Abstract
In the infinite dimensional Heisenberg group, we construct a left invariant weak Riemannian metric that gives a degenerate geodesic distance. The same construction yields a degenerate sub-Riemannian distance. We show how the standard notion of sectional curvature adapts to our framework, but it cannot be defined everywhere and it is unbounded on suitable sequences of planes. The vanishing of the distance precisely occurs along this sequence of planes, so that the degenerate Riemannian distance appears in connection with an unbounded sectional curvature. In the 2005 paper by Michor and Mumford, this phenomenon was first observed in some specific Fréchet manifolds.
Similar content being viewed by others
1 Introduction
Geodesic distances naturally appear in the geometry of infinite dimensional manifolds. A new aspect is that they may also vanish on distinct points. In general, the vanishing of the geodesic distance may occur for certain Riemannian metrics, where no special conditions are assumed, namely for weak Riemannian metrics, [1, Definition 5.2.12]. These metrics are important, since they are the only possible metrics when the manifold is not modelled on a Hilbert space.
Vanishing geodesic (or Riemannian) distances in Fréchet manifolds were first found in [10, 19, 20] and further examples were studied in [3,4,5, 8, 14, 15]. A simple example of vanishing geodesic distance can be also constructed in a Hilbert manifold, [17]. However, one may still wonder whether replacing a weak Riemannian metric with a left invariant weak Riemannian metric with respect to a Hilbert Lie group structure might give a condition to have positive geodesic distance on distinct points.
The answer to this question does not seem intuitively clear. For instance, we observe that connected, simply connected and commutative Banach Lie groups, equipped with a bi-invariant weak Riemannian metric have positive geodesic distance on distinct points. In short, their geodesic distance is actually a distance. The proof of this fact essentially follows from [22, Proposition IV.2.7], observing that the exponential mapping is a local Riemannian isometry.
Thus, the question is whether considering a left invariant weak Riemannian metric on a noncommutative, connected and simply connected Banach Lie group may prevent the vanishing of the geodesic distance. Our first result answers this question in the negative.
Theorem 1.1
There exists a left invariant weak Riemannian metric on the infinite dimensional Heisenberg group \(\mathbb {H}\), whose Riemannian distance is not positive on all couples of distinct points.
In analogy with the finite dimensional case, the infinite dimensional Heisenberg group \(\mathbb {H}\) can be defined starting from the Heisenberg Lie algebra, due to the Baker–Campbell–Hausdorff formula. We use \(\ell ^2\times \ell ^2\times {\mathbb {R}}\) to model \(\mathbb {H}\) as a Hilbert manifold, where \(\ell ^2\) denotes the standard linear space of real square-summable sequences. The Lie product associated to \(\mathbb {H}\) is defined in (1). More information on the Lie group \(\mathbb {H}\) is provided in Sect. 2.1.
Sub-Riemannian distances naturally appears also in infinite dimensional manifolds. We mention for instance [9, Theorem C.2], where strong sub-Riemannian metrics were considered in connection with Wiener spaces, lower bounds on the Ricci curvature and logarithmic Sobolev inequalities. In this case, the space of admissible velocities is strictily contained in the tangent space and we have an a priori smaller family of connecting curves. The possible vanishing of a sub-Riemannian distance between some distinct points of an in infinite dimensional manifold was explicitly mentioned in [12, Remark 2]. The following theorem seemingly provides a first example of such vanishing phenomenon for an infinite dimensional sub-Riemannian manifold.
Theorem 1.2
There exists a left invariant weak sub-Riemannian metric on the infinite dimensional Heisenberg group \(\mathbb {H}\) such that its associated sub-Riemannian distance is not positive on all couples of distinct points.
The sub-Riemannian metric and the sub-Riemannian distance that we consider on \(\mathbb {H}\) are described in Sect. 2.3. Both Theorem 1.1 and Theorem 1.2 are contained in Theorem 3.3 and their proof rather surprisingly relies on the same sequence of length-minimizing curves. The proof of these results shows that both the Riemannian and sub-Riemannian distance are vanishing between points that have the same projection on the subspace \(\ell ^2\times \ell ^2\times \{0\}\). Remark 3.2 completes the picture, showing that when the projections of two points on \(\ell ^2\times \ell ^2\times \{0\}\) are different, then both their Riemannian and sub-Riemannian distance are positive. In sum, all distinct points with vanishing geodesic distance are characterized.
From another perspective, dealing with a left invariant weak Riemannian metric has the advantage to find the sectional curvature by more manageable formulas. In [19], Michor and Mumford proved that in different Fréchet manifolds with a vanishing geodesic distance the sectional curvature is unbounded. Theorem 1.3 below presents the same phenomenon for the left invariant weak Riemannian metric \(\sigma \) defined in (10), in the infinite dimensional Heisenberg group \(\mathbb {H}\).
From the standard formula for the sectional curvature of Lie groups, see for instance [2] and [7], the sectional curvature of \(\mathbb {H}\) with respect to \(\sigma \) can be defined on “many planes” of the Lie algebra \({{\,\textrm{Lie}\,}}(\mathbb {H})\). We wish to point out that for general weak Riemannian metrics the existence of the Levi–Civita (and then of the sectional curvature) is not guaranteed a priori. An example of this fact can be found in [6], where more information on the problem is available. We also observe that the “finite dimensional formula” for the sectional curvature through the structure coefficients of \({{\,\textrm{Lie}\,}}(\mathbb {H})\), [21, Lemma 1.1], converges on the previous planes to the same sectional curvature obtained by [2, Theorem 5]. Broadly speaking, we may think of the convergence of the sectional curvature in Milnor’s paper [21] as a computation of the sectional curvature of \(\mathbb {H}\) through a finite dimensional approximation by an orthonormal basis. On the other side, we also observe that this convergence does not hold on all 2-dimensional subspaces of \({{\,\textrm{Lie}\,}}(\mathbb {H})\), as shown in Remark 4.1.
The next theorem shows that the sectional curvature with respect to the weak Riemannian metric \(\sigma \) is unbounded on a certain sequence of planes.
Theorem 1.3
Let \(\mathbb {H}\) be the infinite dimensional Heisenberg group equipped with the left invariant weak Riemannian metric \(\sigma \). Then there exists two sequences of orthonormal vectors \(a_{1j},a_{2j}\in {{\,\textrm{Lie}\,}}(\mathbb {H})\) and \(b\in {{\,\textrm{Lie}\,}}(\mathbb {H})\) with \(j\ge 1\) such that \(K_\sigma (a_{1j},b)=K_\sigma (a_{2j},b)\),
The numbers \(K_\sigma (a_{1j},a_{2j})\) and \(K_\sigma (a_{1j},b)\) are the sectional curvatures of the planes of \({{\,\textrm{Lie}\,}}(\mathbb {H})\) spanned by the orthonormal bases \((a_{1j},a_{2j})\) and \((a_{1j},b)\).
The proof of this theorem is provided in Sect. 4, where also more information on the vectors \(a_{1j},a_{2j}\) and b can be found. Inspecting the proofs of Theorem 3.3 and Theorem 1.3 another interesting phenomenon appears. The curves whose lengths converge to zero and that connect two distinct points are precisely contained in the span of the planes where the sectional curvature blows-up. We finally mention Proposition 4.2, where we prove that the sectional curvature is unbounded also on some sequences of converging planes.
2 Preliminary notions
In this section, we present an infinite dimensional version of the classical Heisenberg group equipped with either a Riemannian or a sub-Riemannian structure. The construction is well known and it has connections with different areas of Mathematics.
2.1 A short introduction to the infinite dimensional Heisenberg group
We denote by \(\ell ^2\) the linear space of all real and square summable sequences. Its scalar product \(\langle \cdot ,\cdot \rangle \) has the associated norm \(\Vert x\Vert =\sqrt{\sum _{j=1}^\infty x_j^2}\) for any element \(x=\sum _{j=1}^\infty x_je_j\). The set of unit vectors \(\{e_j:j\ge 1\}\) denotes the canonical orthonormal basis of \(\ell ^2\). For each integer \(n\ge 1\), the element \(e_n\) of \(\ell ^2\) has n-th entry equal to 1 and all the others are zero.
We consider \(\ell ^2 \times \ell ^2\times {\mathbb {R}}\) endowed with its standard structure of product of Hilbert spaces and we introduce the continuous and skew-symmetric function
defined on \((\ell ^2\times \ell ^2)\times (\ell ^2\times \ell ^2)\). Then we introduce a continuous Lie product on \(\ell ^2\times \ell ^2\times {\mathbb {R}}\):
that makes \(\ell ^2\times \ell ^2\times {\mathbb {R}}\) an infinite dimensional Lie algebra. Due to the Baker–Campbell–Hausdorff formula (in short BCH), we equip \(\ell ^2\times \ell ^2\times {\mathbb {R}}\) with a noncommutative and analytic Lie group operation:
for all elements \((h_1,h_2,\tau ),(h_1',h_2',\tau ')\in \ell ^2\times \ell ^2\times {\mathbb {R}}\). We denote by \(\mathbb {H}\) the Hilbert Lie group arising from the previous group operation, that is the infinite dimensional Heisenberg group modelled on the Hilbert space \(\ell ^2\times \ell ^2\times {\mathbb {R}}\).
From the viewpoint of Mathematical Physics, the group \(\mathbb {H}\) naturally appears in the theory of representations of infinite dimensional Lie algebras, see [13] and the references therein. In the theory of infinite dimensional Lie groups, \(\mathbb {H}\) can be seen as a special instance of more general BCH-Lie groups. They are infinite dimensional Lie groups with a local exponential mapping that is also a bianalytic diffeomorphism around the origin, [11, 22]. For infinite dimensional nilpotent Lie algebras, the BCH formula defines a global group operation, [23]. This viewpoint was followed in [16] to define infinite dimensional Banach homogeneous groups as suitable direct sums of Banach spaces, equipped with an analytic structure and an everywhere converging BCH formula. In the same work, several examples of Banach homogeneous groups were provided. We mention infinite producs of Engel groups using either \(\ell ^p\) or \(L^p\) spaces, an infinite product of Heisenberg groups modelled on \(\ell ^2\times \ell ^1\) and other analogous analytic constructions.
In relation to the understanding of hypoellipticity in infinite dimensions [18], infinite dimensional Heisenberg-like groups based on a Wiener space, along with their Brownian motion were introduced and studied in [9]. In this connection, also some Ricci curvature lower bounds are obtained, using a left invariant Riemannian metric.
2.2 Weak Riemannian metrics on \(\mathbb {H}\)
For each \(p\in \mathbb {H}\), we denote by \(L_p:\mathbb {H}\rightarrow \mathbb {H}\) the left multiplication by p, defined as \(L_p(r)=p\cdot r \) for all \(r\in \mathbb {H}\). The group operation (2) gives a simple formula for the differential of \(L_p\) at a point q, namely
for every \(v=(v_1,v_2,v_3) \in T_q\mathbb {H}\), with \(p=(p_1,p_2,\tau )\). Notice that we have identified \(T_q\mathbb {H}\) with \(\mathbb {H}\), using the Hilbert space structure \(\mathbb {H}\). We also notice that our formula for the differential \((dL_p)_q\) does not depend on the point q. We consider a scalar product
on the tangent space \(T_0\mathbb {H}\) of \(\mathbb {H}\) at the origin, which is continuous with respect to the product topology of \(T_0\mathbb {H}\times T_0\mathbb {H}\).
Then for every \(p \in \mathbb {H}\) and \(v,w \in T_p\mathbb {H}\) the following scalar product
defines a left invariant weak Riemannian metric \(\sigma \)on \(\mathbb {H}\). The associated Riemannian norm is denoted by \(\Vert \cdot \Vert _\sigma \).
If for any piecewise smooth curve \(\gamma :[0,1]\rightarrow \mathbb {H}\) we define its Riemannian length as
then the associated geodesic distance \(d: \mathbb {H}\times \mathbb {H}\rightarrow [0,+ \infty )\) between \(p,q\in \mathbb {H}\) is
Clearly d is left invariant, symmetric and it satisfies the triangle inequality.
2.3 Weak sub-Riemannian metrics on \(\mathbb {H}\)
Identifying \(\mathbb {H}\) with \(T_0\mathbb {H}\), the set \(\ell ^2 \times \ell ^2 \times \{0\}\) can be seen as a closed subspace of \(T_0\mathbb {H}\), that we denote by \(H_0\mathbb {H}\). We may obtain a left invariant horizontal subbundle, denoted by \(H\mathbb {H}\), introducing the fibers
for every \(p=(p_1,p_2,\tau )\in \mathbb {H}\). We note that \(v=(v_1,v_2,v_3) \in H_p\mathbb {H}\) if and only if
and the previous condition corresponds to the equality
We have a precise formula to define the horizontal curves associated to \(H\mathbb {H}\). They are continuous and piecewise smooth curves \(\gamma :[0,1] \rightarrow \mathbb {H}\) of the form \(\gamma =(\gamma _1,\gamma _2,\gamma _3)\in \mathbb {H}\), such that for almost every \(t \in [0,1]\) we have
The previous differential constraint means that \(\dot{\gamma }(t) \in H_{\gamma (t)}\mathbb {H}\).
On the horizontal fibers \(H_p\mathbb {H}\) of \(H\mathbb {H}\) we can fix a scalar product. A left invariant weak sub-Riemannian metric g on \(H\mathbb {H}\) is defined by a continuous inner product
so that for all \(p \in \mathbb {H}\) and \(v,w \in H_p\mathbb {H}\) we have
The associated sub-Riemannian norm is denoted by \(\Vert \cdot \Vert _g\) and the length of a horizontal curve \(\gamma :[0,1] \rightarrow \mathbb {H}\) is defined by
For any couple of points in \(\mathbb {H}\), it is easy to construct a piecewise smooth horizontal curve that connects them, hence the following sub-Riemannian distance
is finite for every couple of points \(p,q\in \mathbb {H}\), hence we have \(\rho : \mathbb {H}\times \mathbb {H}\rightarrow [0,+ \infty )\). One immediately notices that \(\rho \) is left invariant, symmetric and it satisfies the triangle inequality.
3 Degenerate geodesic distances in the infinite dimensional Heisenberg group
This section is devoted to the construction of special left invariant weak Riemannian and sub-Riemannian metrics that yield degenerate geodesic distances.
We introduce the linear and continuous operator \(A:\ell ^2 \rightarrow \ell ^2\), which associates to each \(x\in \ell ^2\) of components \((x_k)_{k\ge 1}\) the element \(Ax \in \ell ^2\), whose k-th component is \((Ax)_k= x_k/k\). Then we define the scalar product \(\eta : \ell ^2 \times \ell ^2 \rightarrow {\mathbb {R}}\) as
for all \(v,w\in \ell ^2\) and its associated norm
We use \(\eta \) to define the new scalar product
for every \((v_1,v_2),(w_1,w_2)\in \ell ^2\times \ell ^2\). By our identification, \(g_0\) can be seen as a scalar product on \(H_0\mathbb {H}\), so that using (6) we obtain a left invariant weak sub-Riemannian metric g on \(\mathbb {H}\). We follow the notation of the previous section, denoting by \(\rho \) the special sub-Riemannian distance associated to this choice of g through formula (7).
To obtain a left invariant weak Riemannian metric \(\sigma \) on \(\mathbb {H}\), we extend \(g_0\) as follows
for every \((v_1,v_2,v_3),(w_1,w_2,w_3)\in T_0\mathbb {H}\), where \(\sigma _0:T_0\mathbb {H}\times T_0\mathbb {H}\rightarrow {\mathbb {R}}\). From (3), the scalar product in (10) immediately defines a left invariant weak Riemannian metric \(\sigma \) on \(\mathbb {H}\). The Riemannian distance associated to \(\sigma \) through (4) will be denoted by d.
Remark 3.1
Let us consider the sub-Riemannian metric g, the Riemannian metric \(\sigma \) and their associated geodesic distances d and \(\rho \) defined above (the Riemannian distance and the sub-Riemannian distance, respectively). The family of piecewise smooth curves connecting two points also contains the horizontal curves connecting the same points. Since the restriction of \(\sigma \) to the horizontal subbundle \(H\mathbb {H}\) coincides with g, the infimum defining d is taken over a larger family, hence \(d\le \rho \).
Remark 3.2
It is easy to notice that both d and \(\rho \) are not everywhere vanishing on \(\mathbb {H}\). We consider \((p_1,p_2,\tau ), (q_1,q_2,s)\in \mathbb {H}\) with \((p_1,p_2) \ne (q_1,q_2)\) and we choose any piecewise smooth curve \(\gamma =(\gamma _1,\gamma _2,\gamma _3):[0,1] \rightarrow \mathbb {H}\) with \(\gamma (0)=(p_1,p_2,\tau )\) and \(\gamma (1)=(q_1,q_2,s)\). Let \(i_0\in \{1,2\}\) be such that \(p_{i_0}\ne q_{i_0}\) and let \(k_0\ge 1\) such that \(p_{i_0k_0}\ne q_{i_0k_0}\), where
We consider the component \(\gamma _{i_0}=\sum _{j=1}^\infty \gamma _{i_0j}e_j\) and the following inequalities
In particular, we have shown that
The previous computation also shows that both d and \(\rho \) are actually distances, if restricted to any hyperplane \(\ell ^2 \times \ell ^2 \times \{\kappa \}\) with \(\kappa \in {\mathbb {R}}\).
We are now in a position to prove the following theorem.
Theorem 3.3
There exist a left invariant weak sub-Riemannian metric and a left invariant weak Riemannian metric on \(\mathbb {H}\) such that their associated geodesic distances are not positive on all couples of distinct points.
Proof
For each \(p \in \mathbb {H}\), we denote the norm of a horizontal vector
with respect to g as follows
where the last equality is due to (5) and \((v_1,v_2,0)\) is identified with a vector of \(H_0\mathbb {H}\).
Since the subspaces \(\ell ^2 \times \{0\} \times \{0\}\) and \(\{0\} \times \ell ^2 \times \{0\}\) of \(H_0\mathbb {H}\) are orthogonal with respect to \(g_0\), the previous equalities give
where \(\Vert \cdot \Vert _\eta \) is defined in (8). Thus, the length of a horizontal curve \(\gamma :[0,1] \rightarrow \mathbb {H}\) with respect to g satisfies the formula
where \(\gamma (t)= (\gamma _1(t), \gamma _2(t), \gamma _3(t))\).
Next, we wish to show that whenever \((p_1,p_2,s_1), (p_1,p_2,s_2)\in \mathbb {H}\), then
To do this, the main point is to prove that for all \(s >0\), we have \(\rho ((0,0,0), (0,0,s))=0\). We will construct a sequence of horizontal curves connecting (0, 0, 0) to (0, 0, s), whose length converges to zero. Such sequence is obtained by gluing different sequences of horizontal curves. We fix \(c >0\) and consider \(\gamma ^n:[0,1] \rightarrow \mathbb {H}\) defined by
where the unit vector \(e_n\) is the n-th vector of the fixed orthonormal basis \(\{e_j:j\ge 1\}\) of \(\ell ^2\). By definition (8), we get
From the form of \(\gamma ^n\), it is immediate to check that the differential constraint
is satisfied for all \(t\in [0,1]\), hence \(\gamma _n\) is horizontal. Thus, formula (12) holds and the expressions of (14) immediately prove that \(\ell _g(\gamma ^n)\rightarrow 0\) as \(n\rightarrow +\infty \).
Now we define the sequence of curves \(\alpha ^n:[0,1] \rightarrow \mathbb {H}\) as
We immediately obtain
and the differential constraint
is satisfied for all \(t\in [0,1]\). All curves \(\alpha ^n\) are horizontal, hence combining (12) and (15), we conclude that \(\ell _g(\alpha ^n)\rightarrow 0\) as \(n \rightarrow +\infty \). We note that
for all \(n \in {\mathbb {N}}\), hence we can consider the gluing \(\alpha ^n *\gamma ^n:[0,1] \rightarrow \mathbb {H}\) of \(\alpha ^n\) and \(\gamma ^n\), that is a piecewise smooth curve. Clearly \(\alpha ^n*\gamma ^n\) is a horizontal curve and for all \(n \in {\mathbb {N}}\) we have
and \(\ell _g(\alpha ^n *\gamma ^n) = \ell _g(\alpha ^n) + \ell _g(\gamma ^n) \rightarrow 0\) as \(n\rightarrow \infty \). We have proved that
hence \(\rho ((0,0,0),(0,0,s))=0\) for all \(s>0\). By the left invariance of \(\rho \), we have
therefore \(\rho ((0,0,0),(0,0,t))=0\) for every \(t\in {\mathbb {R}}\). We conclude that
that proves (13). According to Remark 3.1, the inequality \(d\le \rho \) implies that for all \((p_1,p_2,s_1), (p_1,p_2,s_2)\in \mathbb {H}\), we have
This concludes the proof. \(\square \)
4 On the sectional curvature of a weak Riemannian Heisenberg group
In this section, we study the sectional curvature of \(\mathbb {H}\) equipped with a specific left invariant weak Riemannian metric. Following Sect. 3, we define the unique left invariant weak Riemannian metric \(\sigma \), such that
for \((v_1,v_2,v_3),(w_1,w_2,w_3)\in T_0\mathbb {H}\), according to (10). We recall the formula
and \(Ax=\sum _{k=1}^\infty x_k/k\), \(x=\sum _{k=1}^\infty x_ke_k\in \ell ^2\). For every positive integer j, we use the notation
to indicate the standard orthonormal basis of \(\mathbb {H}\) seen as the Hilbert space \(\ell ^2\times \ell ^2\times {\mathbb {R}}\).
One easily realizes that the natural linear isomorphism between \({{\,\textrm{Lie}\,}}(\mathbb {H})\) and \(\mathbb {H}\) is also an isomorphism of Lie algebras, where we equip \(\mathbb {H}\) with the Lie product (1). Thus, by slight abuse of notation, the left invariant vector fields of \({{\,\textrm{Lie}\,}}(\mathbb {H})\) isomorphically associated with the basis \(e^1_j,e^2_j,e^3\) are denoted by the same symbols. It follows that for all \(i,j\ge 1\) and \(l=1,2\) we have
where \(e^1_i,e^1_j,e^3\) are now understood as left invariant vector fields of \({{\,\textrm{Lie}\,}}(\mathbb {H})\).
Now we consider a left invariant weak Riemannian metric \(\upsilon \) on \(\mathbb {H}\). The associated scalar product on \({{\,\textrm{Lie}\,}}(\mathbb {H})\) is denoted by \(\left\langle \cdot ,\cdot \right\rangle _\upsilon \). We consider two orthonomal vectors \(X,Y\in {{\,\textrm{Lie}\,}}(\mathbb {H})\) with respect to \(\left\langle \cdot ,\cdot \right\rangle _\upsilon \). By virtue of [2, Theorem 5], the sectional curvature \(K_\upsilon (X,Y)\) of the plane in \({{\,\textrm{Lie}\,}}(\mathbb {H})\) spanned by X and Y can be obtained by the adjoint operator \({{\,\textrm{ad}\,}}(Y)^\top (X)\), that we now introduce. We define \({{\,\textrm{ad}\,}}(Y)(Z)=[Y,Z]\) and consider (in case it exists) the unique vector \({{\,\textrm{ad}\,}}(Y)^\top (X)\in {{\,\textrm{Lie}\,}}(\mathbb {H})\) that satisfies the equalities
for every \(Z\in {{\,\textrm{Lie}\,}}(\mathbb {H})\). Then we define
In the case \(\upsilon \) is a strong Riemannian metric, [1, Definition 5.2.12], the existence of \(B_\upsilon (X,Y)\) is always ensured, but not for any weak Riemannian metric. For instance, in Remark 4.1 below, we show the nonexistence of \(B_\sigma (X,Y)\) for a specific choice of X and Y, where \(\sigma \) is the weak Riemannian metric introduced at the beginning of this section.
From formula (53) of [2], we have
where we define
The proof of Theorem 1.3 follows from the application of (20) with respect to \(\sigma \) on suitable choices of planes. We denote by \(\left\langle \cdot ,\cdot \right\rangle _\sigma \) the scalar product in \({{\,\textrm{Lie}\,}}(\mathbb {H})\) induced by the left invariant weak Riemannian metric \(\sigma \). The associated norm on \({{\,\textrm{Lie}\,}}(\mathbb {H})\) is denoted by \(\Vert \cdot \Vert _\sigma \). We assume that for \(X,Y\in {{\,\textrm{Lie}\,}}(\mathbb {H})\) the adjoint
with respect to \(\sigma \) exists. In this case, its scalar product with a vector \(Z\in {{\,\textrm{Lie}\,}}(\mathbb {H})\) is assigned by the following formula
as a consequence of (1), where \(\pi :\mathbb {H}\rightarrow \ell ^2\times \ell ^2\) is the canonical projection defined by
We use the fixed orthonormal basis \(e^1_j,e^2_j,e^3\) of \(\mathbb {H}\) with respect to the standard Hilbert product of \(\ell ^2\times \ell ^2\times {\mathbb {R}}\), getting
Formula (23) yields
for arbitrary \(Z=Z^3e^3+\sum _{j=1}^\infty Z^1_je^1_j+Z^2_je^2_j\). In the case \(X=\pi (X)\), formula (24) shows the existence of \({{\,\textrm{ad}\,}}(Y)^\top (\pi (X))\) and yields
In the case \(X=e^3\), again (24) for \(Z=e^1_j\) and \(Z=e^2_j\) respectively, gives
For \(Z=e^3\), applying (24) we get
Assuming the existence of \({{\,\textrm{ad}\,}}(Y)^\top (e^3)\), we have shown that
Writing \(Y=Y^3e^3+\sum _{j=1}^\infty (Y^1_j e^1_j+Y^2_je^2_j)\), we finally get
Then the assumption about the existence of \(B_\sigma (e^3,Y)\) corresponds to the convergence of its series. The next remark shows a choice of Y for which the series (28) does not converge.
Remark 4.1
If we consider the vector
then it is easy to check that the series (28) representing \(B_\sigma (e^3,W)\) does not converge. As a consequence, the adjoint \({{\,\textrm{ad}\,}}(W)^\top (e^3)\) cannot be defined. In addition, Arnold’s formula (20) for the sectional curvature of the plane \({{\,\textrm{span}\,}}\left\{ W,e^3 \right\} \) does not apply.
Proposition 4.2
We consider the orthonormal elements \(W_k,e^3\in {{\,\textrm{Lie}\,}}(\mathbb {H})\) with \(k\ge 1\) and
As the subspace \({{\,\textrm{span}\,}}\left\{ W_k,e^3 \right\} \) converges to \({{\,\textrm{span}\,}}\{W_\infty ,e^3\}\) for \(k\rightarrow \infty \), with
it follows that
The convergence of \({{\,\textrm{span}\,}}\left\{ W_k,e^3 \right\} \) to \({{\,\textrm{span}\,}}\{W_\infty ,e^3\}\) is considered in the Grassmannian of the 2-dimensional planes contained in \({{\,\textrm{Lie}\,}}(\mathbb {H})\).
Proof
First of all, the pointwise convergence of \(W_k\) to \(W_\infty \) implies the convergence of \({{\,\textrm{span}\,}}\left\{ W_k,e^3 \right\} \) to \({{\,\textrm{span}\,}}\{W_\infty ,e^3\}\). To compute \(K_\sigma (W_k,e^3)\), we first apply (25), getting
for all \(k\ge 1\). From (28), it follows that \(B_\sigma (e^3,e^1_j)=2je^2_j\), hence
The bilinearity of \(B_\sigma (\cdot ,\cdot )\) yields
From (21), taking \(\delta =\left( B_\sigma (W_k,e^3)+B_\sigma (e^3,W_k) \right) /2\), we obtain
From (21), (22), (25) and (28), we find
Finally, by formula (20), we have proved that
as \(k\rightarrow \infty \). This concludes the proof. \(\square \)
Proof of Theorem 1.3
Following the notation of the present section, we define
of \({{\,\textrm{Lie}\,}}(\mathbb {H})\), that are orthonormal with respect to \(\left\langle \cdot ,\cdot \right\rangle _\sigma \) and do not commute. To apply (20) for finding \(K_\sigma (a_{1j},a_{2j})\), we use (21) and (22). Due to (25), we get
As a result, we have
Now we wish to compute \(K_\sigma (a_{1j},e^3)\) and \(K_\sigma (a_{2j},e^3)\). We first apply (25) and (28), getting
for all \(l=1,2\) and \(k\ge 1\). From (21), taking \(\delta =\left( B_\sigma (a_{1j},e^3)+B_\sigma (e^3,a_{1j}) \right) /2\), we obtain
From (21), (22), (25) and (28), we find
Due to the formula for the sectional curvature (20), we have established that
In analogous setting \(\delta =\left( B_\sigma (a_{2j},e^3)+B_\sigma (e^3,a_{2j}) \right) /2\), we obtain
Again (21), (22), (25) and (28) imply that
Due to (20), we have also proved that
Taking into account (42) and (45), setting \(b=e^3\), we have completed the proof. \(\square \)
Remark 4.3
A direct verification shows that the computations of sectional curvature, to prove Theorem 1.3, could be also carried out extending the finite dimensional formula of [21, Lemma 1.1] for the countable structure coefficients of \({{\,\textrm{Lie}\,}}(\mathbb {H})\). These coefficients are obtained from the orthonormal vectors \(\sqrt{j}e^1_j,\sqrt{j}e^2_j,e^3\) of \({{\,\textrm{Lie}\,}}(\mathbb {H})\) with respect to \(\left\langle \cdot ,\cdot \right\rangle _\sigma \).
Following the notation of the this section, the sequence of curves whose length converges to zero in the proof of Theorem 3.3 can be written as
It is interesting to notice that all such curves are contained in the span of the planes
When these planes are seen in the Lie algebra, Theorem 1.3 shows that their sectional curvature blows-up, as the length of the curves converges to zero.
Change history
05 April 2023
Missing Open Access funding information has been added in the Funding Note.
04 April 2023
A Correction to this paper has been published: https://doi.org/10.1007/s13163-023-00463-4
References
Abraham, R.H., Marsden, J.E., Ratiu, T.S.: Manifolds, tensor analysis, and applications, volume 75 of Applied Mathematical Sciences. Springer-Verlag, New York, second edition, (1988)
Arnold, V.I.: Sur la géométrie différentielle des groupes de Lie de dimension infinie et ses applications à l’hydrodynamique des fluides parfaits. Ann. Inst. Fourier (Grenoble) 16, 319–361 (1966)
Bauer, M., Bruveris, M., Harms, P., Michor, P.W.: Vanishing geodesic distance for the Riemannian metric with geodesic equation the KdV-equation. Ann. Global Anal. Geom. 41(4), 461–472 (2012)
Bauer, M., Bruveris, M., Harms, P., Michor, P.W.: Geodesic distance for right invariant Sobolev metrics of fractional order on the diffeomorphism group. Ann. Global Anal. Geom. 44(1), 5–21 (2013)
Bauer, M., Bruveris, M., Michor, P.W.: Geodesic distance for right invariant Sobolev metrics of fractional order on the diffeomorphism group II. Ann. Global Anal. Geom. 44(4), 361–368 (2013)
Bauer, M., Bruveris, M., Michor, P.W.: Homogeneous Sobolev metric of order one on diffeomorphism groups on real line. J. Nonlinear Sci. 24(5), 769–808 (2014)
Bauer, M., Bruveris, M., Michor, P.W.: Overview of the geometries of shape spaces and diffeomorphism groups. J. Math. Imaging Vis. 50(1–2), 60–97 (2014)
Bauer, M., Harms, P., Preston, S.C.: Vanishing distance phenomena and the geometric approach to SQG. Arch. Ration. Mech. Anal. 235(3), 1445–1466 (2020)
Driver, B.K., Gordina, M.: Heat kernel analysis on infinite-dimensional Heisenberg groups. J. Funct. Anal. 255(9), 2395–2461 (2008)
Eliashberg, Y., Polterovich, L.: Bi-invariant metrics on the group of Hamiltonian diffeomorphisms. Internat. J. Math. 4(5), 727–738 (1993)
Glöckner, H.: Fundamental problems in the theory of infinite-dimensional Lie groups. J. Geom. Symmetry Phys. 5, 24–35 (2006)
Grong, E., Markina, I., Vasil’ev, A.: Sub-Riemannian geometry on infinite-dimensional manifolds. J. Geom. Anal. 25(4), 2474–2515 (2015)
Hörmann, G.: Representations of the Infinite Dimensional Heisenberg Group. PhD thesis, University of Vienna, (1993)
Jerrard, R.L., Maor, C.: Geodesic distance for right-invariant metrics on diffeomorphism groups: critical Sobolev exponents. Ann. Global Anal. Geom. 56(2), 351–360 (2019)
Jerrard, R.L., Maor, C.: Vanishing geodesic distance for right-invariant Sobolev metrics on diffeomorphism groups. Ann. Global Anal. Geom. 55(4), 631–656 (2019)
Magnani, V., Rajala, T.: Radon-Nikodym property and area formula for Banach homogeneous group targets. Int. Math. Res. Not. IMRN 23, 6399–6430 (2014)
Magnani, V., Tiberio, D.: A remark on vanishing geodesic distances in infinite dimensions. Proc. Am. Math. Soc. 148(8), 3653–3656 (2020)
Malliavin, P.: Hypoellipticity in infinite dimensions. In Diffusion processes and related problems in analysis, Vol. I (Evanston, IL, 1989), volume 22 of Progr. Probab., pages 17–31. Birkhäuser Boston, Boston, MA, (1990)
Michor, P.W., Mumford, D.: Vanishing geodesic distance on spaces of submanifolds and diffeomorphisms. Doc. Math. 10, 217–245 (2005)
Michor, P.W., Mumford, D.: Riemannian geometries on spaces of plane curves. J. Eur. Math. Soc. (JEMS) 8(1), 1–48 (2006)
Milnor, J.: Curvatures of left invariant metrics on Lie groups. Adv. Math. 21(3), 293–329 (1976)
Neeb, K.-H.: Towards a Lie theory of locally convex groups. Jpn. J. Math. 1(2), 291–468 (2006)
Omori, H.: Infinite-dimensional Lie groups, volume 158 of Translations of Mathematical Monographs. American Mathematical Society, Providence, RI, 1997. Translated from the 1979 Japanese original and revised by the author
Funding
Open access funding provided by Università di Pisa within the CRUI-CARE Agreement.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The original version of this article was revised to update the name of the author in reference 23.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Magnani, V., Tiberio, D. On the Michor–Mumford phenomenon in the infinite dimensional Heisenberg group. Rev Mat Complut 36, 973–989 (2023). https://doi.org/10.1007/s13163-022-00430-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s13163-022-00430-5
Keywords
- Hilbert manifold
- Infinite dimensional Heisenberg group
- Weak Riemannian metric
- Geodesic distance
- Sectional curvature