arXiv reader

Long time behavior of leafwise heat flow for Riemannian foliations

Jesus A. Alvarez Lopez and Yuri A. Kordyukov

arXiv
dg-ga/9612010v3 · PDF · 14 May 2025
Categories
dg-ga, math.DG
Licence
CC BY 4.0 The site shows the text.

Abstract

For any Riemannian foliation F on a closed manifold M with an arbitrary bundle-like metric, leafwise heat flow of differential forms is proved to preserve smoothness on M at infinite time. This result and its proof have consequences about the space of bundle-like metrics on M, about the dimension of the space of leafwise harmonic forms, and mainly about the second term of the differentiable spectral sequence of F.

Items 43

Theorem

Theorem 1.1

939N

With the above notation, suppose that, on C^\infty(E), there is a transversely elliptic first order differential operator A, and there are zero order differential operators G, H, K and L such that

Ad\pm dA=Gd+dH\;,\quad A\delta\pm\delta A=K\delta+\delta L\;.

Then \Pi defines a continuous operator on C^\infty(E), we have the leafwise Hodge decomposition

C^\infty(E)=\ker{\Delta}\oplus\overline{\mbox{\rm im}\;{\Delta}} =(\ker d\cap\ker\delta)\oplus\overline{\mbox{\rm im}\; d}\oplus\overline{\mbox{\rm im}\;\delta}\;,

and (t,\phi)\mapsto e^{-t{\Delta}}\phi defines a continuous map [0,\infty]\timesC^\infty(E)\rightarrowC^\infty(E).

Used by Remark 2.10
Corollary

Corollary C

T13Q

With the above notation, assume that \widehat\mathcal{F} is a Riemannian foliation. Then \Pi defines a continuous operator on \Omega(\mathcal{F},V), we have the leafwise Hodge decomposition

\Omega(\mathcal{F},V)=\ker\Delta_\mathcal{F}\oplus\overline{\operatorname{im}\Delta_\mathcal{F}} =(\ker d_\mathcal{F}\cap\ker\delta_\mathcal{F})\oplus\overline{\operatorname{im} d_\mathcal{F}}\oplus \overline{\operatorname{im}\delta_\mathcal{F}}\;,

and (t,\alpha)\mapsto e^{-t\Delta_\mathcal{F}}\alpha defines a continuous map [0,\infty]\times\Omega(\mathcal{F},V)\to\Omega(\mathcal{F},V). Thus {\mathcal H}(\mathcal{F},V) can be canonically identified with \ker\Delta_\mathcal{F}, and, if \mathcal{F} is oriented, then the leafwise Hodge star operator on \ker\Delta_\mathcal{F} induces an isomorphism {\mathcal H}^v(\mathcal{F},V)\cong{\mathcal H}^{p-v}(\mathcal{F},V^\ast).

Theorem

Theorem 1.2

EK8A

Let {\cal F} be a Riemannian foliation on a closed manifold M with a bundle-like metric. Then \Pi defines a continuous operator on \Omega, we have the leafwise Hodge decomposition

\Omega=\ker{\Delta}_0\oplus\overline{\mbox{\rm im}\;{\Delta}_0} =(\ker d_{0,1}\cap\ker\delta_{0,-1})\oplus\overline{\mbox{\rm im}\; d_{0,1}} \oplus\overline{\mbox{\rm im}\;\delta_{0,-1}}\;,

and (t,\alpha)\mapsto e^{-t{\Delta}_0}\alpha defines a continuous map [0,\infty]\times\Omega\rightarrow\Omega.

Example

Example 1 (S. Goette)

7W9N

Consider \mathbb{S}^1\equiv\mathbb{R}/2\pi\mathbb{Z}, and let [x]\in \mathbb{S}^1 denote the class of every x\in\mathbb{R}. Take the 2-torus M = \mathbb{S}^1\times \mathbb{S}^1 with the foliation \mathcal{F} whose leaves are the fibers of the first factor projection to \mathbb{S}^1. It is the quotient of the foliation \widetilde\mathcal{F} on \widetilde M :=\mathbb{R}^2 whose leaves are the fibers of the first factor projection to \mathbb{R}. Let V be the complex line bundle over M that is the quotient of the trivial complex line bundle \widetilde V := \widetilde M\times\mathbb{C} over \widetilde M by the action of \mathbb{Z}^2 defined by the transformations

T_{m,n}(x,y,z) = (x+2\pi m, y+2\pi n, e^{2\pi n \sin(x) i} z),

where i = \sqrt{-1} and m,n \in\mathbb{Z}. Consider also the trivial Hermitean structure on \widetilde V. Then the trivial flat \widetilde\mathcal{F}-partial connection \widetilde\nabla on \widetilde V is a Hermitean connection, and projects to a flat \mathcal{F}-partial Hermitean connection \nabla on V. Via the canonical identity \mathbb{C} \equiv \mathbb{R}^2, \widetilde V and V become oriented real vector bundles of rank 2, with trivial Riemannian structures so that \widetilde\nabla and \nabla are flat \mathcal{F}-partial Riemannian connections. Then \Omega^0(\mathcal{F}, V) = C^\infty(M; V) (the space of smooth sections of V), which can be considered as the subspace of C^\infty(\widetilde M; \widetilde V) \equiv C^\infty(\widetilde M, \mathbb{C}) consisting of the smooth functions f: \widetilde M \to \mathbb{C} satisfying the equivariance condition

f(x+2\pi m, y+2\pi n) = e^{2\pi n \sin(x) i} f(x, y)\quad(m,n\in\mathbb{Z})\;.

Such a function f satisfies d_{\widetilde\mathcal{F}}(f)=0 if and only if it depends only on x; say f = f(x). Then Example 1 means that

f(x) = e^{2\pi n \sin(x) i} f(x)\quad(n\in\mathbb{Z})\;,

If \sin(x) is irrational, then the set \{\,e^{2\pi m \sin(x) i}\mid m\in\mathbb{Z}\,\} is dense in the complex unit circle \{\,z\in\mathbb{C}\mid|z|=1\,\}. So the corresponding section s \in C^\infty(M; V) is invariant by multiplication by complex numbers of modulus 1 on the leaves \{[x]\} \times \mathbb{S}^1 with \sin(x) irrational. So s vanishes on those leaves, whose union is dense in M, and therefore s vanishes on M. This shows that \ker d_\mathcal{F} = 0 in \Omega^0(\mathcal{F}, V).

On the other hand, the function f(x,y) = e^{y \sin(x) i} satisfies Example 1 and is equal to 1 on 2\pi \mathbb{Z} \times \mathbb{R}. So it defines a section s\in C^\infty(M;V) satisfying \Delta_\mathcal{F} s = 0 on the leaf L_0 = \{ [0] \} \times \mathbb{S}^1. Then s_t := e^{-t\Delta_\mathcal{F}}s remains independent of t>0 on L_0, and therefore the restriction to 2\pi \mathbb{Z} \times \mathbb{R} of its lift to \widetilde M remains equal to f for all t. If the limit of s_t as t \to \infty were defined in C^\infty(M;V), it would be in \ker d_\mathcal{F} and different from zero on L_0, a contradiction.

In this case, the principal O(2)-bundle F of orthonormal frames of V over M is the quotient of the bundle \widetilde F := \widetilde M\times \{\,z\in\mathbb{C}\mid|z|=1\,\} over \widetilde M by the action of \mathbb{Z}^2 defined by the transformations T_{m,n} (m,n \in\mathbb{Z}). The foliation \widehat\mathcal{F} on F is the quotient of the foliation \widetilde\mathcal{F} on \widetilde F whose leaves are the fibers of the first factor projection to \mathbb{R}. \widehat\mathcal{F} is not Riemannian because its leaf closures do not define a stratified structure of F (see Section 5.4 of [Molino1988]).

Corollary

Corollary 1.3

VHY9

Let {\cal F} be a Riemannian foliation on a closed manifold M, and let V be any Riemannian vector bundle with a flat Riemannian {\cal F}-partial connection. Fix any Riemannian metric on the leaves, smooth on M. Then \Pi defines a continuous operator on \Omega({\cal F},V), we have the leafwise Hodge decomposition

\Omega({\cal F},V)=\ker{\Delta}_{\cal F}\oplus\overline{\mbox{\rm im}\;{\Delta}_{\cal F}} =(\ker d_{\cal F}\cap\ker\delta_{\cal F})\oplus\overline{\mbox{\rm im}\; d_{\cal F}}\oplus \overline{\mbox{\rm im}\;\delta_{\cal F}}\;,

and (t,\alpha)\mapsto e^{-t{\Delta}_{\cal F}}\alpha defines a continuous map [0,\infty]\times\Omega({\cal F},V)\rightarrow\Omega({\cal F},V). Thus {\cal H}({\cal F},V) can be canonically identified with \ker{\Delta}_{\cal F}, and, if {\cal F} is oriented, then the leafwise Hodge star operator on \ker{\Delta}_{\cal F} induces an isomorphism {\cal H}^v({\cal F},V)\cong{\cal H}^{p-v}({\cal F},V^\ast).

Acknowledgment

Acknowledgment

0YLG

We thank S. Goette for providing us with the example.

Corollary

Corollary 1.4

XR8L

Let {\cal F} be a Riemannian foliation on a closed manifold M. Then, with respect to the C^\infty topology, the space of bundle-like metrics on M is a deformation retract of the space of all metrics on M.

Corollary

Corollary 1.5

GW17

Let {\cal F} be a Riemannian foliation on a closed manifold M, and let V any Riemannian vector bundle with a flat Riemannian {\cal F}-partial connection. For any metric on the leaves, smooth on M, if there is a nontrivial integrable harmonic r-form on some leaf with coefficients in V, then \Omega^r({\cal F},V)\cap\ker{\Delta}_{\cal F} is of infinite dimension.

Theorem

Theorem 1.6

GZJK

Let {\cal F} be a Riemannian foliation on a closed manifold M. For any bundle-like metric on M, the operators {\Delta}_1 and \widetilde{\Delta}_1 are essentially self-adjoint on L^2{\cal H}_1 and L^2\widetilde{\cal H}_1, respectively. Moreover L^2{\cal H}_1 and L^2\widetilde{\cal H}_1 have complete orthonormal systems, \{\phi_i : i=1,2,\ldots\}\subset{\cal H}_1 and \{\tilde\phi_i : i=1,2,\ldots\}\subset\widetilde{\cal H}_1, consisting of eigenvectors of {\Delta}_1 and \widetilde{\Delta}_1, respectively, so that the corresponding eigenvalues satisfy 0\leq\lambda_1\leq\lambda_2\leq\cdots, 0<\tilde\lambda_1\leq\tilde\lambda_2\leq\cdots, with \lambda_i\uparrow\infty if \dim{\cal H}_1=\infty, and \tilde\lambda_i\uparrow\infty if \dim\widetilde{\cal H}_1=\infty; thus all of these eigenvalues have finite multiplicity. In particular we have

{\cal H}_1=\ker{\Delta}_1\oplus\mbox{\rm im}\;{\Delta}_1\;,\quad \widetilde{\cal H}_1=\mbox{\rm im}\;\widetilde{\Delta}_1\;.
Corollary

Corollary 1.7 (X. Masa)

JVHA

Under the same conditions we have H(\bar 0_1)=0 and E_2\cong H({\cal E}_1), which is of finite dimension and, if M is orientable, satisfies the duality E_2^{u,v}\cong E_2^{q-u,p-v}, u,v\in{\bf Z}, where p=\dim{\cal F} and q=\mbox{\rm codim}\;{\cal F}.

Acknowledgment

Acknowledgment Acknowledgment

8VI6

The first author would like to thank S. Hurder and J. Roe for helpful conversations about leafwise heat flow. The second author gratefully acknowledges the hospitality and support of the University of Santiago de Compostela.

Proposition

Proposition 2.1 (J. Roe)

MLKP

The functional calculus map f\mapsto f(D) given by the spectral theorem restricts to a continuous homomorphism of algebras and {\bf C}[z]-modules from \cal A to the space of bounded endomorphisms of each W^k(E), and thus to the space of continuous endomorphisms of C^\infty(E). In particular, e^{-t{\Delta}} defines a bounded operator on each W^k(E) and a continuous operator on C^\infty(E), which depends smoothly on t\in[0,\infty).

Used by Proof
Proof

Proof (property (vi) for k=0)

R431

Since the image of d is orthogonal to the image of \delta in L^2(E), we get \| D\phi\|_0^2= \| d\phi\|_0^2+\|\delta\phi\|_0^2 for any \phi\inC^\infty(E). Thus $$ d_0+_02  D_0 .

Lemma

Lemma 2.2

XFWL

There are bounded operators B_1,\ldots,B_5 on W^l(E) such that

\begin{aligned} Ad\pm dA&=B_1d+dB_2\;,\\ A\delta\pm\delta A&=B_1\delta+\delta B_2\;,\\ {[A,{\Delta}]}&=B_3{\Delta} + {\Delta} B_4 + D B_5 D\;. \end{aligned}
Used by Proof
Lemma

Lemma 2.3

J5MW

For any bounded operator R:W^{l+1}(E)\rightarrow W^l(E), we have the following Duhamel type formula:

[ R , e^{-t{\Delta}} ]= - \int_0^t e^{-(t-s){\Delta}} \, [ R,{\Delta} ] \, e^{-s{\Delta}} \, ds

as a bounded operator W^{l+1}(E)\rightarrow W^l(E).

Lemma

Lemma 2.4

GZR0

[A,e^{-t\Delta}] defines a bounded operator on W^l(E), and there exists \tilde c_{l,1}>0 such that

\|[A,e^{-t\Delta}]\|_l\leq\tilde c_{l,1}\quad\mbox{for all}\quad t>0\;.
Lemma

Lemma 2.5

24S9

AD\,e^{-t{\Delta}}\pm D\,e^{-t{\Delta}} A defines a bounded operator on W^l(E) and there exists \tilde c_{l,2}>0 such that

\left\|AD\,e^{-t{\Delta}}\pm D\,e^{-t{\Delta}}\,A\right\|_l\leq \frac{\tilde c_{l,2}}{\sqrt{t}}\quad\mbox{for all}\quad t>0\;.
Used by Proof
Proof

Proof (property (i) for k=l+1)

SBHV

This property follows from (?) since, for \phi\in W^{l+1}(E), we have the following:

\left\| e^{-t{\Delta}}\phi\right\|_l\leq c_{l,1}\,\|\phi\|_l\leq c_{l,1}e_{l,2}\,\|\phi\|_{l+1}\;,
\left\| D\,e^{-t{\Delta}}\phi\right\|_l= \left\|e^{-t{\Delta}}\,D\phi\right\|_l\leq c_{l,1}\,\|D\phi\|_l\leq c_{l,1}e_{l,2}\,\|\phi\|_{l+1}\;,
\begin{aligned} \left\|A\,e^{-t{\Delta}}\phi\right\|_l&\leq&\left\|\left[A,e^{-t{\Delta}}\right]\phi \right\|_l+ \left\|e^{-t{\Delta}}\,A\phi\right\|_l\\ &\leq&\tilde c_{l,1}\,\|\phi\|_l+ c_{l,1}\,\|A\phi\|_l\\ &\leq&(\tilde c_{l,1}+c_{l,1})e_{l,2}\,\|\phi\|_{l+1}\;, \end{aligned}

where we have used property (i) for k=l, Lemma ?, ? and (?). \Box

Proof

Proof (property (ii) for k=l+1)

V5DO

Again this property follows from (?) since, for \phi\in W^{l+1}(E) we have the following:

\left\| D\,e^{-t{\Delta}}\phi \right\|_l\leq \frac{c_{l,2}}{\sqrt{t}}\|\phi\|_l \leq\frac{c_{l,2}e_{l,2}}{\sqrt{t}}\,\|\phi\|_{l+1}\;,
\left\|{\Delta} \,e^{-t{\Delta}}\phi\right\|_l=\left\|D\,e^{-t{\Delta}}\,D\phi\right\|_l\leq \frac{c_{l,2}}{\sqrt{t}}\,\|D\phi\|_l\leq \frac{c_{l,2}e_{l,2}}{\sqrt{t}}\,\|\phi\|_{l+1}\;,
\begin{aligned} \left\|AD\,e^{-t{\Delta}}\phi\right\|_l&\leq&\left\|\left(AD\,e^{-t{\Delta}}\pm D\,e^{-t{\Delta}}\,A\right)\phi\right\|_l+\left\|D\,e^{-t{\Delta}} \,A\phi\right\|_l\\ &\leq&\frac{\tilde c_{l,2}}{\sqrt{t}}\,\|\phi\|_l+ \frac{c_{l,2}}{\sqrt{t}}\,\|A\phi\|_l\\ &\leq&\frac{(\tilde c_{l,2}+c_{l,2})e_{l,2}}{\sqrt{t}}\,\|\phi\|_{l+1}\;, \end{aligned}

where we have used property (ii) for k=l, Lemma 2.5 and (?). \Box

Proof

Proof (property (iii) for k=l+1)

30OH

This follows directly from property (ii) for k=l+1 since {\Delta} \,e^{-t{\Delta}}=D\,e^{-t{\Delta}/2}\,D\,e^{-t{\Delta}/2} on C^\infty(E). \Box

Lemma

Lemma 2.6

9G7F

The operator \left[A+B,e^{-t{\Delta}}\right] strongly converges to the zero operator on W^l(E) as t\rightarrow\infty. Moreover (t,\phi)\mapsto\left[A+B,e^{-t{\Delta}}\right]\phi extends to a continuous map [0,\infty]\times W^l(E)\rightarrow W^l(E) vanishing on \{\infty\}\times W^l(E).

Proof

Proof (property (iv) for k=l+1)

XQYY

Consider the following bounded compositions:

\mbox{\begin{picture}(166,22) \put(0,0){$W^{l+1}(E)$} \put(43,3){\vector(1,0){30}} \put(46,6){$e^{-t{\Delta}}$} \put(77,0){$W^{l+1}(E)\hookrightarrow W^l(E)\;,$} \end{picture}}
\mbox{\begin{picture}(176,17) \put(0,0){$W^{l+1}(E)$} \put(43,3){\vector(1,0){30}} \put(46,6){$e^{-t{\Delta}}$} \put(77,0){$W^{l+1}(E)$} \put(120,3){\vector(1,0){20}} \put(125,6){$D$} \put(144,0){$W^l(E)\;,$} \end{picture}}
\mbox{\begin{picture}(192,17) \put(0,0){$W^{l+1}(E)$} \put(43,3){\vector(1,0){30}} \put(46,6){$e^{-t{\Delta}}$} \put(77,0){$W^{l+1}(E)$} \put(120,3){\vector(1,0){36}} \put(123,6){$A+B$} \put(160,0){$W^l(E)\;.$} \end{picture}}

By (?), it is enough to prove that the compositions (Proof), (Proof) and (Proof) are strongly convergent as t\rightarrow\infty, and to prove the continuous extension to [0,\infty]\times W^{l+1}(E) of the maps [0,\infty)\times W^{l+1}(E)\rightarrow W^l(E) defined by these time dependent operators. This holds for (Proof) and (Proof) by property (iv) for k=l since these operators are respectively equal to the compositions

The paper has a picture here, which this page does not show.

On the other hand, (Proof) is the sum of the bounded compositions

\mbox{\begin{picture}(196,17) \put(0,0){$W^{l+1}(E)\hookrightarrow W^l(E)$} \put(90,3){\vector(1,0){70}} \put(93,9){$\left[A+B,e^{-t{\Delta}}\right]$} \put(164,0){$W^l(E)\;,$} \end{picture}}
\mbox{\begin{picture}(182,17) \put(0,0){$W^{l+1}(E)$} \put(43,3){\vector(1,0){36}} \put(46,6){$A+B$} \put(83,0){$W^l(E)$} \put(116,3){\vector(1,0){30}} \put(119,6){$e^{-t{\Delta}}$} \put(150,0){$W^l(E)\;.$} \end{picture}}

Now Lemma ?, ? and property (iv) for k=l respectively imply the strong convergence of (?) and (?) as t\rightarrow\infty, as well as the continuous extension to [0,\infty]\times W^{l+1}(E) of the maps [0,\infty)\times W^{l+1}(E)\rightarrow W^l(E) defined by both of these time dependent operators. \Box

Proof

Proof (property (v) for k=l+1)

9IGP

By Corollary 2.7 and property (v) for k=l, \Pi defines a projection of W^{l+1}(E) onto the kernel of {\Delta} in W^{l+1}(E).

Now, for t>0 and s\in{\bf R} let

f_t(s)=\left\{ \begin{array}{cl} (1-e^{-ts^2})/s^2& \mbox{if}\quad s\neq 0\\ t& \mbox{if}\quad s=0\;. \end{array} \right.

It is easy to check that f_t is in the algebra \cal A of Proposition 2.1. Thus f_t(D) defines a bounded operator on W^{l+1}(E) satisfying

\mbox{\rm id}-e^{-t{\Delta}}={\Delta}\,f_t(D)\;.

So property (iv) for k=l+1 yields that, for any \phi\in W^{l+1}(E), \widetilde\Pi\phi is the limit of {\Delta} f_t(D)\phi in W^{l+1}(E) as t\rightarrow\infty. Hence

\widetilde\Pi(W^{l+1}(E))\subset\mbox{\rm cl}_{l+1}(\mbox{\rm im}\;{\Delta})\;.

This is really an equality because the reverse inclusion can be easily proved as follows:

\begin{aligned} \mbox{\rm cl}_{l+1}(\mbox{\rm im}\;{\Delta})&\subset&W^{l+1}(E)\cap\mbox{\rm cl}_0(\mbox{\rm im}\;{\Delta})\\ &=W^{l+1}(E)\cap\widetilde\Pi\left(L^2(E)\right)\\ &=\widetilde\Pi\left(W^{l+1}(E)\right)\;. \end{aligned}

Here we have used property (v) for k=0 and the fact that \widetilde\Pi is a projection. Therefore we have proved the first direct sum decomposition in property (v) for k=l+1.

The second direct sum decomposition follows similarly by using the functions given by

g_t(s)=\left\{ \begin{array}{cl} (1-e^{-ts^2})/s& \mbox{if}\quad s\neq 0\\ t& \mbox{if}\quad s=0 \end{array} \right.

instead of the f_t. \Box

Proof

Proof (property (vi) for k=l+1)

UG3W

Take any \phi\inC^\infty(E). Lemma 2.2, property (vi) for k=l and (?) yield d_l+1+_l+1&& 1e_l,1”(d_l+_l+ Dd_l+D_l.
& &.+(A-B_1)d_l+(A-B_1)_l)
&&1e_l,1”(d_l+_l+ dD_l+D_l.
& &.+d(A+B_2)_l+(A+B_2)_l)
&&c_l,4e_l,1”(D_l+ _l+D(A+B_2)_l)
&&c_l,4e_l,1”(D_l+ _l+(A-B_1)D_l)
&&c_l,4e_l,2”e_l,1”  D_l+1 .

Remark

Remark 2.9

THAU

If A is symmetric in Theorem 1.1, then one of the two equations in (1.1) can be removed in its statement because both of them are equivalent by taking adjoints on M.

Remark

Remark 2.10

G87L

The following more general condition can be used instead of (1.1) in Theorem 1.1 to get the same long time behavior of leafwise heat flow with a similar proof: There is some first order transversely elliptic differential operator A and zero order operators K_1,\ldots,K_{16} such that the operators

\begin{aligned} F_1&=Ad\pm dA-K_1d-dK_2-K_3\delta-\delta K_4 \;,\\ F_2&=A\delta\pm\delta A-K_5\delta-\delta K_6-K_7d-dK_8 \end{aligned}

satisfy

\begin{aligned} \lefteqn{F_1\delta\mp\delta F_1+F_2d\mp dF_2=}\\ &&dK_9d+\delta K_{10}d+dK_{11}\delta+\delta K_{12}\delta+d\delta K_{13}+\delta dK_{14}+K_{15}d\delta+K_{16}\delta d\;. \end{aligned}

Nevertheless, so far we did not find any non-Riemannian foliation with a (non-elliptic) leafwise elliptic differential complex satisfying such a more general condition.

Corollary

Corollary 2.11

S9JO

[A,\Pi] defines a bounded operator on each W^k(E), and thus a continuous operator on C^\infty(E).

Proposition

Proposition 3.1

RX4X

If {\cal F} is a Riemannian foliation and M is endowed with a bundle-like metric, there is a zero order differential operator K on \Omega such that

D_\perp\delta_{0,-1}+\delta_{0,-1} D_\perp =K\delta_{0,-1}+\delta_{0,-1} K\;.
Used by Remark 3.6
Lemma

Lemma 3.2

1ZXO

Over U, the Hodge star operator on {\bigwedge TM^\ast}\equiv\bigwedge T{\cal F}^{\perp\ast}\otimes{\bigwedge T{\cal F}^\ast} is given by

\star=(-1)^{(q-u)v}\star_\perp\otimes\star_{\cal F}: \bigwedge^uT{\cal F}^{\perp\ast}\otimes\bigwedge^vT{\cal F}^\ast\longrightarrow \bigwedge^{q-u}T{\cal F}^{\perp\ast}\otimes\bigwedge^{p-v}T{\cal F}^\ast\;.
Lemma

Lemma 3.3

IBXV

If X\in{{\cal X}\left(U,{{\cal F}_U}\right)} is orthogonal to the leaves, then there is a zero order differential operator R_{U,X} on \Omega^{0,\cdot}(U), depending C^\infty(M/{{\cal F}_U})-linearly on X, such that

[\theta_X,\delta_{0,-1}]=[R_{U,X},\delta_{0,-1}]\quad\mbox{on}\quad \Omega^{0,\cdot}\;.

Moreover the assignment (U,X)\mapsto R_{U,X} can be chosen so that the restriction of R_{U,X} to any subset U'\subset U of triviality for {\cal F} is equal to R_{U',X'}, where X'=X|_{U'}.

Lemma

Lemma 3.4

RZPM

We have

d_{0,1}\equiv(-1)^u\,\mbox{\rm id}\otimes d_{0,1}\;,\quad \delta_{0,-1}\equiv(-1)^u\,\mbox{\rm id}\otimes\delta_{0,-1}

with respect to the canonical decomposition

\Omega^{u,\cdot}(U)\equiv \Omega^u\left(U/{{\cal F}_U}\right)\otimes\Omega^{0,\cdot}(U)

as tensor product of C^\infty\left(U/{{\cal F}_U}\right)-modules.

Remark

Remark 3.5

3J71

Proposition 3.1 is slightly stronger than (1.1) in Theorem 1.1. In this case K=L, and thus

D_\perp D_0+D_0 D_\perp= B_1D_0+D_0B_2

with

B_1=K^\ast P+KQ\;,\quad B_2=QK^\ast+PK

in Lemma 2.2. In particular B_2=B_1^\ast. Observe that, if K is symmetric, then so is B_1, yielding that D_\perp-B_1 commutes with {\Delta}_0 and thus with e^{-t{\Delta}_0}. Therefore the proof of Theorem 1.2 would follow with a much simpler induction argument. Of course K is not symmetric in general.

Remark

Remark 3.6

WWA7

The above proof of Proposition 3.1 gives explicit local descriptions of K and L. But an alternative proof can be made by using Molino’s description of Riemannian foliations: It allows to reduce the proof to the case of transversely parallelizable foliations, where our local arguments can be made globally.

Remark

Remark 3.7

VJ4R

In the setting of this section, Corollary ?, ? states that [D_\perp,\Pi] defines a bounded operator on each W^k\Omega.

Lemma

Lemma 5.1

39HN

Suppose {\cal F} is Riemannian and M is closed with a fixed bundle-like metric. Then the corresponding leafwise heat operator e^{-t{\Delta}_{\cal F}} on \Omega({\cal F},Q) preserves C^\infty(Q^+) for each t\in[0,\infty].

Lemma

Lemma 7.1

40R7

D_1 and \widetilde D_1 are essentially self-adjoint in L^2{\cal H}_1 and L^2\widetilde{\cal H}_1, respectively.

Lemma

Lemma 7.2

U03N

We have the following properties:

  • (i) D\Pi-\Pi D_\perp\Pi defines a bounded operator on L^2\Omega.

  • (ii) For each v\in{\bf Z}, D\widetilde\Pi_{\cdot,v} -\widetilde\Pi_{\cdot,v}D\widetilde\Pi_{\cdot,v} defines a bounded operator on L^2\Omega.

Corollary

Corollary 7.3

LVVC

The restrictions of the kth Sobolev norm \|\cdot\|_k to {\cal H}_1 and \widetilde{\cal H}_1 are respectively equivalent to the norms \|\cdot\|_{D_1,k} and \|\cdot\|_{\widetilde D_1,k}. Thus W^k{\cal H}_1 and W^k\widetilde{\cal H}_1 are the closures of {\cal H}_1 and \widetilde{\cal H}_1 in W^k\Omega, respectively.

Lemma

Lemma 7.4 (V. Sergiescu)

NLSB

The cohomology of the quotient complex {\cal B}_v/{\cal Z}_{v-1} is trivial.

Remark

Remark 7.5

Y8GL

Let E(\lambda) and \widetilde E\left(\tilde\lambda\right) be the eigenspaces corresponding to eigenvalues \lambda of {\Delta}_1 and \tilde\lambda of \widetilde{\Delta}_1. This eigenspaces have bigradings induced by the bigradings of {\cal H}_1 and \widetilde{\cal H}_1. Then, if M is oriented, the Hodge star operator also induces duality E(\lambda)^{u,v}\cong E(\lambda)^{q-u,p-v}, and skew duality \widetilde E\left(\tilde\lambda\right)^{u,v}\cong\widetilde E\left(\tilde\lambda\right)^{q-u-1,p-v+1}.

Remark

Remark 7.6

DI5Z

If we use appropriate zero order modifications of D_\perp, then we get the Hodge theoretic version of the results of [Demetrio], which have important implications about tenseness.

Source and licence

The text on this page is from arXiv:dg-ga/9612010v3 by Jesus A. Alvarez Lopez and Yuri A. Kordyukov, under the licence CC BY 4.0. The corpus made these changes: it read the items out of the LaTeX source, it expanded the macros of the paper, it put each proof after its statement, and it converted the text to HTML. The formulas are the TeX of the paper, and KaTeX renders them. A formula or a reference can look different from the PDF. The page is published under CC BY 4.0.

The record of each item is in analysis/objects/9612/dg-ga-9612010.jsonl.