Continuous leafwise harmonic functions on codimension one transversely isometric foliations
Abstract
Let be a codimension one foliation on a closed manifold which admits a transverse dimension one Riemannian foliation. Then any continuous leafwise harmonic functions are shown to be constant on leaves.
1. Introduction
Let be a closed manifold, and let be a continuous leafwise foliation on This means that . is covered by a finite union of continuous foliation charts and the transition functions are continuous, together with their leafwise partial derivatives up to order 2. Let be a continuous leafwiseFootnote1 leafwise Riemannian metric. In this paper such a triplet is simply referred to as a leafwise foliation. For simplicity, we assume throughout the paper that the manifold and the foliation are oriented. For a continuous leafwiseFootnote2 real valued function on the leafwise Laplacian , is defined using the leafwise metric It is a continuous function of . .
This means that the leafwise partial derivatives up to order 2 of the components of in each foliation chart are continuous in the chart.
The leafwise partial derivatives of up to order 2 in each foliation chart are continuous in the chart.
As an example, if is a foliation by compact leaves, then is Liouville. Moreover, there is an easy observation.
This can be seen as follows. Let (resp. be the maximum (resp. minimum) value of the continuous leafwise harmonic function ) on Assume . takes the maximum value at Then by the maximum principle, . on the leaf which passes through Now the closure of . contains the unique minimal set Therefore . on The same argument shows that . on finishing the proof that , is constant on .
A first example of non-Liouville foliations is obtained by R. Feres and A. Zeghib in a beautiful and simple construction Reference FZ. It is a foliated over a hyperbolic surface, with two compact leaves. There are also examples in codimension one. B. Deroin and V. Kleptsyn -bundleReference DK have shown that a codimension one foliation is non-Louville if is transversely admits no transverse invariant measure, and possesses more than one minimal set, and they have constructed such a foliation. ,
A codimension one foliation is called if the leaf space of its lift to the universal covering space is homeomorphic to -covered See .Reference F or Reference FFP. It is shown in Reference F and Reference DKNP that an foliation without compact leaves admits a unique minimal set. Therefore the above example of a codimension one non-Liouville foliation is not -covered This led the authors of -covered.Reference FFP to the study of the Liouville property for foliations. The purpose of this paper is to generalize a result of -coveredReference FFP.
Notice that a transversely isometric foliation is Our main result is the following. -covered.
In Reference FFP, the above theorem is proved in the case where the leafwise Riemannian metric is negatively curved. Undoubtedly this is the most important case. But the general case may equally be of interest.
If a transversely isometric foliation does not admit a compact leaf, then, being it admits a unique minimal set, and Theorem -covered,1 holds true by Proposition 1.3. Therefore we consider only the case where admits a compact leaf. In this case the union of compact leaves is closed. Let be a connected component of and let , be the metric completion of Then . is a foliated interval bundle, since the one dimensional transverse foliation is Riemannian.
Therefore we are led to consider the following situation. Let be a closed manifold of dimension equipped with a , Riemannian metric Let . where , is the interval Denote by . the canonical projection. Consider a continuous foliation which is transverse to the fibers , Although . is only continuous, its leaf has a differentiable structure as a covering space of by the restriction of Also, . admits a leafwise Riemannian metric obtained as the lift of to each leaf by Such a triplet . is called a leafwise foliated in this paper. Now Theorem -bundle1 reduces to the following theorem.
Sections 2–4 are devoted to the proof of Theorem 2. The proof is by contradiction. Throughout these sections, denotes a leafwise foliated without interior compact leaves, and we assume that there is a continuous leafwise harmonic function -bundle such that , As remarked in .Reference FFP, this is not a loss of generality. Also notice that for any point we have , .
The same phenomenon for discrete group actions is discussed in Reference FR. We give an alternative proof of it in Section 5.
2. Preliminaries
In this section, we recall fundamental facts about Brownian motions, needed in the next section.
Let us denote by the space of continuous leafwise paths For any . a random variable , is defined by Let . be the of -algebra generated by ( As is well known and easy to show, ). coincides with the generated by the compact open topology on -algebra A bounded function . is called a Borel function if is i.e. if for any Borel subset -measurable, the inverse image , belongs to .
For any point the Wiener probability measure , is defined using the leafwise Riemannian metric Notice that . For any bounded Borel function . the expectation of , w.r.t. is denoted by The following proposition is well known. .
Let be the completion of by the measure For any . let , be the generated by -algebra ( Its completion is denoted by ). Notice that unlike . and , and depend strongly on but we suppress the dependence in the notation. ,
The following fact is well known and can be shown easily using the Radon-Nikodým theorem.
The function -measurable is called the conditional expectation of with respect to and is denoted by .
For any let , be the shift map by defined by ,
The following proposition is known as the Markov property. See for example Reference O.
A family of uniformly bounded functions, -measurable is called a bounded , if -martingale
We have the following martingale convergence theorem. See Reference O, Appendix C.
We shall raise two applications of the above facts, which will be useful in the next section. Let be the continuous leafwise harmonic function defined at the end of Section 1. Then is a Borel function defined on .
3. Proof of Theorem 2
Again let be a continuous leafwise harmonic function defined at the end of Section 1. A probability measure on is called stationary if for any continuous leafwise function .
Given let , a neighbourhood of the upper boundary component , Let . be the subset of defined by
Thus consists of those paths which visit at arbitrarily large time. Let be the characteristic function of Clearly . is a Borel function on and satisfies for any Thus by Lemma .2.6, the function defined by is leafwise harmonic.
Another important feature of the function is that is nondecreasing along the fiber , since our leafwise Brownian motion is synchronized; i.e. it is the lift of the Brownian motion on , Notice that . on , .
The proof is the same as the proof of Proposition 9.1 of Reference FFP. In short, if we assume the contrary, we can construct a stationary measure such that which is contrary to Proposition ,3.1. The proof is included in Section 4 for completeness.
4. Proof of Lemma 3.2
The projection has a distinguished role since the leafwise metric is the lift of by Let . be an open ball in the base manifold and consider an inclusion , such that , and is contained in a leaf of , Such an inclusion . or its domain , is called a distinguished chart. ,
Using a distinguished chart, let us define by
Of course does not depend on the choice of the distinguished charts and is defined on the whole .
The function is right semicontinuous on each fiber , and is leafwise harmonic just as , Since . is nondecreasing on each fiber , we get a probability measure , on by
for any Again . does not depend on the choice of the distinguished charts.
Our goal is to show that for any Clearly this shows that . and hence , is constant on , Since we do not use the definition of . in what follows, we write for for simplicity.
We first show that does not have an atom in , Assume on the contrary that there is . such that has an atom in Then in a distinguished chart, there is . such that
Define a positive function on the plaque by
Using other distinguished charts, we can define on the whole leaf which passes through The function . is positive harmonic on Clearly we have .
for any .
Define a measure on by
where is the Dirac mass at and denotes the volume form on Precisely ( .Equation 4.1) means that for any continuous function on we have ,
Then according to a criterion in Reference G, is a stationary measure, contradicting Proposition 3.1.
Now is continuous on since there is no atom of , Next assume that there is . and an interval in such that Let . be a maximal such interval and write it as in a distinguished chart such that Then in that chart the function . is nonnegative harmonic and takes value at Therefore we have .
This equality holds also in neighbouring distinguished charts and therefore over all Thus if we delete the saturation of . from the function , is still well defined and continuous. Deleting all such open saturated sets, we get a new manifold, still denoted by and a new foliation, still denoted by , .
For the new foliation the function , is continuous and strictly increasing on the interior of each fiber It may not be continuous on the boundary. However it is possible to extend the function . to the boundary by continuity, thanks to the leafwise harmonicity of The new function . is still constant on each boundary component of After normalization, one may assume that . on , Let . be an open ball in centered at Define a special kind of distinguished chart . by using the value along the fiber That is, . , Notice that . is the Lebesgue measure in this chart.
Given two values the function , is a positive harmonic function on By the Harnack inequality, there is . independent of and such that
Since this shows that there is , such that
for any and Then . is Lipschitz and thus differentiable everywhere. Precisely we have -almost
On the other hand, since the space of harmonic functions taking values in is compact, (Equation 4.2) shows that the upper partial derivative defined by and the lower partial derivative exists for any and is a harmonic function of Notice also that .
By (Equation 4.3), we have for any Then by Fubini, there is a Lebesgue full measure set . of such that for any ,
holds for all -almost But then ( .Equation 4.4) holds for any since , is harmonic in .
Writing the common value by as usual, we can define a measure , on by
This measure does not depend on the choice of the center of Thus . can be defined on the whole Again by a criterion in .Reference G, is a stationary measure. This contradicts Proposition 3.1, finishing the proof of Lemma 3.2.
5. Discrete group actions
Let be a group with a finite symmetric generating set We put a probability measure . on such that for any ,
For completeness we write if .
Assume acts continuously on preserving the orientation. A continuous function , is called if for any -harmonic we have
The purpose of this section is to give an alternative proof to the following theorem due to R. Feres and E. Ronshausen. Our proof is parallel to a probabilistic proof of the nonexistence of bounded harmonic functions on the real line.
Let us define
Given and the , cylinder set -th is defined by
Denote by the (finite) generated by all the -algebra cylinder sets. Let -th be the generated by -algebra Define a probability measure . on by That is, .
Let be the completion of by the measure For a bounded . function -measurable the conditional expectation by , is denoted by It is a . function, that is, simply a function constant on each -measurable cylinder set. Thus it can be thought of as a function of the -th product -ple of Now any . has the form
and the probability conditioned on is Therefore we have .
If furthermore the function is i.e. a function on -measurable, then we have ,
For define a random variable , by
Now let us embark upon the proof of Theorem 5.1. Assume for contradiction that there is a continuous function -harmonic such that , .
Therefore by the discrete martingale convergence theorem (Reference O, Appendix C) we have the following.
But it is shown in Reference DKNP that and , surely, provided the probability measure -almost satisfies (1) and (2). The contradiction shows Theorem 5.1.
Finally we have the following consequence of Theorem 5.1.
Acknowledgements
The author is grateful to R. Feres and K. Parwani for useful information.