Slices
- f
Thurston
's
Master Teapot
Kathryn Lindsey
Boston College
Master Teapot Kathryn Lindsey Boston College on preprint My talk - - PowerPoint PPT Presentation
Slices of 's Thurston Master Teapot Kathryn Lindsey Boston College on preprint My talk today based is " Master Teapot " A characterization Thurston 's of Wu Chen xi with joint on a previous preprint , This builds " Master
Slices
Thurston
's
Master Teapot
Kathryn Lindsey
Boston College
My talk today
is
based
Thurston's
Master Teapot
"joint
with
Chenxi
Wu
This builds
's
Master Teapot
"
joint with Diana Davis
,Harrison Bray
,Chen Ki Wu
Chenxi Wu
Also
a bitabout a
work
in progressjoint with Giulio
TI 0220
Diana Davis
Hamson Bray
Giulio TI 0220
study topological entropy
periodic
, unimodalself
↳ "
point
"is:*
.
±:
Cutting the interval at postcritical set
⇒ Markov partition
.the
critical point
matrix e.g
. quadratic 'is the gt:=e
polynomials
rate
is
a wreak¥3 A positive
, real , algebraicinteger that
is z the normall
its
Galois
conjugates
.Thurston's
Master Teapot
is
the
set
TIP
: = {ftp.ccx/R:XisthegrowthrateofsomefeF
andzisaGaloisconj#tof]}?
Two plots of finite approximations
)
unimodal
, criticallyperiodic self
Why study the Master Teapot ?
Rich + mysteryYasctare .
gives
a necessary
condition for
a weakPerron
number to
be the growth rate
a critically per
. unimmodapa!Uniform expanders
are
l
I
' i nwhat are the growth
rates of pseudo
surface diffeos ?
are
analogues of the
"Mandelbrot set for
"restricted " iterated functor systemshey7G { ZEE
: Cz ,Etc}
[
a goal of my talk is to explain this
.Iterated
Function Systems (IFS)
"Normal " iterative dynamics studyIN
by
a single function ,e.g
.fncx)
For
an IFS generatedby maps
fi , fz
, . . . , fo , youlook at the action of the
semi
by
the fi 's
. iunder
all words
inthe alphabet f. ,
. . . , fo .Usually , you want
the fi 's to be contractions of a metric space
. A contracting IFShas
a limit , a unique closed , boundedset K
s .t .K
= Ufick) .The
image of
any
closed , bounded set
wider the set of all
length
words
→ Imit set
in Hausdorff topology .Equivalently: for a sequence
w
set Glatz)
fyofwzo
. . . . Cz?(Notice
: Glwit) doesn't actually dependlimit set
= { Gcw , z) : w is a sequence} .Every
unimodal , critically periodic
map
issemi
to
a criticallyf×
periodic
' 'tent map "with the
sameentropy :O
fun branch
g.wana.⇒
.Itineraries :
It,
issequence
corresponding to
which sub interval the
belongs to
under fx
.fo
, >Cx)Ambiguity
when
Ya
.if
, , x Cx)For a critically periodic X
,It,
is periodic : It, = wit for somefinite word w
and there
is a canonicalchoice of
w.
( you choose the word that is minimal land has positive cumulative sign ) under the twisted lexicographical ordering . . This agrees with itinerary from kneading theory .)Twisted lexicographic ordering
so
, BN . ✓ µ, ⇒ ,For
v = yvzv ,
. . -w=w, wzw ]
. . . .if
Prefix ( v)
but
w ,⇐, = ,then {
Kew
if
E.iri
is even -WCEV
if
isI
Thm : 7<72 ⇒ It, Le Itx
,The Parry polynomial for a word
w=w
, wz . . . wewith positive cumulative sign
PwC⇒ ' = fay , z
Notice :
if X
is a critically periodicslope ,
then It, 4)
=w
and
PwC X)
= f-we , ×
Ci)
So
Pw
is
a polynomialwith integer cuffs
. sPw G)
= O .→ growth
rate X
isa root of Pw
.Cand
ball Galois conjugates of X)
( Pw
is the product of a cycloTomic polynomial and the "kneading polynomial ")PwC# ydotomic factor
may or may
not be irreducible
in ZEES
.Note
:For
aGalois conjugate
f-
we ,>
,
, pro"'
.This
is
a reason why
Galois conjugates
"matter "follow
under
words
in fi
, y 'sI call the
image
Cl of the
Master
Teapot
the
"Thurstonset " and denote
it r?
cpTicino proved that RIP n D
is theR
set of all roots of all power
series
Zwith
coeffsEquivalently
, it is the set of allZED
set
.the limit set Az
generated by fqz Cx) = Zx
{f ,,zCxI= 2- Zx
contains thepointt
For any word
w -and
Z E G , F- (w , z): = fwm
, zFor any
sequence
w = w , wz wz . . . and ZEB ,Image by
W . ThurstonGcw ,z) := highs fw
, ,zDIP AD
= { ZED :G (wit)
w}
Some results from
"the Shape of Thurston'sMaster Teapot
" (Beagconnected
.teapot contains the unit cylinder
S
' x Cl , 23 .The part of slices
inside the unit cylinder
are
monotone
increasing with the height a
,X
< 72⇒
¥
,E Hz
.↳In particular
, ri r ⑤ = Kzn ID .↳ which
we have seen iss
' u { ZEID :Gcw , 2-7=1 for some
seq
w}TheoremILL
)
:for any X e Cl
, 27 , the part of the slice Ex insidethe closed disk D-
can be characterized
as
Azn ⑤
= S 'U { ZEB :
GCW
, z) = I for
somela6k sequence w}Equivalently ,
Ze tf DD iff the
limit set
IFS
" in which you Inly allowwords that are 7-suitable
contains the pt 't .
X
a combinatorial condition
For each X E Cl , 27
,let My
= { XIt
isa
closed condition
For each
7 ,
My
= ?,µ× .( this implies the Persistence Theorem
)
Def
:A sequence
w is A-suitable if(1)
Reverse (Prefix CWDEE Prefix (Ita
,)FX
' > 7(2) If
equality Uther it has negative cumulate sign
.then
wdoes not contain
KH consecutive Os
.(3) If
It
= I ly
X
. K O's4) If AFL
, then a "renormalization of X satisfies theseconditions)
(
i.e .
if K
is the integer so that 52<7%2 ,then
w = DKCW ') forsome
sequence
w ' ,where D
is the renormalizationreplaces
It
and
Iwith 01
,and
for every X
'
> 7,2 "if It,y
then
w 'does not
KO
'scontain KH
Os
.(From
Bray
Renormalization
(Z , X) with
KATZ
is in the teapot ⇐( zzk
, y")
is in theteapot ,
where
k
isthe
integer
s -f .52 E X'
"s 2
.)Theorem 2
:For any XE Cl , 27
, the part of the slice It,
74,3 ④
={ z e El ID
:H ( Ita , z)
= 03 whereHlw
, z) : = Living⇐
' wiz
Theorems
I
and 2give
a way to verify that a point is not in a slice Hy .This
can be implemented algorithmically .form
,For points
z
with
Hal
.you look at expressions
far bigger
and if the value of the expression
isever
"toobig
" ,then z
can't
be
in X-p
.For points
z
with IZKI and Xt
' D :satisfy
Let UN
, × = { length N words thatconditions
CD - CD of Def . of 7-stability}Proposition
:z¢E, ⇐
F N
s .t . for every word wTest bigger and bigger N's
,fun
,I ' ° . . . Ofw ,,z" G) > ¥z,
The height
I -8 sliceIt
,.gr④ plotted 2 waysconstructive plot all roots of Parry
polynomials of critically
pggnwidhirazapsei.us
'ttahdusing
theorem I
algorithm
critical period
E 29 .white pts
areshown to
be in
black pts
in teapotthe complement of *
i.sowhite pls undetermined
black pts undetermined
.(Tested all this
, µfor
N E 18)
As
an applicationTheorem 3
:the part of the
Master Teapot
inside the unit cylinder
is not symmetricalwith respect to reflection across the imaginary
axis
.Galois conjugates
that
Cxtiy
, 7) trip ⇐Cx - iy , x)
c- NiP .Theorem
3
is surprising because the Thurston set REPthe projection of RCP to e
under
z to
Pf
: RIP ND is the roofs of all power series with± I
coeffs .If
Z
is aroot of such a power
series ,
the power
series formed
by flipping signs of
degree coefficients
.Because of renormalization ,
we
know this asymmetry
is confined
to
the part of the teapot
above height
52
.is to prove that
roots
in
TD of
reducible Parry polynomials don't matter
roots are
in the teapot
:Theorem :
Fox K X
c 2 .For each I > X
,define
YI
: =S
'U { ZED
: Pw Cz) = o for some admissible wordw
Then
Han DT
=Yj
.set
.w- Ee Iti?
This
is usefulbecause
you don't have to worry about whether roots
are
Galois conjugates of the growth rate
.After that
,a key technical result
is that ifWo
isa
"dominant "word such that
It, E WF
and
v
is a prefixsequence ( and
some other conchtons) ,then
Wo
. Reverse (v)is
admissible
.this fact is the motivation for the definition
X
The
reason this is useful
is that the beginning of a word determinesthe
leading
root
and the end determines roots in D
.Th I says
:the
slice It,
n D
= { z e ID : l Elimit set of 7- suitable seqs}
analogy
isIt , →
Mandelbrot sets (this
is * the connectednesslocus)
limit set
→
filled Julia set
fo CZ) = Etc
filled Julia set Rlfc)
= Ezek : fo# → a}seq
JCfc)
= 2K#c)Mandelbrot set M
= Ec :OE K Cfc)}
.Theoremasbyymptaoylea.iysas.ysmi.gr?orcaMisiearwiczpt
, M and Iare
at c
.B
.Solomyak investigated asymptotic similarity for the connectedness
locus
and
limit sets .
Proved asymptotic self
locus and limit sets for parameters z that are
" like "Misiearwkz pts.
Verified that
a few " landmark " pts fall in this category .Conjecture :
For each slice It,
and
each point
z e 2¥ ,
H
at z
isasymptotically similar to the restricted
IFS
limit set .at I
.slice
slice
limit
set
limit limit set
set
we hope to prove ldlsprore the
slice
conjecture
in future work .A related
, ongoing project(with
G
. Tiozzo) : "Bottchercoordinates
" for the Thurston set REPAIDcomplex dynamics IFS Thurston set
.For
ZED
,define
/
141mm #it:S :
# Hui're
. . . otwi ,d l l
, Az)Bottcher coordinatesfor
the Mandelbrot set M
"polar
" coordinatesequipotential
measure "rateat which critical pt → -
"Clear that 114mm G) = O
⇐
ZERIP
.Working to
showlevel sets
141mm
areJordan curves
,114mm
is differentablemost places ,
ele
.Thank you !