A note on pliability and the openness of the multiexponential map in Carnot groups

Frédéric Jean, Mario Sigalotti, Alessandro Socionovo · 2025

Read A note on pliability and the openness of the multiexponential map in Carnot groups by Frédéric Jean, Mario Sigalotti, Alessandro Socionovo (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-13049.txt.

Introduction

Frédéric Jean and two co-authors compare in 2025 rival notions of pliability for horizontal vectors in Carnot groups. Carnot groups are the tangent worlds of sub-Riemannian geometry, where motion is allowed only along certain directions. Several definitions of non-rigid vectors have been proposed lately, driven by questions about monotone sets and Whitney extension.

The note lines the definitions up side by side and tests which imply which. A linked question is whether the multiexponential map, which sends control data to reachable points, is open. Clarifying the web of definitions keeps later theorems from resting on look-alike but unequal concepts, a quiet kind of progress that geometry depends on.

Catalog record

CREATORFrédéric Jean, Mario Sigalotti, Alessandro Socionovo
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE24 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-13049

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A note on pliability and the openness of the multiexponential map in Carnot groups Frédéric Jean, Mario Sigalotti, Alessandro Socionovo Abstract . In recent years, several notions of non-rigidity of horizontal vectors in Carnot groups have been proposed, motivated, in particular, by the characterization of monotone sets and Whitney extension properties. In this note we compare some of these notions, proving that pliability, strong pliability, and the openness property of the multi-exponential map known as (H)-condition are all equivalent. 1. Introduction Consider a Carnot group \mathbb{G} with stratified Lie algebra {\mathfrak{g}}={\mathfrak{g}}_{1}\oplus\dots\oplus{\mathfrak{g}}_{s} (see Definition 2.1 ). The endpoint map E:L^{\infty}([0,1],{\mathfrak{g}}_{1})\to\mathbb{G} (see Definition 2.2 ) plays a crucial role for studying metric and topological properties of \mathbb{G} . The goal of this paper is to explore and clarify the connections among several openness and submersivity properties of the endpoint map and of the multiexponential map (1.1) \Gamma^{(p)}:({\mathfrak{g}}_{1})^{p}\to\mathbb{G},\qquad\Gamma^{(p)}(Y_{1},\dots,Y_{p}):=\exp(Y_{p})\cdots\exp(Y_{1}), where p\in\mathbb{N} and \exp:{\mathfrak{g}}_{1}\to\mathbb{G} is the exponential map, i.e., the restriction of E to {\mathfrak{g}}_{1} , seen as the subspace of L^{\infty}([0,1],{\mathfrak{g}}_{1}) made of constant functions. In particular, given a vector X\in{\mathfrak{g}}_{1} , we are interested in the following properties of the multiexponential map and of E_{X}(\cdot):=E(X+\cdot) , the endpoint map based at X : (H) there exists p\in\mathbb{N} such that the map \Gamma^{(p)} is open at (X,\dots,X)\in({\mathfrak{g}}_{1})^{p} ; (SbH) there exists p\in\mathbb{N} such that the map \Gamma^{(p)} is a submersion at (X,\dots,X)\in({\mathfrak{g}}_{1})^{p} ; (P) the map E is open at X (i.e., E_{X} is open at 0). One of the first appearances of these conditions in the literature is the remark by Cheeger and Kleiner in [ 3 ] that in the three-dimensional Heisenberg group every X\in{\mathfrak{g}}_{1} satisfies condition (SbH) with p=2 . As a consequence, they deduced the following cone property : if a horizontally convex set C contains a point p in its closure and an open ball B , then it also contains the cone over B with vertex at p . In [ 2 ] Arena, Caruso, and Monti generalized the submersivity of the map (Y_{1},Y_{2})\mapsto\exp(Y_{2})\exp(Y_{1}) to the case where \mathbb{G} is of step two of Métivier type and Morbidelli proved in [ 8 ] that the cone property holds for all Carnot groups of step two. Then Montanari and Morbidelli realized in [ 7 ] that the cone property can be obtained assuming only that every X\in{\mathfrak{g}}_{1} satisfies condition (H) (which is clearly implied by (SbH)). In addition, they showed that if a vector X\in{\mathfrak{g}}_{1} satisfies condition (SbH), then the Carnot–Carathéodory distance from the unit element of \mathbb{G} is Pansu differentiable at \exp(X) . (They also showed that (SbH) implies the related and rather involved notion of deformable direction introduced by Pinamonti and Speight in [ 9 ] ). The condition (H) appears again in the paper [ 10 ] by Rigot and it is used to characterise precisely monotone subsets of \mathbb{G} . We adapt the terminology introduced in [ 10 ] , by distinguishing between the (H)-condition and the submersive (H)-condition (SbH) introduced above. On the other hand, the condition (P), called pliability , appeared for the first time in [ 4 ] , where Juillet and the second author proved the C^{1} Whitney extension property for horizontal curves in Carnot groups where every vector in {\mathfrak{g}}_{1} satisfy the property (P) (see also the recent preprint [ 12 ] by Speight and Zimmerman for a generalization to horizontal curves tangent to pliable directions). The first question addressed in this paper is the connection between pliability and (H)-conditions. Later, Sacchelli and the second author introduced in [ 11 ] the notion of strong pliability (SP) , and they improved the results of [ 4 ] for horizontal curves in sub-Riemannian manifolds where every vector in {\mathfrak{g}}_{1} of the tangent Carnot group satisfies (SP). The condition (SP) and the analogous notion of strong (H)-condition (SH) are defined as follows: (SP) for all \eta>0 there exists Y\in{\mathfrak{g}}_{1} such that \|Y\|_{\infty}<\eta , E_{X}(Y)=E_{X}(0) , and the map E_{X} is a submersion at Y ; (SH) for all \eta>0 there exist an integer p\geq 2 and W_{1},\dots,W_{p}\in{\mathfrak{g}}_{1} such that \|W_{i}-X\|<\eta for i=1,\dots,p , \Gamma^{(p)}(W_{1},\dots,W_{p})=\Gamma^{(p)}(X,\dots,X) , and \Gamma^{(p)} is a submersion at (W_{1},\dots,W_{p}) . Finally, note that (SP) holds true in particular for X\in{\mathfrak{g}}_{1} that are regular points of the endpoint map (those X are the ones such that the straight line t\mapsto\exp(tX) is not abnormal). This property motivates a last condition: (Reg) the map E is a submersion at X (i.e., E_{X} is a submersion at 0 ). Our main result is the following. Theorem 1.1 . The following holds: • pliability, strong pliability, (H)-condition, and strong (H)-condition are equivalent; • regularity (Reg) and submersive (H)-condition are equivalent and imply the preceding four conditions. We notice that the equivalence between (H) and submersive (H)-conditions does not hold, see Remark 3.12 . Theorem 1.1 is proved in Section 3 in a slightly more general version, see Theorem 3.4 . Acknowledgments. We are thankful to Séverine Rigot for some useful discussions we had about the topics discussed in this paper. Research funding. This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 101034255. This project has received funding from the Fondazione Ing. Aldo Gini of Padova, Italy. 2. Preliminaries Let us recall some basic facts about Carnot groups, endpoint, and exponential maps. Definition 2.1 . A Carnot group is a nilpotent, connected, and simply connected Lie group \mathbb{G} whose Lie algebra {\mathfrak{g}} admits a stratification, i.e., a direct sum decomposition (2.1) {\mathfrak{g}}={\mathfrak{g}}_{1}\oplus\dots\oplus{\mathfrak{g}}_{s}, such that [{\mathfrak{g}}_{1},{\mathfrak{g}}_{k}]={\mathfrak{g}}_{k+1} for k=1,\dots,s-1 and [{\mathfrak{g}}_{1},{\mathfrak{g}}_{s}]=0 . The left translation by g\in\mathbb{G} is the map L_{g}:\mathbb{G}\to\mathbb{G} defined by L_{g}(x):=gx . Through this map, we have the identification (2.2) (L_{g})_{*}:{\mathfrak{g}}\to T_{g}\mathbb{G},\quad{\mathfrak{g}}\ni X\mapsto(L_{g})_{*}X\in T_{g}\mathbb{G}. Every vector X\in{\mathfrak{g}} defines the left-invariant vector field X(g):=(L_{g})_{*}X . With abuse of notation, we identify the vector and the induced left-invariant vector field, in particular X=X(1_{\mathbb{G}}) . From now on, I:=[0,1]\subset\mathbb{R} denotes the unit interval. A Lipschitz continuous curve \gamma:I\to\mathbb{G} is said to be horizontal if \dot{\gamma}(t)\in{\mathfrak{g}}_{1} for a.e. t\in I , that is, (2.3) (L_{\gamma(t)})^{*}\dot{\gamma}(t)=Y(t)\qquad\mbox{a.e.,} for some function Y\in L^{\infty}(I,{\mathfrak{g}}_{1}) , called the control of \gamma . Definition 2.2 . The endpoint map is the mapping (2.4) E:L^{\infty}(I,{\mathfrak{g}}_{1})\to\mathbb{G},\quad E(Y):=\gamma(1), where \gamma is the unique solution to the Cauchy problem (2.5) \begin{cases}(L_{\gamma(t)})^{*}\dot{\gamma}(t)=Y(t),&t\in I,\\ \gamma(0)=1_{\mathbb{G}}.\end{cases} Given X\in{\mathfrak{g}}_{1} , the endpoint map based at X is the mapping E_{X}:L^{\infty}([0,1],{\mathfrak{g}}_{1})\to\mathbb{G} defined by E_{X}(\cdot):=E(X+\cdot) . In the following, we need the following well known result, which is proved, for instance, in [ 1 , Chapter 8] . Lemma 2.3 . For every X\in{\mathfrak{g}}_{1} , the endpoint map E_{X}:L^{\infty}(I,{\mathfrak{g}}_{1})\to\mathbb{G} is of class C^{1} as soon as L^{\infty}(I,{\mathfrak{g}}_{1}) is endowed with the L^{2} topology. If X\in{\mathfrak{g}} and \gamma:I\to\mathbb{G} is the integral curve of (the left-invariant vector field identified with) X with \gamma(0)=1_{\mathbb{G}} , then the exponential map is defined by (2.6) \exp:{\mathfrak{g}}\to\mathbb{G},\quad\exp(X):=\gamma(1). Notice that \gamma(t)=\exp(tX) . Such a curve is called a straight line . We denote the flow of a left-invariant vector field X by e^{tX}:\mathbb{G}\to\mathbb{G} . We have that e^{tX}(g)=g\exp(tX). For every \lambda\in\mathbb{R} , we define the dilation on the Lie algebra \delta_{\lambda}:{\mathfrak{g}}\to{\mathfrak{g}} as the linear map such that (2.7) \displaystyle\delta_{\lambda}(X)=\lambda^{i}X,\quad\text{for all }X\in{\mathfrak{g}}_{i}. By abuse of notation, we use the

Read the complete text