in the subject. Cohomological liftability, categorical liftability, and honest differentiable liftability are separate notions. The failure of a \(\Diff\)-valued lift does not imply failure of cohomological splitting, and conversely the existence of a cohomological section does not produce a genuine geometric lift.

## 6. Related lifting constructions and broader mathematical context

The notion of lifting braids extends beyond coverings of surfaces. In the symplectic and contact setting, Casals and Murphy construct an embedding
\[
\rho_{\Symp}:B_{m+1}\longrightarrow \pi_0\Symp_c(Q,\omega_Q)
\]
for the \(A_m\)-Milnor fiber \(Q\), and combine it with a natural lifting homomorphism
\[
\ell:\pi_0\Symp_c(Q)\to \pi_0\Cont_c(Q\times S^1)
\]
defined by
\[
\phi^*\theta_Q=\theta_Q+dH_\phi,\qquad \widehat\phi(p,z)=(\phi(p),z+H_\phi(p)).
\]
The composition
\[
\rho_{\Cont}=\ell\circ\rho_{\Symp}:B_{m+1}\to \pi_0\Cont_c(Q\times S^1)
\]
remains injective, yielding a braid-group embedding into a contact mapping class group [1511.00879]. Although this is not the standard “liftable braid group” of a cover or fibration, it is a closely related lifting mechanism in which braid-theoretic generators persist under passage to a richer geometric category.

A different extension appears in loop-braid theory. Damiani, Faria Martins, and Martin construct an injection of the extended loop braid group into the automorphism group of a \(\pi\)-module \(\mathbb M_n=(F_n,M_n,\triangleright)\), extending Artin’s classical representation. In their formulation, an ordinary braid is “liftable” to a loop-braid automorphism precisely when its action preserves the \(\pi_2\)-module structure, equivalently when the element
\[
M=K_1+\cdots+K_n
\]
is invariant [1912.11898].

Taken together, these developments show that liftability is not a single theorem but a recurring structural theme. It links braid groups to monodromy factorizations, branched covers, mapping class groups, symplectic and contact topology, and higher homotopical automorphism data. The central invariant throughout is compatibility: a braid is liftable exactly when it preserves the geometric or algebraic datum that defines the covering, fibration, or enhanced target category.

Source: https://www.emergentmind.com/topics/liftable-braid-group