Upload
vuphuc
View
220
Download
1
Embed Size (px)
Citation preview
Lyapunov exponents ofTeichmüller flows
Marcelo Viana
IMPA - Rio de Janeiro
Lyapunov exponents ofTeichmuller flows – p. 1/61
Abelian differentials
Abelian differential = holomorphic
�
-form ��� � � ��� � on a(compact) Riemann surface.
Adapted local coordinates: then
near a zero with multiplicity :
then
Adapted coordinates form a translation atlas: coordinatechanges near any regular point have the form
Lyapunov exponents ofTeichmuller flows – p. 3/61
Abelian differentials
Abelian differential = holomorphic
�
-form ��� � � ��� � on a(compact) Riemann surface.
Adapted local coordinates:
� � �� � � � � then �� � �
near a zero with multiplicity �:
� � �� � � � �
�� � �
then � � � � � ��
Adapted coordinates form a translation atlas: coordinatechanges near any regular point have the form
Lyapunov exponents ofTeichmuller flows – p. 3/61
Abelian differentials
Abelian differential = holomorphic
�
-form ��� � � ��� � on a(compact) Riemann surface.
Adapted local coordinates:
� � �� � � � � then �� � �
near a zero with multiplicity �:
� � �� � � � �
�� � �
then � � � � � ��
Adapted coordinates form a translation atlas: coordinatechanges near any regular point have the form
� � � � ��� �� � �� � �
Lyapunov exponents ofTeichmuller flows – p. 3/61
Translation surfaces
The translation atlas defines
a flat metric with a finite number of conical singularities:
PSfrag replacements
��� � �� � �
a parallel unit vector field (the “upward” direction) on thecomplement of the singularities.
Conversely, the flat metric and the parallel unit vector fieldcharacterize the translation structure completely.
Lyapunov exponents ofTeichmuller flows – p. 4/61
Geometric representation
Consider any planar polygon with even number of sides,organized in pairs of parallel sides with the same length.
WE
N
S
PSfrag replacements
Identifying sides in the same pair, by translation, yields atranslation surface. Singularities arise from the vertices.
Every translation surface can be represented in this way,but not uniquely.
Lyapunov exponents ofTeichmuller flows – p. 5/61
Moduli spaces
� � moduli space of Abelian differentials on Riemannsurfaces of genus � � �
a complex orbifold of dimension
�� �� � � � �
a fiber bundle over the moduli space � of Riemannsurfaces of genus � � �
stratum of Abelian differentials havingzeroes, with multiplicities
complex orbifold with
Lyapunov exponents ofTeichmuller flows – p. 6/61
Moduli spaces
� � moduli space of Abelian differentials on Riemannsurfaces of genus � � �
a complex orbifold of dimension
�� �� � � � �
a fiber bundle over the moduli space � of Riemannsurfaces of genus � � �
�� � ��� � � � � � � � stratum of Abelian differentials having �
zeroes, with multiplicities � ��� � � � � � �
���� �
� � � � � ��
complex orbifold with
�� �� �� � �� � � � � � � � � � � � �
Lyapunov exponents ofTeichmuller flows – p. 6/61
Teichmüller geodesic flow
The Teichmüller flow is the natural action
�
on the fiberbundle � by the diagonal subgroup of
� � � ��
� :
� � � � � � � �� �� � � � � � �� ��
Geometrically:
PSfrag replacements�
The area of the surface
is constant on orbits of this flow.
Lyapunov exponents ofTeichmuller flows – p. 7/61
Ergodicity
Masur, Veech:Each stratum of � carries a canonical volume measure,which is finite and invariant under the Teichmüller flow.
The Teichmüller flow
�
is ergodic on every connectedcomponent of every stratum, restricted to any hypersurfaceof constant area.
Kontsevich, Zorich:Classified the connected components of all strata. Eachstratum has at most
connected components.
Eskin, Okounkov, PandharipandeComputed the volumes of all the strata.
Lyapunov exponents ofTeichmuller flows – p. 8/61
Lyapunov exponents
Fix a connected component of a stratum. The Lyapunovspectrum of the Teichmüller flow has the form
� � � � �� ��� � � � � � ��� � � �� � � � � � � ��� �� � � � � �� � �
� � � �� �� � � � � � � � � � �� � � � � � � ��� �� � � � � � � � �
Theorem (Avila, Viana).
� � � ��� � � � � � �
.
Conjectured by Zorich, Kontsevich.
Veech proved � � �
(flow is non-uniformly hyperbolic)
Forni proved � � �
(exactly
� � �
exponents � � �
)
Before discussing the proof, let us describe an application.
Lyapunov exponents ofTeichmuller flows – p. 9/61
Asymptotic cycles
Given any long geodesic segment � in a given direction,“close” it to get an element
� � � of
� � � �
�
:PSfrag replacements
Kerckhoff, Masur, Smillie: The geodesic flow in almostevery direction is uniquely ergodic.
Then
� � � �� ��
converges uniformly to some � � � � � � �
�
when the length
� �� � , where the asymptotic cycle � �
depends only on the surface and the direction.
Lyapunov exponents ofTeichmuller flows – p. 10/61
Asymptotic flag in homology
Corollary (Zorich). There are subspaces
� � � � � ��� � � � � � of
� � � �
� � � � �
with
� � � � � � �
, such that
the deviation of
� � � from
� � has amplitude
� �� �� � �for all
� � �
�� � �
� � ��� ���� � �� � � �� ��� ��� ��� �
�� � � ! "� � �
the deviation of
� � � from
� � is bounded ( � � genus of
).
PSfrag replacements
# �
� � �
Lyapunov exponents ofTeichmuller flows – p. 11/61
Zippered rectangles
Almost every translation surface may be represented in theform of zippered rectangles (minimal number of rectanglesrequired is
� � � � �
):
PSfrag replacements
�
�
�
�
�
�
�
�
Lyapunov exponents ofTeichmuller flows – p. 12/61
Coordinates in the stratum
This defines local coordinates
��� ��
� �� �
in the stratum:
� �
� � � �
� � � � describes the combinatorics of
the associated interval exchange map
� � � �� ��� �� � ����
are the horizontal coordinates ofthe saddle-connections (= widths of the rectangles)
� � � �� � � � �� � ��
are the vertical components of thesaddle-connections
� � � � has complex dimension
� � � � � �
The heights� � � �� �� ��� ���
of the rectangles arelinear functions of �
Lyapunov exponents ofTeichmuller flows – p. 13/61
Poincaré return map - 1st step
We consider the return map of the Teichmüller flow to thecross section
�
��
� � � �
.
PSfrag replacements
�
�
�
�
�
�
�
�
Lyapunov exponents ofTeichmuller flows – p. 14/61
Poincaré return map - 1st step
We consider the return map of the Teichmüller flow to thecross section
�
��
� � � �
.
PSfrag replacements
�
�
�
�
�
�
�
�
In this example: the top rectangle is the widest of the two.
Lyapunov exponents ofTeichmuller flows – p. 15/61
Poincaré return map - 1st step
We consider the return map of the Teichmüller flow to thecross section
�
��
� � � �
.
PSfrag replacements
�
�
�
�
�
�
�
�
Lyapunov exponents ofTeichmuller flows – p. 16/61
Rauzy-Veech induction
This 1st step corresponds to
�� ��
� �� � �� � ��� ��
� �� � �� � �
, with
� � �
� � � �
� � � � (top case)
� �� � �
� except that
� �� � ��� ��
same for �
� �� � �
� except that� �� � �� � ��
Write
� � � � � �
where� � ����� � is the linear operator
defined in this way. Then
� � � � � � �
and analogously for �.
Lyapunov exponents ofTeichmuller flows – p. 17/61
Poincaré return map - 2nd step
We consider the return map of the Teichmüller flow to thecross section
�
��
� � � �
.
PSfrag replacements
�
�
�
�
�
�
�
�
Lyapunov exponents ofTeichmuller flows – p. 18/61
Poincaré return map - 2nd step
We consider the return map of the Teichmüller flow to thecross section
�
��
� � � �
.
PSfrag replacements
�
�
�
�
�
�
�
�
Lyapunov exponents ofTeichmuller flows – p. 19/61
Rauzy-Veech renormalization
This 2nd step corresponds to
��� ��
� �� � �� � � � � ��� � ��
� � �� � � �� � � �
,with
� � � � � �
� � � � � �� �
� � � � � � �
and
� � � � � � �
where � � �� �
� is the normalizing factor.
The Poincaré return map
��� ��� ��
� � � � ��� � ��
� � �� � � � is called
(invertible) Rauzy-Veech renormalization.
Lyapunov exponents ofTeichmuller flows – p. 20/61
Rauzy-Veech cocycle and long geodesics
Consider the linear cocycle over the map
�
defined by
��� � �� ��
� �� � �� � � � ��� ��
� � � � � � �
� � ��� � �
PSfrag replacements
Consider a geodesic segment crossing each rectanglevertically. “Close” it by joining the endpoints to some fixedpoint in the cross-section. This defines some .
Lyapunov exponents ofTeichmuller flows – p. 21/61
Rauzy-Veech cocycle and long geodesics
Consider the linear cocycle over the map
�
defined by
��� � �� ��
� �� � �� � � � ��� ��
� � � � � � �
� � ��� � �
PSfrag replacements
�
�
�
�
Consider a geodesic segment crossing each rectangle �
vertically. “Close” it by joining the endpoints to some fixedpoint in the cross-section. This defines some � �
� � � � �
�
.
Lyapunov exponents ofTeichmuller flows – p. 21/61
Rauzy-Veech cocycle and long geodesics
Consider the linear cocycle over the map
�
defined by
��� � �� ��
� �� � �� � � � ��� ��
� � � � � � �
� � ��� � �
PSfrag replacements
� �
�
�
except that
In other words, . “builds” long geodesics.
Lyapunov exponents ofTeichmuller flows – p. 22/61
Rauzy-Veech cocycle and long geodesics
Consider the linear cocycle over the map
�
defined by
��� � �� ��
� �� � �� � � � ��� ��
� � � � � � �
� � ��� � �
PSfrag replacements
� �
�
�
��
� � � � except that ��� � �� � ��
In other words, �� � � �
�
.
��� “builds” long geodesics.
Lyapunov exponents ofTeichmuller flows – p. 22/61
Lyapunov exponents and asymptotic flag
So, it seems likely that
the behavior of long geodesics be described by theasymptotics of the Rauzy-Veech cocycle, the flag� � � � � �� � � � � � being defined by the Oseledetssubspaces of
��� with positive Lyapunov exponents.
the Lyapunov exponents of� � (for an appropriate
invariant measure) be related to the Lyapunovexponents of the Teichmüller flow.
This suggests we should try to prove that
� � has � positiveLyapunov exponents and they are all distinct. There is onetechnical difficulty, however:
Lyapunov exponents ofTeichmuller flows – p. 23/61
Zorich cocycle
The Rauzy-Veech renormalization does have a naturalinvariant measure (on each stratum) related to the invariantvolume of the Teichmüller flow. But this measure is infinite.
Zorich introduced an accelerated renormalization and anaccelerated cocycle
and
where is smallest such that the Rauzy iterationchanges from “top” to “bottom” or vice-versa.
This map has a natural invariant volume probability (oneach stratum) and this probability is ergodic.
Lyapunov exponents ofTeichmuller flows – p. 24/61
Zorich cocycle
The Rauzy-Veech renormalization does have a naturalinvariant measure (on each stratum) related to the invariantvolume of the Teichmüller flow. But this measure is infinite.
Zorich introduced an accelerated renormalization and anaccelerated cocycle
� ��� ��
� � � � � ��� ��
� � and��� ��� ��
� �� � � � �� ��� ��
� �� �
where � � � �� ��
is smallest such that the Rauzy iterationchanges from “top” to “bottom” or vice-versa.
This map
�
has a natural invariant volume probability (oneach stratum) and this probability is ergodic.
Lyapunov exponents ofTeichmuller flows – p. 24/61
Zorich cocycle
The Lyapunov spectrum of the Zorich cocycle� � has the
form
� � �� � � � �� � � �� � � � � � �� �� � � � � �
and it is related to the spectrum of the Teichmüller flow by
� �� �
� �� � �
� � � � � � �
Theorem (Avila, Viana). .
Forni: proved (cocycle is non-uniformly hyperbolic)
Lyapunov exponents ofTeichmuller flows – p. 25/61
Zorich cocycle
The Lyapunov spectrum of the Zorich cocycle� � has the
form
� � �� � � � �� � � �� � � � � � �� �� � � � � �
and it is related to the spectrum of the Teichmüller flow by
� �� �
� �� � �
� � � � � � �
Theorem (Avila, Viana).� � � � � ��� � � � �
� � �
.
Forni: proved
�� � �
(cocycle is non-uniformly hyperbolic)
Lyapunov exponents ofTeichmuller flows – p. 25/61
Lecture # 2
Simplicity criterium for Lyapunov spectra
Lyapunov exponents ofTeichmuller flows – p. 26/61
Linear cocycles
A linear cocycle over a map
� � � is an extension
� � � � � � � � ��
� ��� � � � � � ���
�� ���
��
where
� � � � � ��
�
. Note
� � ��� � � � � � � ���
�� � ��
�
with
� � ��� � � � � � � � ��� � � � � � � ��� � � ��
.
Let � be an
�
-invariant ergodic probability on such that
� � � � � � � � � � . Then the Oseledets theorem applies.
We say
�
has simple Lyapunov spectrum if all subspaces inthe Oseledets decomposition are
�
-dimensional.
Lyapunov exponents ofTeichmuller flows – p. 27/61
Hypotheses
Assume that
� � � � �
is the shift map on
� � � �
, where thealphabet
�
is at most countable
Denote
� � � � � � � � ������ � �� � � � � � � � for all � � � �
� � � � � � � � � ��� ���� � � � � � � � � � for all � � � �
Denote by
� � � � � � �� � � � � �� � � � �� ��
the cylinder of sequences
� � �
such that � � � � � for all
� � �� � � � � ��
�� � � � � �.
Lyapunov exponents ofTeichmuller flows – p. 28/61
Hypotheses
Assume that
the ergodic probability � has bounded distortion: � ispositive on cylinders and there exists
� � �
such that
��
�
� � � � � � � � �� � � � � �� � � � � � � �� ��
� � � � � � � � �� � � � � � � � �� � � � � � � �� �� � �
for every
� � � � � �� �� � � � �� � and every � � � � �.
is locally constant:
More generally: is continuous and admits stable andunstable holonomies.
Lyapunov exponents ofTeichmuller flows – p. 29/61
Hypotheses
Assume that
the ergodic probability � has bounded distortion: � ispositive on cylinders and there exists
� � �
such that
��
�
� � � � � � � � �� � � � � �� � � � � � � �� ��
� � � � � � � � �� � � � � � � � �� � � � � � � �� �� � �
for every
� � � � � �� �� � � � �� � and every � � � � �.
�
is locally constant:� �
� � � �� � ��� � �� � � � � � � � � � � �
More generally:�
is continuous and admits stable andunstable holonomies.
Lyapunov exponents ofTeichmuller flows – p. 29/61
Pinching
Let
� � � � ��
�
be the monoid generated by the� � �
,� � �
. We call the cocycle
�
(pinching) if for every
� � �
there exists some
� � �
such that the
�
-eccentricity
�� � � ��
� � �
���� �
� �for every
� � � � .
1
PSfrag replacements
�
� �
��
� �
Lyapunov exponents ofTeichmuller flows – p. 30/61
Twisting
Let
� � � � ��
�
be the monoid generated by the� � �
,� � �
. We call the cocycle
�
twisting if for any
� � ��� � � � � ��
� �
and any finite family� ��� � � � ���� � ��� � � � � �
�� �
there exists some
� � �
such that
� � � � � � � � � �
for all� � �
� � � � �
�
.
simple if is both pinching and twisting.
Lemma. is simple if and only if is simple.
Lyapunov exponents ofTeichmuller flows – p. 31/61
Twisting
Let
� � � � ��
�
be the monoid generated by the� � �
,� � �
. We call the cocycle
�
twisting if for any
� � ��� � � � � ��
� �
and any finite family� ��� � � � ���� � ��� � � � � �
�� �
there exists some
� � �
such that
� � � � � � � � � �
for all� � �
� � � � �
�
.
simple if
�
is both pinching and twisting.
Lemma.
�
is simple if and only if
� � � � � � � �� � � � �
is simple.
Lyapunov exponents ofTeichmuller flows – p. 31/61
Criterium for simplicity
Theorem 1 (Bonatti-V, Avila-V). If the cocycle
�
is simple then itsLyapunov spectrum is simple.
Guivarc’h, Raugi obtained a simplicity condition for productsof independent random matrices, and this was improved byGol’dsheid, Margulis.
Later we shall apply this criterium to the Zorich cocycles.
Right now let us describe some main ideas in the proof.
Lyapunov exponents ofTeichmuller flows – p. 32/61
Comments on the strategy
If the first
�
exponents are strictly larger, the correspondingsubbundle
� ��� �� ��� � � � � �
is an “attractor” for the actionof
�
on the Grassmannian bundle
� � ��� � � � � ��
� �
:
PSfrag replacements�
�� � � � �
���� � � �
�� � � � �
��� �� � ���� � � � � �
� � � � �
Notice that
� �� is constant on local unstable sets
������
.
Lyapunov exponents ofTeichmuller flows – p. 33/61
Invariant section
Proposition 1. Fix
�
. Assume the cocycle
�
is simple. Then there
exists a measurable section
� � � � ��� � � � � ��
� � such that
1.
�
is constant on local unstable sets and
�
-invariant, that is,� ��� � �� � � � � ���
at �-almost every point.
2.
�� � � ��
� � � � � � ��� � and the image
� ��� � � of the most
expanded
�
-subspace converges to� ���
3. for any
� � ��� � � � � ��
� � , we have that
� ���
is transverse to
�
at �-almost every point.PSfrag replacements
� � � �� � � � � � � � � �� � �
Lyapunov exponents ofTeichmuller flows – p. 34/61
Proof of Theorem 1
Applying the proposition to the inverse cocycle and tothe dimension
�
, we find another invariant section
� � � � � ��� � � � � ��
� �
with similar properties for
� � �
.
Recall
�
is simple if and only if� � �
is simple.
Keeping in mind that is constant on unstable setsand is constant on stable sets, part 3 implies that
at -almost every point.
Lyapunov exponents ofTeichmuller flows – p. 35/61
Proof of Theorem 1
Applying the proposition to the inverse cocycle and tothe dimension
�
, we find another invariant section
� � � � � ��� � � � � ��
� �
with similar properties for
� � �
.
Recall
�
is simple if and only if� � �
is simple.
Keeping in mind that�
is constant on unstable setsand
� � is constant on stable sets, part 3 implies that
� ��� � � � �� � � �at �-almost every point.
Lyapunov exponents ofTeichmuller flows – p. 35/61
Proof of � �
�PSfrag replacements
� �
� �
�� �� � �
�� � � � �
Suppose the claim fails on a set� � �
with � � � � �
.
By local product structure, there exist points � � �
such that�� � � �
has positive �-measure, where
� � ������ � � �
.
Define
� � � ��� �. Then
� � � � �
for all � ������ �
and
� �
is not transverse to
�
on
� � � �
. This contradicts part 3.
Lyapunov exponents ofTeichmuller flows – p. 36/61
Proof of Theorem 1
From part 2 we get that the Lyapunov exponents of� � � �
are strictly larger than the ones of� � � .
PSfrag replacements
� � � � �
� � � �� � �
� �Using that
� � ����
� �
is close to
� ��� �
for large �, one finds aninvariant cone field around
� �for the cocycle induced by
�
on some convenient subset (with large return times).
Hence, there are�
Lyapunov exponents (for either cocycle)which are larger than the other ones. Theorem 1 follows.
Lyapunov exponents ofTeichmuller flows – p. 37/61
-states
A probability � on
� � �� �� � � ��
� � �
is a �-state if it projectsdown to � and there is
� �
such that
� �� �� � � � � � �� �� � � � �� ��� � �
� ���� �� � � � �� � � � � � � � �� � � � � � �� �� � � � ��� ��� � �
� � � �� �� � � � ��� � �
for every
"! � � � � �
$#� � � � �
"% �
&#� � � � �
&"' and
( ) *,+ �� � �-�
� . �
.
PSfrag replacements
/ �
/0
12$34 5 5 54 2 0 6 7
1 2984 5 5 54 2$: 7 1; 84 5 5 54 ;$< 7
Lyapunov exponents ofTeichmuller flows – p. 38/61
-states
Equivalently, � is a �-state iff it may be disintegrated as
� ��
��� ��� ��� ��
��� on* + �� � �-
�
� . ��
where � � is equivalent to �� whenever � � ��� ���� �
, withderivative uniformly bounded by
�.
-states always exist: for instance, for anyprobability in the Grassmannian.
any Cesaro limit of iterates of a -state is invariantunder the cocycle .
Lyapunov exponents ofTeichmuller flows – p. 39/61
-states
Equivalently, � is a �-state iff it may be disintegrated as
� ��
��� ��� ��� ��
��� on* + �� � �-
�
� . ��
where � � is equivalent to �� whenever � � ��� ���� �
, withderivative uniformly bounded by
�.
�-states always exist: for instance, � � � � � for anyprobability � in the Grassmannian.
any Cesaro limit of iterates of a �-state is invariantunder the cocycle
� � * + �� � �-�
� . � � � � * + � � � �-�
� . �
.
Lyapunov exponents ofTeichmuller flows – p. 39/61
Proof of Proposition 1
Proposition 2. Let � be an invariant �-state and � be its projection to* + �� � �-�
� . �
. Then
the support of � is not contained in any hyperplane section of theGrassmannian
the push-forwards � � �� �
of � under� � �� � ��� � �
converge almost
surely to a Dirac measure at some point
� � ��� �
.PSfrag replacements
�
�0 � � � �
��� � � � �
�� � � � �
� �
� � � � �
Lyapunov exponents ofTeichmuller flows – p. 40/61
Proof of Proposition 1
Proposition 1 follows from Proposition 2 together with
Lemma. Let
� � � � . � � .
be a sequence of linear isomorphisms and
� be a probability measure on
*,+ �� � �-�
� . �
which is not supported in ahyperplane section
� � � * + �� � �-�
� . � � �� � � � � ��
� � * + �� � � � - �
� . �
If the push-forwards
� � ���� � converge to a Dirac measure
�� then the
eccentricity
��� � �-�
� �� � � and the images
� � � � ��
of the mostexpanded
-
-subspace converge to
�
.
Lyapunov exponents ofTeichmuller flows – p. 41/61
Proof of Proposition 2
Part 1: � is not supported in any hyperplane section.
This follows from the twisting hypothesis.
Part 2: the sequence � �
converges to a Dirac measure.
The proof has a few steps. The most delicate is to showthat some subsequence converges to a Dirac measure:
Lyapunov exponents ofTeichmuller flows – p. 42/61
Proof of Proposition 2
For almost every � there exist �� � � such that � �� �� �
converges to a Dirac measure.
By hypothesis, there is some strongly pinching . Theare chosen so that for some
convenient and large . This uses ergodicity.
PSfrag replacements
Lyapunov exponents ofTeichmuller flows – p. 43/61
Proof of Proposition 2
For almost every � there exist �� � � such that � �� �� �
converges to a Dirac measure.
By hypothesis, there is some strongly pinching
�# � �
. The
�� are chosen so that
� �� �� � �� ��� � �� ��� �%# ��� � '# for some
convenient
��� and large �
�. This uses ergodicity.
PSfrag replacements
/
/0 �
�0 �� � � �
Lyapunov exponents ofTeichmuller flows – p. 43/61
Proof of Proposition 2
Let � � � � ��� �
denote the projection to the Grassmannian ofthe normalized restriction of � to the cylinder
� � �
� �
�
that contains � .
The sequence � � � � ��� �
converges almost surely to someprobability � ��� �
on
* + � � � �-
� . �.
This follows from a simple martingale argument.
Moreover, for all , with .
This is a consequence of the definition of -state.
Thus, exists and is a Dirac measure, as stated.
Lyapunov exponents ofTeichmuller flows – p. 44/61
Proof of Proposition 2
Let � � � � ��� �
denote the projection to the Grassmannian ofthe normalized restriction of � to the cylinder
� � �
� �
�
that contains � .
The sequence � � � � ��� �
converges almost surely to someprobability � ��� �
on
* + � � � �-
� . �.
This follows from a simple martingale argument.
Moreover,
�� �
� � ��� �
� � � � ��� � � �
for all � , with
� � � � � � .
This is a consequence of the definition of �-state.
Thus, exists and is a Dirac measure, as stated.
Lyapunov exponents ofTeichmuller flows – p. 44/61
Proof of Proposition 2
Let � � � � ��� �
denote the projection to the Grassmannian ofthe normalized restriction of � to the cylinder
� � �
� �
�
that contains � .
The sequence � � � � ��� �
converges almost surely to someprobability � ��� �
on
* + � � � �-
� . �.
This follows from a simple martingale argument.
Moreover,
�� �
� � ��� �
� � � � ��� � � �
for all � , with
� � � � � � .
This is a consequence of the definition of �-state.
Thus,
� � � �� � � ��� �
exists and is a Dirac measure, as stated.
Lyapunov exponents ofTeichmuller flows – p. 44/61
Lecture # 3
Zorich cocycles are pinching and twisting
Lyapunov exponents ofTeichmuller flows – p. 45/61
Strata and Rauzy classes
On each stratum � � � �
� � � we introduced a system oflocal coordinates
� � � � � , where
� �
�#�� is a pair of permutations of the alphabet
�
and � live in convenient subsets of
� .
,
� � ��� � �
.
The pair � determines the stratum and even its connectedcomponent.
An (extended) Rauzy class is the set of all � correspondingto a given connected component of a stratum (there is alsoan intrinsic combinatorial definition).
Lyapunov exponents ofTeichmuller flows – p. 46/61
Zorich cocycle
Fix some (irreducible) Rauzy class
�
and consider � � �
.
Previously, we introduced the Zorich renormalization andthe Zorich cocycle for translation surfaces
� � � � � � � � � ��� � � � � � � � � �
��� � � � � � � � � � ��� � � � � � � � � � � �
, with � � � � � � �
Dropping , we find the Zorich renormalization and Zorichcocycle for interval exchange maps
For our purposes, is equivalent to the inverse limit of .
Lyapunov exponents ofTeichmuller flows – p. 47/61
Zorich cocycle
Fix some (irreducible) Rauzy class
�
and consider � � �
.
Previously, we introduced the Zorich renormalization andthe Zorich cocycle for translation surfaces
� � � � � � � � � ��� � � � � � � � � �
��� � � � � � � � � � ��� � � � � � � � � � � �
, with � � � � � � �
Dropping �, we find the Zorich renormalization and Zorichcocycle for interval exchange maps
� � � � � �� � � � � � � �
��� � � � �
� ��� � � � � � �
� � �
For our purposes,
�
is equivalent to the inverse limit of
�
.
Lyapunov exponents ofTeichmuller flows – p. 47/61
Checking the hypotheses
�
is a Markov map: There exists a finite partition
�#
and a countable partition
� � �# such that� �� � � �#
for every
� � �
.
There exists a -invariant probability which is ergodicand equivalent to volume . Moreover, has boundeddistortion.
The cocycle is constant on every atom of , and it isintegrable with respect to .
Lyapunov exponents ofTeichmuller flows – p. 48/61
Checking the hypotheses
�
is a Markov map: There exists a finite partition
�#
and a countable partition
� � �# such that� �� � � �#
for every
� � �
.
There exists a
�
-invariant probability � which is ergodicand equivalent to volume
� �
. Moreover, � has boundeddistortion.
The cocycle is constant on every atom of , and it isintegrable with respect to .
Lyapunov exponents ofTeichmuller flows – p. 48/61
Checking the hypotheses
�
is a Markov map: There exists a finite partition
�#
and a countable partition
� � �# such that� �� � � �#
for every
� � �
.
There exists a
�
-invariant probability � which is ergodicand equivalent to volume
� �
. Moreover, � has boundeddistortion.
The cocycle
� � is constant on every atom of
�
, and it isintegrable with respect to � .
Lyapunov exponents ofTeichmuller flows – p. 48/61
Zorich cocycle
There is an
� � -invariant subbundle
� �� � � � � � ����� � � of
the vector bundle
� � � � � � � � .
, such that the
fibers have dimension
��� and they depend only on thepermutation pair:
��� � � � ��
Lyapunov exponents of the cocycle
� � in the directionstransverse to
�� vanish (in a trivial fashion)
restricted Zorich cocycle�� �
preserves a symplecticform � �
� � ��
on this subbundle
Thus, the Lyapunov spectrum of the Zorich cocycle has theform
�� �� �� � � � � � �� � � � � � � � � � �� ��
Lyapunov exponents ofTeichmuller flows – p. 49/61
Zorich cocycles are simple
The goal is to prove that the
��� are all distinct and positive.
By Theorem 1 and previous observations, all we need is
Theorem 2. Every restricted Zorich cocycle is twisting and pinching.
The proof is by induction on the complexity ( � number ofsingularities, � � genus of the surface) of the stratum:
we look for orbits of
�
that spend a long time close to theboundary of each stratum (hence, close to “simpler” strata).
Lyapunov exponents ofTeichmuller flows – p. 50/61
Hierarchy of strata
Let � � � � �
� � � be any stratum. Collapsing two ormore singularities of an � � together (multiplicities add)one obtains an Abelian differential in another stratum, onthe boundary of � � � �
� � � .
PSfrag replacements
6 � � � � � � � �
We write � � � �
���� � � � � �
� � � if
� � �
����
canbe attained from
� � �
� � � by collapsing singularities.
�
�
� � � �
�
�
� �
but � � � � � � � �
�
� �
.
Lyapunov exponents ofTeichmuller flows – p. 51/61
Starting the induction
The case � � �
, � �
,
� � �
is easy:
There is only one permutation pair � �
� �
� � .
Top case is
��� � ��� and bottom case is
� � � �� , with
�� % �
� �
� � and
��� � �
� �� �
is hyperbolic and so its powers
have arbitrarily large eccentricity. This proves pinching.
Lyapunov exponents ofTeichmuller flows – p. 52/61
Starting the induction
The case � � �
, � �
,
� � �
is easy:
There is only one permutation pair � �
� �
� � .
Top case is
��� � ��� and bottom case is
� � � �� , with
�� % �
� �
� � and
��� � �
� �� �
� � �� % � � � �
� �
� � is hyperbolic and so its powers
have arbitrarily large eccentricity. This proves pinching.
Lyapunov exponents ofTeichmuller flows – p. 52/61
Starting the induction
To prove twisting, consider
�
and
��
��� � * + �� � � �
�� �
.
PSfrag replacements
��
��
Fix
�
large enough so that no
� � � % � ��
�
coincides with any ofthe eigenspaces of
�.
Then
� � � � � � � � � % � ��
��� � �
, that is,
� � % � � � � � � �� �
� � �
for all
and any sufficiently large �.
Lyapunov exponents ofTeichmuller flows – p. 53/61
Inductive step
Fix any pair � � �
and denote by
� � the submonoid of
�
corresponding to orbit segments
� �# �# �
� � � �
suchthat �# � � � � .
It suffices to prove that the action of� � on the space
�� issimple. The proof is by induction on the complexity of thestratum, with respect to the order � and to the genus.
This is easier to implement at the level of interval exchangetransformations. In that setting, approaching the boundarycorresponds to making some coefficient
��� very small.
Then it remains small for a long time, under iteration by therenormalization operator.
Lyapunov exponents ofTeichmuller flows – p. 54/61
Reduction and extension
We consider operations of simple reduction/extension:
� ��
�
� 6 � � � �� 0 6 � �� 6 � � � � � � � � � ��
� 6 � � � � � � � � � �� 0 6 � �� 6 � � � ��
�
� � ��
�
� 6 � � � � � 0 6 �� 6 � � � � � � ��
� 6 � � � � � � �� 0 6 �� 6 � � � ��
extension: inserted letter can not be last in either row and can notbe first in both rows simultaneously
Lemma. Given any � there exists � � such that � is a simple extension of
� �. Moreover, either � � � � � � � � � � or � � � � � � � � � � � �
.
Lyapunov exponents ofTeichmuller flows – p. 55/61
Proof of Theorem 2
For proving twsting and pinching, we take advantage of thesymplectic structure:
Proposition 3. The action of
� � twists isotropic subspaces of
� � .
That is, the twisting property holds for the subspaces
�
onwhich the symplectic form vanishes identically.
To compensate for this weaker twisting statement, we provea stronger form of pinching:
Lyapunov exponents ofTeichmuller flows – p. 56/61
Proof of Theorem 2
Proposition 4. The action of
� � on
�� is strongly pinching.
That is, given any
� � �
there exist
�
in the monoid for
which
��� � � � � �
and
� � � ��
��� � �� � �
� �
for all� � & � � .
For symplectic actions in
� � �, twisting = isotropic twisting
and pinching = strong pinching. In any dimension,
Lemma. isotropic twisting & strong pinching
�
twisting & pinching.
This reduces Theorem 2 to proving the two propositions.
Lyapunov exponents ofTeichmuller flows – p. 57/61
Proof of Proposition 3
Take � � such that � is a simple extension of � �.
If � � � � � � � � � � then there is a symplectic isomorphism�� � � �� that conjugates the action of� � � on
�� � to theaction of
� � on
�� .
Therefore, by induction,
� � does twist isotropic subspaces.
Lyapunov exponents ofTeichmuller flows – p. 58/61
Proof of Proposition 3
If � � � � � � � � � � � �
, there is some symplectic reduction� �� of
�� and some symplectic isomorphism
�� � � � ��
that conjugates the action of
� � � on�� � to the action
induced by
� � on
� �� .
symplectic reduction:
� � � quotient by � of the symplecticorthogonal of �, where � � � � � �
.� � � � � � � � � � �
induced action: the action on� �
of the stabilizer of �, thatis, the submonoid of elements of
�
that preserve �.
Lemma. If the action of on is minimal and its induced action ontwists isotropic subspaces, then twists isotropic subspaces of .
The action of on the projective space is indeedminimal, that is, all its orbits are dense.
Lyapunov exponents ofTeichmuller flows – p. 59/61
Proof of Proposition 3
If � � � � � � � � � � � �
, there is some symplectic reduction� �� of
�� and some symplectic isomorphism
�� � � � ��
that conjugates the action of
� � � on�� � to the action
induced by
� � on
� �� .
symplectic reduction:
� � � quotient by � of the symplecticorthogonal of �, where � � � � � �
.� � � � � � � � � � �
induced action: the action on� �
of the stabilizer of �, thatis, the submonoid of elements of
�
that preserve �.
Lemma. If the action of�
on� � � �
is minimal and its induced action on� �
twists isotropic subspaces, then
�
twists isotropic subspaces of
�
.
The action of� � on the projective space
� � ���
is indeedminimal, that is, all its orbits are dense.
Lyapunov exponents ofTeichmuller flows – p. 59/61
References
A. Avila, M. Viana, Simplicity of Lyapunov spectra: proof of the Zorich-Kontsevichconjecture. Preprint (www.preprint.impa.br) 2005.
A. Avila, M. Viana, Simplicity of Lyapunov spectra: a sufficient criterion. Preprint(www.preprint.impa.br) 2006.
C. Bonatti, M. Viana, Lyapunov exponents with multiplicity�
for deterministic productsof matrices. Ergod. Th. & Dynam. Syst. 24 (2004), 1295–1330.
G. Forni, Deviation of ergodic averages for area-preserving flows on surfaces of highergenus. Ann. of Math. 155 (2002), 1–103.
I. Ya. Gol’dsheid, G. A. Margulis, Lyapunov exponents of a product of randommatrices. Uspekhi Mat. Nauk 44 (1989), 13–60; translation in Russian Math. Surveys44 (1989), 11–71.
Y. Guivarc’h, A. Raugi, Products of random matrices: convergence theorems. Randommatrices and their applications, Contemp. Math. 50 (1986), 31–54.
S. Kerckhoff, H. Masur, J. Smillie, Ergodicity of billiard flows and quadratic differentials.Ann. of Math. 124(1986), 293–311.
Lyapunov exponents ofTeichmuller flows – p. 60/61
References
M. Kontsevich, A. Zorich, Connected components of the moduli spaces of Abeliandifferentials with prescribed singularities. Invent. Math. 153 (2003), 631–678.
F. Ledrappier, Positivity of the exponent for stationary sequences of matrices.Lyapunov exponents, Lecture Notes in Math. 1186 (1986), 56–73.
H. Masur, Interval exchange transformations and measured foliations. Ann. of Math.115 (1982), 169–200.
G. Rauzy, Echanges d’intervalles et transformations induites. Acta Arith. 34 (1979),315–328.
S. Schwartzman, Asymptotic cycles. Ann. of Math. 66 (1957), 270–284.
W. Veech, Gauss measures for transformations on the space of interval exchangemaps. Ann. of Math. 115 (1982), 201–242.
A. Zorich, Finite Gauss measure on the space of interval exchange transformations.Lyapunov exponents. Ann. Inst. Fourier 46 (1996), 325–370.
A. Zorich, How do the leaves of a closed
�
-form wind around a surface?Pseudoperiodic topology, Amer. Math. Soc. Transl. Ser. 2, 197 (1999), 135–178.
Lyapunov exponents ofTeichmuller flows – p. 61/61