The Geometry of Solvable Groups-Amin Saied

Spring 2013
MATH7320:
The Geometry of Solvable Groups
lecturer: Tim Riley
by: Amin Saied
Contents
I Tims Lectures 1
1 Jan 22, 2013 1
1.1 Basic Dentions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2 Jan 31, 2013 3
2.1 Examples of Solvable Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
3 Feb 05, 2013 5
3.1 Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
4 Feb 7, 2013 8
4.1 Idea behind Gromovs proof . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
4.1.1 Background: -algebras and Probability Measures . . . . . . . . . . . . . . . . . . . 9
4.1.2 Ultralters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4.1.3 Construction of Y . (van den Dries-Wilke) . . . . . . . . . . . . . . . . . . . . . . . . 11
5 Feb 12, 2013 11
5.1 Properties of Asymptotic Cones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
6 Feb 14, 2013 13
6.1 Filling Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
7 Feb 19, 2013 15
7.1 Combings of Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
8 Feb 21, 2013 18
8.1 Quasi-Isometry Invariants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
9 Feb 26, 2013 21
9.1 The Area Function of Nilpotent Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
10 Feb 29, 2013 23
10.1 Lower Bounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
11 Mar 5, 2013 25
11.1 Some Open Problems Relating to Dehn Functions . . . . . . . . . . . . . . . . . . . . . . . 25
11.2 HNN-Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
11.3 Baumslag-Solitar Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
12 Mar 7, 2013 28
12.1 The Geometry of Baumslag-Solitar Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
12.1.1 Normal Form for BS(1, n) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
12.1.2 Dehn Function of BS(1, m) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
12.1.3 Cayley Graph of BS(1, 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
12.1.4 Treebolic Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
II Student Talks 35
13 Geometry of Sol (Chenxi Wu) 35
13.1 Dehn Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
13.2 Quasi-Rigidity of Sol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
13.3 Area function becomes quadratic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Tim Riley The Geometry of Solvable Groups 3
14 Geometry of the Lamplighter Group (Margarita Amchislavska) 36
14.1 Dead-end Depth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
14.2 Horocyclic Product of Two Z-branching Trees . . . . . . . . . . . . . . . . . . . . . . . . . . 36
15 Generalisations of the Lamplighter Group (Margarita Amchislavska) 38
15.1 Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
15.2 Some Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
15.3 Horocyclic Product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
16 Geometry of the Magnus Embedding (Andrew Sale) 40
16.1 Magnus Embedding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
16.2 Applications of the Magnus Embedding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
16.3 Magnus Representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
16.4 Fox Calculus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
16.5 Wreath Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
16.5.1 Geometric Interpretation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
16.6 Geometric Denition of the Magnus Embedding . . . . . . . . . . . . . . . . . . . . . . . . . 42
17 Finitely Presented Metabelian Groups (Amin Saied) 44
17.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
17.2 Baumslag/Margaritas Example: Embedding
1
= Z / Z into
2
. . . . . . . . . . . . . . . . 45
17.3 Proof of Baumslag-Remeslennikov Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
18 Sigma Invarients (Teddy Einstein) 51
18.1 -invariants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
18.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
19 Sigma Invarients (Yash Lodha) 55
19.1 Finiteness Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
20 Automatic Groups (Scott Messick) 58
20.1 Closure Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
20.2 Biautomatic Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
20.3 Open Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
21 Quasi-Isometric Rigidity (Iian Smythe) 62
22 Quasi Isometric Rigidity (Kristen Pueschel) 65
22.1 Proof Form for QI Rigidity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
22.2 X
n
and its Boundary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
23 Random Walks on Solvable Groups (Johannes Cuno) 67
23.1 Random Walks on Graphs and Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
23.2 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
23.3 Flows and Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
23.4 Growth of Groups and Return Probability . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
23.5 Advertisement for Next Talks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
24 Boundaries of Random Walks (Mathav Murugan) 70
24.1 Martin Boundary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
24.2 Poisson Boundary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
25 Rate of Escape of Random Walks on Groups (Tianyi Zheng) 73
Part I
Tims Lectures
1 Jan 22, 2013
Two perspectives on studying solvable groups
1. Classical perspective; of Phillip Hall and Gilbert Bammslang
-Algebraic
-Representations
-Commutative Algebra
-Cohomological methods
-Bieri-Strebel Invarients
2. Geometric perspective;
-Growth
-Isoperimetry
-Rigidity, quasi-isometry
-Lattices
-Random walks
The aim of this course is to try and create a bridge between these two perspectives.
1.1 Basic Dentions
Denition A group G is solvable if it has a nite abelian series - that is,
G = G
n
G
n1
G
0
= 1 (1)
such that the factor groups are all abelian. The smallest such n is called the derived length of G.
1. Derived length 0: trivial groups
2. Derived length 1: abelain groups
3. Derived length 2: metabelian
e.g. G = G
2
G
1
1
Get the short exact sequence
G
1
G G/G
1
where G
1
and G/G
1
are abelian.
The class of solvable groups is closed under taking extensions, subgroups, homomorphic images. (Exersize:
show this for subgroups and homomorphic images: for subgroups just take intersection of the whole series,
for homomorphic images, just hit the whole series with the homomorphism f). For extensions consider,
1 K G H 1
K and H solvable. Now H

= G/K. Now H is solvable, so take its abelain series. Lift this to one for G,
and append it with the one for K. (also an exercise).
Denition The derived subgroup (or commutator subgroup) of a group G is G
= [G, G]. The

derived series of G is
G = G
(0)
G
(1)

where G
(i+1)
= (G
(i)
)
. Clearly G
(n)
/G
(n+1)
is abelian (this is the abelianisation!). So if this terminates
then we have a nite abelian series for G, and this G is solvable.
Claim If we have a nite abelian series (with the notation in 1) then G
(i)
G
ni
Pf: By induction. For i = 0 it is immediate. And G
(i+1)
= (G
(i)
)
(G
ni
)
G
n(i+1)
where the last two
inequalities come from the induction hypothesis and because the last group has abelian factors (resp.)
Corollary The derived length of a group is the length of the derived series.
Denition A group G is nilpotent if it has a (nite) central series, that is
G = G
n
G
n1
G
0
= 1
such that the factor groups G
i+1
/G
i
is contained in the center of G/G
i
(for all i).
Note// Since the center of a group is abelian (!) we see that nilpotent groups are examples of solvable
groups.
We call the minimal length n of the central series the class of G
1. Class 0: trivial groups
2. Class 1: abelain groups
Denition The upper central series of a group G is
1 = Z
0
(G) Z
1
(G) Z
2
(G)
dened by Z
n+1
(G)/Z
n
(G) = Z(G/Z
n
(G)). (In particular, since Z
0
(G) = 1, we see that Z
1
(G) = Z(G), is
the center) (Also note that this is clearly a central series, by denition)
The lower central series of a group G:
1
(G) = G, then
n+1
(G) = [
n
(G), G]
Notice that
n
(G)/
n+1
(G) Z(G/
n+1
(G)). (Exersize: check that this really is a central series) and
n
(G) G.
Theses series are related in the following way:
1.
i
(G) G
ni+1
2. G
i
Z
i
(G)
These can be proved by induction on i
Z
n
(G) Z
n1
Z
0
(G) = 1
| |
G = G
n
G
n1
G
0
= 1
| |
1
(G)
2
(G)
n+1
(G)
In this way we see that the lengths of these series agree if the central series was of minimal length.
Denition A group G is polycyclic if it has a nite normal series with cyclic factors.
Denition The max condition on subgroups if a group G is any one of the following equivalent condi-
tions:
1. Every subgroup is nitely generated (So F
2
doesnt satisfy the max condition)
2. Every ascending chain of subgroups
H
1
H
2
H
3

stabilises i.e. there exists N such that H
N
= H
N+1
= (Noetherian condition)
3. Every non-empty set of subgroups has a maximal element (a subgroup that is not a proper subgroup
of any other one of these subgroups)
We say G is max.
Properties of Max Notice that cyclic groups are max. Max is inherited when taking extensions (if H,
K are Max,and K G H SES, then G is Max, exersize). So polycyclic groups are Max.
Theorem Polycyclic groups are precisely the solvable groups that are Max.
Pf: If G is solvable and Max, then its subgroups are nitely generated, and so the factors in the derived
series are nitely generated abelian groups, and so the derived series can be rened to a cyclic series
2 Jan 31, 2013
The idea today is to ll in this map of dierent classes of groups with some examples.
Theorem (Hirsch) Finitely generated nilpotent groups are polycyclic.
Pf Recall
1
(G) = G,
n+1
(G) = [G,
n
(G)]. Let
c
(G) be the last non-trivial term in the lower central
series of a nitely generated nilpotent group G. By induction on c, we may assume that G/
c
(G) is poly-
cyclic (because by quotienting the lower central series for of G by
c
(G) gives a central series for G/
c
(G)
with one fewer term, so by induction, polycyclic).
Suppose G = x
1
, . . . , x
m
. As G/
c
(G) is polycyclic,
c1
/
c
(G) is nitely generated (by max condition).
Write
c1
(G) = y
1
, . . . , y
n
c
(G). Then
c
(G) = [x
1
, . . . , x
m
, y
1
, . . . , y
n
] = [x
i
, y
j
][1 i m, 1 j n
and so is nitely generated. But
c
(G) is abelian, and so is polycyclic. So as G/
c
(G) and
c
(G) are
polycyclic, and hence G is polycyclic too.
2.1 Examples of Solvable Groups
1. Abelian group: Z
k
the free abelian group
2. R a commutative ring with 1, U
n
(R) = (n n) upper triangular matrices over R, with 1s on the
diagonal.
_
_
_
_
_
1
1
.
.
.
0 1
_
_
_
_
_
These is nilpotent of class n 1. Indeed,
[e
ij
, e
kl
] =
_
1 j ,= k
e
il
j = k
e
ij
= I
n
+E
ij
(The idea is that terms get pushed upwards and away from the diagonal)
In particular, U
3
(Z) is the thee-dimensional integral Hiesenburg group, sometimes denoted
1
3
(Z) = a, b, c[[a, b] = c, [a, c] = [b, c] = 1
Notice that this is metabelian, as the map which kills c takes 1
3
(Z) ZZ. It is of course not abelian.
Theorem. (P.Hall, 1969) Every nitely generated torsion free nilpotent group embeds in U
n
(Z)
for some n.
3. Baumslag-Solitar BS(1, m) = a, b[b
1
ab = a
m
. b generates Z and a is the unit in Z[
1
m
]. We have
the presentation,
Z Z[
1
m
]
where a generator for the Z-factors acts on Z[
1
m
] by multiplication by m. This is metabelian by
construction.
But this is not a polycyclic group: indeed Z[
1
m
] is not nitely generated, and is clearly (isomorphic
to) a subgroup. This contradicts the max condition.
4. Sol = Z
2
A
Z, where A =
_
2 1
1 1
_
. This is metabelian and polycyclic by construction.
5. Lamplighters
Before dening lamplighters we require the more general denition of a wreath product. For groups
A, B denote A
(B)
:= nitely supported functions B A =
B
A
Denition. The restricted wreath product A wr B = A/ B = A
(B)
B with the action
(f, b)(g, c) := (fg
b
, bc)
where g
b
(x) = g(b
1
x).
Denition. A lamplighter is a restricted wreath product of the form
Z/mZ wr Z
k
For example Z wr Z or Z/2Z wr Z. Lets look at Z/2Z wr Z more closely to see how the name
lamplighter arises.
Example: A = Z/2Z wr Z
An element in A looks like (f, b) where f is a nitely supported function from Z/2Z to Z. If we think
of Z lamps arranged on a street, then f is a nite set of these lamps which are turned on. We could
then think of b as being someone at a given position on the street, say lighting the lamps. So
(f, b) =

1 0 1 2 3 4 5

means that f lights 1, 3, 4, and b = 1 is the position of the lamplighter. So lets consider the group
product. Consider the element
(g, c) =

1 0 1 2 3 4 5

where g lights 0, 1, 2 and the lamplighter is at position c = 3. The denition of the group product in
a wreath product uses g
b
(x) = g(b
1
x). NB: that we are using multiplicative notation, but the group
operation in Z is addition. So g
b
(0) = g(1) = 0, g
b
(1) = g(0) = 1 etc., thus we get
g
b

f
fg
b

1 0 1 2 3 4 5
Now we have (f, b)(g, c) = (fg
b
, bc). fg
b
is just addition (mod 2) of f and g
b
, and bc is just regular
addition in Z of b and c, hence
(f, b) (g, c) = (fg
b
, bc) =

1 0 1 2 3 4 5

To summarize this, we think of g
b
as a translating g by b, then fg
b
is just addition mod 2 of these
lights. The lamplighter just moves to the sum of other lamplighter positions.
Finite generation of Z/2Z wr Z:
f : Z Z/2Z by 0 1, r 0 for all r ,= 0 Now (f, 0) and (0, 1) generate Z/2Z wr Z.
General result: A wr B is nitely generated if both A and B are.
For Z wr Z =
Z
Z is not polycyclic because the rst factor is not nitely generated.
6. F
r
a free group of rank r. Then F
r
/F
(d)
r
is a solvable group, where F
(d)
r
is the dth term in the derived
series(check?).
3 Feb 05, 2013
3.1 Growth
Denition. G a nitely generated group with G A where [A[ < . The word metric on G wrt A is
d(g, h) is the length of the shortest word on AA
1
representing g
1
h. Minimal length words representing
a given group element are called geodesic words.
Example The Lamplighter group Z
2
wr Z
2
Below is an example of a group element. Here o denotes the origin (0, 0), denotes a lit light, and a
non-lit light. is the position of the lamplighter.

o

A generating set here is A = a, x, y where x and y move you left and up, respectively, and a lights the
lamp in that position. So the element above would be given by
yayxxax
1
The problem of identifying a geodesic representing a given group element here is NP-hard since it amounts
to travelling salesman problem on a grip.
Denition. Growth function f is
f(n) : = #(A A
1
1)
n
= #g G : d(1, g) n
Examples
1. Z wrt 1 then f(n) = 2n + 1
n
. .
n
0 n
. .
n
. .
2n+1
2. Z
2
wrt (1, 0), (0, 1) then f(n) n
2
3. F
2
wrt a, b notice f(0) = 1, f(1) = 5, f(2) = 17 in general f(n) = 2 3
n
1
The above example featured exponential growth. In fact this is the fastest that any group can grow:
Proposition. For all groups f(n) (2[A[ + 1)
n
4. Z
2
wrZ

n 1 0 1 n
. .
4n
Words of length 6n represent all group elements where the lamplighter is at the origin and the
illuminated lights are in n, , n. That is 2
2n+1
group elements. So
f(6n) 2
2n+1
f(n) 2
n
5. BS(1, 2) = Z[
1
2
]
2
Z
We have 2
m
b
m
ab
m
0
1 +
1
2
1
+ +
k
2
k
where
i
0, 1
k
i=0
(b
i
ab
i
)
i
and this has length k, so
f(n) 2
n
6. Sol = Z
2
A
Z where A =
_
2 1
1 1
_
Exponential growth
_
_
_
2 1
1 1
_
k
p
q
0 0 1
_
_
7. The Heisenburg Group 1
3
=
_
_
_
_
_
1 Z Z
1 1 Z
0 0 1
_
_
_
_
_
with generators
a =
_
_
1 1 0
1 1 0
0 0 1
_
_
b =
_
_
1 0 0
1 1 1
0 0 1
_
_
We have a
m
=
_
_
1 m 0
1 1 0
0 0 1
_
_
and b
n
=
_
_
1 0 0
1 1 n
0 0 1
_
_
and [a
m
, b
n
] =
_
_
1 0 mn
1 1 0
0 0 1
_
_
= c
mn
where
c = [a, b] =
_
_
1 0 1
1 1 0
0 0 1
_
_
So if w is a word of length at most k. It corresponds to a matrix of the form
_
_
1 p r
1 1 q
0 0 1
_
_
where [p[, [q[ k and [r[ k
2
. Therefore
f(n) n
4
Remark: If w is a word on a, b, c
1
then there exists a word w representing the same group element
of the form a
p
b
q
c
r
where if l(w) = n then [p[, [q[ n and [r[ n
2
. This follows from the (nice) normal
form of elements in 1
3
coming form the fact that
[a, b] = c, [a, c] = 1, [b, c] = 1
This says that c is central, and that that we can swap as and bs at the expense of adding a c.
Therefore we can bring our as to the front, and get some cs oating about, but because c is central,
we can just send them all to the back.
Theorem (Dixmier-Wolf-Bass-Guivarch)
For a nitely generated nilpotent group with lower central series
G =
1
(G)
2
(G)
n+1
(G) = 1
the growth f(n) n
d
where
d =
k torsion-free rank(
k
(G)/
k+1
(G))
Example 1
3
1
3
Z 1
So d = 1 2 + 2 1 = 4, in agreement with example 7 above.
Some Facts About Growth
For f, g : N N write f g when there exists c > 1 such that f(n) cg(cn + c) + cn + c. Write f g
when f g and g f. For example, n
for , 1 i = , but
n

n
for all , > 1.
If A and B are nitely generated sets for a group G then f
A
f
B
. In fact the functions are Lipschitz
equivalent i.e. there exists a c > 0 such that
f
A
(n/c) f
B
(n) f
A
(cn)
Lemma. If H is a nite index subgroup of a nitely generated group G then (H is also nitely generated
(!) and) their growth functions are equivalent.
Denition. A map : X Y between metric space is a (, )-quasi-isometry (with 1, 0)
when
1
d(a, b) d((a), (b)) d(a, b) +

and for all y Y there exists x X such that
d((x), y)
Examples
1. G = A = B with [A[, [B[ . Then Id : (G, d
A
) (G, d
B
) is a quasi-isometry
2. H G is a quasi-isometry
Lemma. Quasi-isometry is an equivalence relation. (Exersize)
Lemma. If G
1
and G
2
are nitely generated quasi-isometric groups (i.e. there exists a quasi-isometry
between them) then their respective growth functions are equivalent
f
G
1
f
G
2
Note. Proving this lemma in combination with example 2 above is sucient to prove the lemma above.
4 Feb 7, 2013
Recall,
Lemma. If G and H are nitely generated quasi-isometric groups (i.e. there exists a quasi-isometry
between them) then their respective growth functions are equivalent
f
G
f
H
Pf. : G H a (, )-quasi-isometry. Consider a ball of radius n about e G. The image of this ball in
H under has radius n + (since is a quasi-isometry).
There exists some c > 0 such that if x, y G have d(x, y) > c then (x) ,= (y). Now take a maxi-
mal subset of G such that no two points are c apart. Call this subset a mesh, M. Now [M[ [B(n, e)[.
But [M[ [B(n, e)[ [B(c, e)[[M[. Dene [B(c, e)[ = K, some constant. We have [(M)[ = [M[, by
construction of M. But [M[ f
H
(n +) Putting this together gives,
f
G
(n) = [B(n, e)[ K[M[ Kf
H
(n +)
Now since quasi-isometry is an equivalence relation there is an inverse quasi-isometry. Doing the same as
above will give the other direction. .
Denition. A group is virtually nilpotent when it has a nite index subgroup which is nilpotent.
Theorem. (Gromov) A nitely generated group G has polynomial growth i G is virtually nilpotent.
Corollary. Being virtually nilpotent is a geometric property among nitely generated groups (that is,
if G is quasi-isometric to a virtually nilpotent group H, then G is virtually nilpotent).
4.1 Idea behind Gromovs proof
Theorem. (Milnor-Wolf ) A nitely generated solvable group is of polynomial growth i it is virtually
nilpotent. It is of exponential growth otherwise.
Sketch of Proof of Gromovs Theorem.
Suppose G is an innite nitely generated group of polynomial growth. Let d be a word metric on G with
respect to a nite generating set. Let
Y := lim
n
(G,
1
n
d)
(This is a clever idea: Think of Z
n
lattice. The idea of Y is that as we move further away from this it starts
to look like 1
n
)
Properties of Y :
1. Homogeneous
2. Connected, locally connected
3. Complete
4. Locally compact, nite dimensional (topological dimension)
Gleason-Montgomery-Zippins Solution to Hilberts 5th Problem It can be shown that Isom(Y )
is a Lie group. Gromov uses this to build a homomorphism from a nite index subgroup of G onto Z. (This
is hard)
The kernel of this homomorphism is of polynomial growth of lower degree. Induction will allow us to
deduce that G has a solvable subgroup of nite index. Then Milnor-Wolf nishes the proof.
4.1.1 Background: -algebras and Probability Measures
Denition. Let X be a set, and T(X) = 2
X
its power set. Then a subset 2
X
is a -algebra if it
satises:
1. is non-empty
2. is closed under compliment
3. is closed under countable unions
It thus follows by de Morganss laws that it is also closed under nite intersection.
Proposition. X and are both contained in for any -algebra (of X).
Proof. By 1. is non-empty so there exists A X in . By 2. it is closed under taking compliments, so
XA . By 3. it is closed under countable unions, so A XA = X . And nally it is closed under
compliment, so XX = .
Proposition. Given a nite collection of subsets of X, S = X
i
, there exists a unique smallest -algebra
containing S.
Proof. Indeed, consider
:= E 2
X
: E is a -algebra of X containing S
First note that is non-empty, as 2
X
. Now take the intersection of every element in . Since each
element contains S, so too will the intersection. Moreover, since consists of -algebras, and since they are
closed under intersections, then this intersection will itself be a -algebra, and thus is the minimal element
in .
Denition. Denote this intersection (S) and call it the -algebra generated by S.
Example. Let X = 1, 2, 3. The -algebra generated by 1 is (1) = , 1, 2, 3, 1, 2, 3
Denition. Let X be a set and a -algebra over X. A function : 1 is called a measure
if it satises:
1. Non-negativity: (E) 0 for all E
2. Null empty set: () = 0
3. Countably additive: for all countable collections E
i
of pairwise disjoint sets in
_
_
i
E
i
_
=
i
((E
i
))
Denition. A probability measure is a measure with total measure 1, that is, (X) = 1.
4.1.2 Ultralters
Denition. A non-principle ultralter : T(N) 0, 1 is a nitely additive probability measure on
N taking values in 0, 1 and giving all singleton sets measure 0.
Remark. These exists by Zorns lemma: Consider the set of functions f : T(N) 0, 1 such that
f
1
(1) is closed under taking nite intersections and supersets, and doesnt include the empty set, but does
include all conite sets. Example: f
1
(1) = the conite sets. Now I want to take a maximal element using
Zorns lemma, under the relation f

f when f
1
(1)

f
1
(1). This maximal element is an example of a
non-principle ultralter.
Denition. Let : T(N) 0, 1 be a non-principle ultralter. Then a 1 is an -ultralimit of a
sequence (a
n
) 1 when
> 0 (i : [a a
i
[ < ) = 1
and say is an -ultralimit when
i : a
i
> N = 1 N > 0
and say is an -ultralimit when
i : a
i
< N = 1 N > 0
Lemma. Every sequence (a
n
) in 1 has a unique -ultralimit in 1 . Call this lim
a
n
.
Sketch. Assume (a
n
) is bounded, then within the bounded set of 1 (so an interval) there is a sequence of
points. Cut this interval in half. One interval will have measure 1, the other measure 0. On the half with
measure 1 repeat this process.
4.1.3 Construction of Y . (van den Dries-Wilke)
Denition. (X, d) a metric space, a non-principle ultralter, e = (e
n
) a sequence of basepoints in X,
s = (s
n
) a sequnce in 1
>0
with s
n
as n . Then an asymptotic cone is
Cone
(X, e, s) :=
_
(x
n
) X : lim
d(e
n
, x
n
)
s
n
<
_
/
where (x
n
) (y
n
) i lim
d(x
n
,y
n
)
s
n
= 0. The metric on Cone
(X, e, s) is
d((x
n
), (y
n
)) := lim
d(x
n
, y
n
)
s
n
(hence quotienting out by , as this identies sequences which would be distance 0 under the metric above).
This is lim
ia
(X,
d
s
i
), the limit of these metric spaces.
5 Feb 12, 2013
5.1 Properties of Asymptotic Cones
Proposition. Given , e in X, s, then a quasi-isometry : X Y induces a bi-Lipshitz homeomorphism
Cone
(X, e, s) Cons
(Y, (e), s)
where (e) = ((e
i
)).
Corollary. Topological invarients of cones give quasi-isometry invarients of spaces e.g. nitely generated
groups. (So for instance, cones being contractable, say, or simply connected, is a quasi-isometry invariant)
Proof of Prop. Suppose (a
i
) and (b
i
) represents points in Cone
(X, e, s). Then

1
d((a
i
), (b
i
)) d((a
i
), (b
i
)) d((a
i
), (b
i
))
(which is bi-Lipshitcz). Lets try and justify this:
d((a
i
), (b
i
)) = lim
d((a
i
), (b
i
))
s
i
lim
d(a
i
, b
i
) +
s
i
= lim
d(a
i
, b
i
)
s
i
+ lim
s
i
= d((a
i
), (b
i
))
since lim
s
i
= 0 as is nite and s
i
. The other inequality is similar.
Suppose (y
i
) representa a point in Cone
(Y, (e), s). As is a quasi-isometry (and in particular we are

including quadi-onto in that denition), for all i there exists x
i
X such that
d((x
i
), y
i
)
We claim that
1. (x
i
) represents a point in Cone
(X, e, x)
2. ((x
i
)) = (y
i
)
for (2.)
d(((x
i
)), (y
i
)) = lim
d((x
i
), y
i
)
s
i
lim
s
i
= 0
For (1.)
d((e
i
), (x
i
)) = lim
d(e
i
, x
i
)
s
i
lim
d((x
i
), (e
i
)) +
s
i
lim
d((x
i
), y
i
) +d(y
i
, (e
i
)) +
s
i
lim
+d(y
i
, (e
i
)) +
s
i
= d((y
i
), ((e
i
))) <
Proposition. For a nitely generated group , the asymptotic cone,

( := Cone
(, e, s)
is:
1. homogeneous
2. geodesic
3. complete
Proof.
1. Suppose (g
i
), (h
i
) represents points in (. Then
(h
i
g
1
i
) : ( (
(a
i
) (h
i
g
1
i
a
i
)
is an isometry taking (g
i
) to (h
i
).
2. If = A with [A[ < then Cay
A
() (its Cayley graph) is a (1,
1
2
)-quasi-isometry (you map
to the vertices of its Cayley graph, you miss the edges but, thats ne because we take edges to have
length 1). So ( = Cone
(, e, s)
= Cone
(Cay(), e, s). So I can just prove it for the latter space.

Claim. If X is a geodesic space, then Cone
(X, e, s) is a geodesic space.

Indeed, suppose (a
i
), (b
i
) represent points in Cone
(X, e, s). Then there exists a geodesic path
i
: [0, d(a
i
, b
i
)] Cone
(X, e, s) from a
i
to b
i
. Dene
: [0, d(a, b)] Cone
(X, e, s)
by (r) = (
i
(s
i
r)). For r [0, d(a, b)],
i
(s
i
r) is well dened for all i in some set of -measure 1.
3. So Cauchy sequences converge. Use an appropriate diagonal argument.
Remark. ( geodesic implies that it is both connected and locally connected. However, there are examples
where they are not locally simply connected, so that can look very ugly locally...
Theorem. (Drutu-Gromov) For a group with a word metric,
( = Cone
(, e, s)
is proper (closed balls are compact) for all and s i has polynomial growth.
Theorem. (Pansu) If is nitely generated and virtually nilpotent, then all its asymptotic cones are
isometric to the same graded nilpotent Lie group with Carnot-Cartheodory metric (i.e. its Lie algebra has
a decomposition
i
g
i
where [g
i
, g
j
] = g
i+j
for all i, j).
e.g. Real Heisenburg Group 1
3
(1) =
_
_
1 x y
1 z
1
_
_
. Lie algebra is 3-dimensional x, y, z, where the
only non-trivial bracket is [x, y] = z, so the grading is g
1
g
2
where g
1
= x, y, g
2
= z.
A Carnot-Cartheodory metric (here on a C-C metric) on a manifold is a metric in which the distance
between two points is the inmal length of all at paths between the points, that is, paths that stay
tangent to some specied chosen sub-bundle of the tangent bundle.
e.g. Thinking still of the Heisenburg group 1
3
the paths can only travel in the x and y directions (so
that is the sub-bundle).
6 Feb 14, 2013
Theorem. For a nitely generated nilpotent group, the topological dimension of Cone
(, e, s) is the
cohomological dimension of /tor()
Remark. Torsion ruins cohomological dimension, so kill that (in a nilpotent group torsion forms a normal
subgroup).
Here is how to think of cohomological dimension: think of the Cayley graph, and add cells in dimen-
sion 1, say, then 2 etc. until the space is contractable. The cohomological dimension is the dimension of
the cells at which this stops. For example, think of Z
n
, add 2 cells, 3 cells, etc. up to n-cells, when we get
1
n
which is contractable. Thus cohomological dimension of Z
n
is n.
e.g. n=3,
Here we think of adding 2-cells to the lattice Z
3
, but this is still not contractable, so add 3-cells to ll it
in, and we are left with 1
3
, which is contractable.
Corollary. If a group is quasi-isometric to Z
n
then it is virtually Z
n
.
Proof. is quasi-isometric to Z
n
is of polynomial growth is virtually nilpotent (by Gromov).
Let
0
be a nite-index subgroup of which is nilpotent. The topological dimension of the cones of a group
quasi-isomteric to Z
n
is n. So
0
/tor(
0
) has cd = n. But the growth of
0
is f(r) = r
n
, a polynomial of
degree n. So
0
/tor(
0
) is Z
n
.
(To see this last step recall the formula k torsionfreerank(
k
(G)/
k+1
(G)). We see that k has to
be one as we also have n = torsionfreerank(
k
(G)/
k+1
(G)).)
Denition. We say that Z
n
is quasi-isometry-rigid.
6.1 Filling Functions
Given a nite presentation x
1
, . . . , x
m
[r
1
, . . . , r
n
of a group .
Denition. The presentation 2-complex, K has:
1. 0-cells: 1
2. 1-cells: a directed edge for each generator
3. 2-cells: glue in corresponding to relations
Denition. The Cayley 2-complex is the universal cover of K.
1
(K) = by Seifert-van Kampen.
Can identify with

K
(0)
, and identify

K
(1)
with the Cayley graph of with respect to x
1
, . . . , x
m
.
Example. Z
2
= a, b[[a, b], then K is the rose with 2 petals, R
2
with a square with edges labelled
a, b, a
1
, b
1
glued to R
2
in such a way that the edges and their orientations match up. The result of this
is a torus. The universal cover of of the torus is 1
2
. We see that
1
(Torus) = Z
2
acts on 1
2
by translations
(freely).
Denition. A van Kampen diagram of = x
i
[R is a nite contractable planar 2-complex with
all the edges directed and labelled by some x
i
in such a way that around each 2-cell one reads the dening
relator.
Equivalently, a van Kampen diagram is a nite contractable planar 2-complex together with a combina-
torial map f :

K (meaning that it sends i-cells homeomorphically to i-cells of the Cayley 2-complex).
1. The area of is the number of 2-cells of , Area().
2. The diameter of is Diam() := maxd(p, q) : p, q in
(0)
, where distance here is the combinatorial
distance in
(1)
.
Denition. Let X be a topological space with basepoint e, let : (S
1
, ) (X, e), be rectiable (nite
length). The lling length of is
FL() := infL 0 : there is a nullhomotopy H
t
(s) of with l(H
t
) L
Below are two contrasting examples: in the plane, a nullhomotopy can always be performed via paths of
length strictly less then that of the original loop. However, on a sphere with a puncture the loops can have
to get much larger then the original loop during the homotopy e.g. see the red loop below:
3. The lling length of (, ), where is a choice of basepoint on , FL(, ), is the minimal L such
that there is a combinatorial nullhomotopy (
i
) of =
0
to =
r
, with max
i
l(
i
L).
A combinatorial nullhomotopy is thought of as reducing our planar 2-complex by obvious moves,
e.g. can cast of a free edge (think of this as contracting a free edge) or removing a 2-cell which
shares a boundary component with the boundary. Then we think of the length of the boundary of
the 2-complex at each stage in this process.
Denition. For M = Area, Diam or FL, dene M : N N by
M(n) = maxM() : an edge loop of length n in

K
where
M() = minM() : is a van Kampen diagram such that =
Example Z
2
Area(n) = n
2
, Diam(n) = n, FL(n) = n.
7 Feb 19, 2013
7.1 Combings of Groups
Denition. G a group with nite generating set X. A combing is a section : G (X X
1
) of the
natural surjection (X X
1
) G.
: G (X X
1
)
g
g
can be considered as a unit speed path in the Cayley graph Cay
X
(G) from e to g.
1. is synchronously k-fellow-travelling when g, h G : d
X
(g, h) = 1 t [0, )
d(
g
(t),
h
(t)) k
2. A reparametrisation is an unbounded function N N with (0) = 0 and (n+1) (n), (n)+1
for all n
3. is asynchronously k-fellow travelling when g, h G : d
X
(g, h) = 1 reparametrisations ,
such that
t N d(
g
((t)),
h
(
(t)))
4. The length function
L(n) = maxl(
g
) : g G, d(e, g) n
Examples. (Combable Groups)
1. Finite groups all have synchronously k-fellow-travelling combings with L(n) constant. (Just take k
to be the diameter).
2. The free group F
n
. The geodesics form a synchronously 1-fellow travelling combing with L(n) = n.
3. Z
n
= a
1
, . . . , a
n
: [a
i
, a
j
] = 1 i, j. Then
g
= a
r
1
1
a
r
2
2
a
r
n
n
4. Baumslag-Solitar group, BS(1, 2) = a, b : b
1
ab = a
2
b
r
ua
s
: r, s Z, u ab
1
, b
1
and the rst letter of u is not b

1
denes an asynchronously fellow-travelling combing. The re-write rules:

(a) ab ba
2
(b) a
1
b ba
2
(c) a
2
b
1
b
1
a
(d) a
1
b
1
ab
1
a
1
(e) aa
1
(f) a
1
a
(g) bb
1
(h) b
1
b
Using these rules one can convert any given word into a combing word. e.g.
ab
n
ba
2
b
n1
b(ab)a
2
b
n2
b
n
a
(2
n
)
Here L(n) 2
n
.
Denition. Synchronously automatic groups are nitely generated groups with synchronous k-fellow-
travelling combings for which the language
g
: g G is regular (low complexity).
For example: Z
n
with the rst combing is automatic (language can be realised by an automaton), but
if one uses the straight line combing then it is not automatic. Other examples of synchronously automatic
groups are:
1. Hyperbolic groups
2. Braid groups
3. Many 3-manifold groups
4. Almost no nilopotent groups are automatic: Exceptions: the only torsion free nilpotent groups which
are automatic are Z
n
.
Remark. The length functions of synchronously automatic groups satisfy L(n) Cn for a constant C.
Remark. CAT(0) groups admit synchronously k-fellow travelling combings with L(n) Cn for a constant
C.
Theorem. If a group G with nite generating set X has a combing that is asynchronously k-fellow-
travelling and has length function L(n), then G is nitely presentable and
Area(n) CnL(n)
Diam(n) FL(n) Cn
Proof. Suppose w (X X
1
)
represents 1 in G. Then w represents a loop in the Cayley graph.

Moreover there is a combing path between e and every group element in w:
Consider this as a planar graph and send this into the Cayley graph in a combinatorial way (which now
might not be planar). (Note: there is a subtle point here that the combing paths drawn above are really
just representations of the combing paths in the Cayley graph, and we draw them in such a way that they
are disjoint. When we map this into Cay
X
(G) they may intersect). Faces have perimeter at most 2k + 2.
Let R be the set of words in (XX
1
)
which represent 1 in Gand which have length 2k+2. Note [R[ < .
Claim 1: is a van Kampen diagram for w wrt X[R
Claim 2: X[R = G
Indeed claim 1 is true by construction. Will come back to Claim 2.
1. Area claim: There are l(w) icicles (i.e. combing paths) each with at most 2L(n) 2-cells. So in total
Area(n) 2nL(n).
2. Diameter claim: Will show that Diam(n) Cn. Take a general vertex. Travel orthogonal to the
icicles. There are at most k vertices between you and the icicle adjacent to you, and there are at most
n paths, hence the diameter is kn.
3. It is a general fact that Diam FL With that point 2 becomes redundant.
4. Filling length claim: null-homotope by contracting orthogonal to the icicles
Open Question. Do nitely generated class c nilpotent groups all admit (a)synchronous k-fellow-travelling
combings with L(n) Cn for a constant C. (And if its asynchronous can you do it with a regular language).
8 Feb 21, 2013
We can view a nite presentation = X[R as a short exact sequence,
1 R F(X) 1
so a word w on X X
1
represents 1 in if and only if w can be written as a product,
w =
N
i=1
u
1
i
r
i
i
u
i
()
with r
i
R,
i
= 1, u
i
words on X X
1
. (Note: () reduces to 1 in the quotient F(X)/R)
Lemma. (van Kampen) A word w on XX
1
represents 1 in if and only if w admits a van Kampen
diagram over X[R. Moreover, Area(w) is the minimum N amongst all expressions ().
Proof. Suppose is a van Kampen diagram for w
For the reverse implication, start with the lollipop diagram associated to the expression in (). Morally we
should perform the above operation in reverse.
We have, however, to be careful about maintaining planarity.
That said, if we are careful, then it works.
Lemma. If w = 1 if and only if it can be reduced to the empty word using the moves
1. free reduction
x(a
1
a)y xy
2. free expansion
xy x(aa
1
)y
3. applying relations
xay xby where a
1
b R
Such a sequence of moves is called a null-sequence.
Example. Z
n
= a
1
, . . . , a
n
[[a
i
, a
j
] i, j
The Dehn functions Area(n) n
2
and FL(n) n.
Lemma. w = 1
1. Area(w) is the minimal N such that there is a null-seuqnce for w using at most N applications of
relations.
2. FL(w) is the minimum of max
i
l(w
i
) such that there is a null-sequence
w = w
0
w
1
w
2
w
n
=
for w.
Remark. This is a computer science type of notion. Re-label (1.)=Time, (2.)=Space. That is, thinking
of performing the moves 1, 2, 3 to a word w in a non-deterministic way (so no prescribed rule for when to
apply a given word), the idea above is that counting 3. tells us how long the algorithm will take to get the
the empty word, and counting the lling length will tell us how many moves it will take. Hence time and
space.
Proposition. The following are equivalent:
1. Area(n) is recursive
2. There is an algorithm solving the word problem
3. a recursive function f(n) such that Area(n) f(n) for all n
8.1 Quasi-Isometry Invariants
Proposition. Area(n), Diam(n) and FL(n) are all quasi-isometry invariants (up to ) among nitely
generated groups.
Corollary 1. Among nitely generated groups being nitely presentable is a quasi-isometry invariant
Corollary 2. If P
1
and P
2
are nite presentations for the same group, then
1. Area
P
1
Area
P
2
2. Diam
P
1
Diam
P
2
3. FL
P
1
FL
P
2
Proofs. Suppose that
1
= X
1
[R
1
nite presentation, and let
2
be a group with nite generating set
X
2
. Suppose,
:
2

1
is a (, )-quasi-isometry. Suppose w a word on X
2
X
1
2
and represents 1
2
. Aim to show that there
exists a van Kampen diagram for w in
2
. The idea is to
1. Represent the word w in Cay
X
2
(
2
)
2. Use the quasi-isometry to push the loop from Cay
X
2
(
2
) to Cay
X
1
(
1
)
3.
1
has a nite presentation, so construct a van Kampen diagram in Cay
X
1
(
1
).
4. Use quasi-inverse

to pull this van Kampen diagram back to Cay
X
2
(
2
)
Given a van Kampen diagram for w we have a nite set of relations that reduce w to 1
2
. Thus we
have shown that
2
is nitely presented. This will prove Corollary 1. In fact it is just the same idea to
prove the Proposition, we just need to analyse the van Kampen diagram obtained for w in Cay
X
2
(
2
),
and compare it to that in Cay
X
1
(
1
).
Use to push the word w from Cay
X
2
(
2
) to Cay
X
1
(
1
). Here we have a nite presentation, and so
have a van Kampen diagram representing the word:
Now is a quasi-isometry, so it has a quasi inverse

such that d(x,
(x)) C, d(x,
(x)) C and

is
a quasi-isometry. Push this van Kampen diagram back to Cay
X
2
(
2
) with

:
Since

is only a quasi-inverse, the van Kampen diagram might get shifted about some. But the idea is
that it only gets shifted by a bounded (in particular, nite) amount. Indeed - consider a loop in the van
Kampen diagram in Cay
X
1
(
1
). This has a perimeter length bounded above by the largest length of a
dening relation in R
2
(which is nite, so this perimeter is bounded). Hence when we push diagram back
to Cay
X
2
(
2
), the perimeter length is bounded. The only remaining problem is complete the diagram (the
red bit) by adding cells (the light blue bits) to get a van Kampen diagram for w in Cay
X
2
(
2
). But the
fact that

is a quasi-isometry means that this added bit are bounded, hence we are done.
9 Feb 26, 2013

9.1 The Area Function of Nilpotent Groups
Theorem. (Gromov; Gersten-Holt-Riley) If G is a nitely generated nilpotent group of class c then
Area(n) n
c+1
and FL(n) n
Moreover, these bounds are realisable simultaneously.
Proof. (Will prove this for the group 1
3
as this captures all the ideas of the general proof)
1
3
= x, y, z[[x, y] = z, [x, z] = 1, [y, z] = 1
Suppose w = 1 1
3
. Let n = l(w). We will show that there is a null-sequence w
i
for w using n
3
dening relations and with max
i
l(w) n (hence simultaneously).
Dene compression words u(s) representing z
s
for s 0:
for 0 s n
2
1, u(s) = z
s
0
[x
n
, y
s
1
] where s = s
0
+s
1
n and s
0
, s
1
0, 1, . . . , n 1
u(n
2
) = [x
n
, y
n
]
u(A+Bn
2
) = u(A)u(n
2
)
B
for all integers A, B with 0 A n
2
1 and B > 0.
Note: l(u(A+Bn
2
)) K
0
n for some K
0
depending on B.
Lemma. Fix K
1
> 0. Then there exists K
2
> 0 such that for all 0 s K
1
n
2
, there is a concatenation
of sequences converting z
s
to u(s):
z
s
z
s
u(0) z
s1
u(1) z
s2
u(2) zu(s 1) u(s)
using K
2
n
3
relations.
Proof of Lemma Key idea:
z
n
[x
n
, y
k
] = z
n
x
n
y
k
x
n
y
k
= z
n
x
n
y
k
(y
1
y)x
n
y
k
(insert 1 = y
1
y)
= z
n
x
n
y
(k+1)
(xz
1
)
n
y
k
+ 1 ([x, y] = z n times)
= [x
n+1
, y
k+1
] (n(2n +k + 1) relations)
So the total cost is n
2
.
Remark. Finitely generated nilpotent groups are all nitely presentable, a fact which we will prove later.
Induction on class c : Base case c = 1 i.e. abelian group. (easy)
Induction step: Consider G/
c
(G) which is a nitely presentable nilpotent group of class c 1. Let X be
a nite generating set for G and X its image in G/
c
(G). Let X[R be a nite presentation. Let X
be a
nite set of length c commutators (entries in X) in G that generate
c
(G) (a fact that itself can be proved
by induction on c). Let R be a set of [R[ relations (words on X) each expressing elements of R as equal to
a word on X
. So,
X X
[R, X
central
is a nite presentation for G.
Suppose w is a word on X and w = 1 in G. Let w be its image as a word on X. w = 1 in G/
c
(G),
so there is a null-sequence,
w = w
0
w
k
=
with l(w
i
) n and the number of relations used n
c
(by induction hypothesis). We seek to lift this to a
null-sequence for w. Prove this for 1
3
, but the idea is the same in general.
Let G = 1
3
we have G = 1
3
/
2
(1
3
) = 1
3
/z = x, y[[x, y] = 1.
First
w (z
)
1
w
0
z
by carrying all the zs to the right end, and all the z

1
s to the left.
Each move w
i
w
i+1
that applies the relation [x, y] = 1 lifts to a move which applies [x, y] = z, so
introduces a z
1
. Then shue all zs to the right, and all z
1
s to the left. Compress the powers of z at
each end as the z
1
s arrive. Eventually we get
u(s)
1
w
k
..
=1
u(t)
for some s, t 0. In 1
3
this is
z
s
z
t
= z
ts
= 1 (2)
so s = t. The total cost of this procedure is:
n
2
- the initial step
n
2
n - lifting the w
i
null-sequence and carrying the zs to the ends
Kn
3
- compression at each end
Total: n
2
+n
2
n +Kn
3
10 Feb 29, 2013

Question. (Open) Do class c, nitely generated, nilpotent groups have (a)synchronous fellow-travelling
combings with L(n) n
c
Remark. This would imply the n
c+1
Dehn function and n lling length result we proved last lecture.
Question. (Open) Does 1
3
admit a synchronously fellow-travelling combing?
Remark. There is an alternative proof to the result from last lecture using asymptotic cones.
Asymptotic cones approach to n
c+1
: Suppose G is a nitely generated nilpotent group of class c,
Cone
(X, e, s)
are graded (homogeneous) nilpotent Lie groups. In this case loops can be lled with discs of area n.
The problem is how do we pull this result back to the group. There is a partial result of Papasoglu
Area(n) n
c+1+
, which is close.
10.1 Lower Bounds
These can be tricky. Here is a warm up problem: Prove that for Z Z
= a, b : [a, b] the Area(n) n

2
. A
guess might be that Area([a
n
, b
n
]) n
2
.
Gerstens Lemma If the Cayley 2-complex is contractable then a van Kampen diagram which is 1-1 on
the compliment of its 1-skeleton is of minimal area.
Mad Proof that Area([a
n
, b
n
]) n
2
Suppose W =
N
i=1
u
1
i
[a, b]
i
u
i
, (where
i
= 1, u
i
are words on a, b, a
1
, b
1
) equals [a
n
, b
n
] in F(a, b).
We want to show that N is at least n
2
. We have W =
N
i=1
c
i
in
1
3
= a, b, c[[a, b] = c, [a, c] = 1, [b, c] = 1
(since [a, b] = c and c is central, so the u
i
s cancel). But in 1
3
we have [a
n
, b
n
] = c
n
2
, and c has innite
order. Now W = [a
n
, b
n
], by hypothesis, and so equating gives,
W = [a
n
, b
n
] = c
n
2
=
N
i=1
c
i
= W
whence comparing the powers of c gives N n
2
. (Recall,
i
= 1, so we get the lower bound n
2
precisely
when these are all +1, but it will be bigger then n
2
is some of them are 1s)
Centralised Isoperimetric Functions

P = A[R a nite presentation of , 1 = R in F(A).
= F(A)/1
Suppose w = 1 in .
Denition. The centralised area Area
cent
(w) = minN[w (
N
i=1
u
1
i
r
i
i
u
i
)[1, F]
This is just, the number of times, up to sign, that each relator occurs in a product of conjugates freely equal
to w, ignoring conjugation involved. This is a calculation in
1/[1, F]
which is nitely generated abelian group, as it is generated by 1, say r
1
, . . . , r
k
. Write w = r
s
1
1
r
s
k
k
in
1/[1, F]. Then the centralised area, Area
cent
(w) = min
k
i=1
[s
i
[
Lemma. Suppose B = y
1
, . . . , y
m
generates the abelian group,
1/[1, F]
For a word w(A) = 1 in , let l
B
(w) be the minimal length of a word on B representing w[1, F]. Let
K = maxl
B
(r) : r 1, then
1. l
B
(W) K Area
cent
(w) K Area(w)
2. c > 0 such that if w[1, F] = y
m
[1, F], where y[1, F] has innite order in 1/[1, F] then
m c Area
cent
(w)
Proof.
1.
2. 1/[1, F] = T Z
k
for some nite T. Let B = T a basis for Z. Then m l
B
(y
m
) (by projecting
to the Z
k
factor). Then m l
B
(y
m
) = l
B
(w) K Area
cent
(w).
Denition. Let F = F(a, b). Write F

1
= F, F
i+1
= [F
i
, F]. Then F/F
c+1
is the free nilpotent group
of class c on two generators .
Theorem. F/F
c+1
has Dehn function n
c+1
(so n
c+1
)
Proof.
Let u = [
..
c
[[a, b], a], , a]
Lemma u F
c
1, and u
k
c
= [ [[a
k
, b
k
], a
k
], . . . , a
k
] mod F
c+1
This is just a more elaborate version of the same notion we showed in the example with 1
3
above. Not
going to prove it.
Let 1 = F
c+1
. Then [1, F] = F
c+k
and 1/[1, F] is F
c+1
/F
c+2
is generated by simple commutators. Now
write u
k
c
= [
..
c
[[a
k
, b
k
], a
k
], . . . , a
k
] mod F
c+1
= w
c,k
. Now w
c+1,k
F
c+1
and is a k
c+1
power of a simple
commutator (e.g. w
c+1,1
). So w
c+1,k
is a word representing 1 in F/F
c+1
.
Area(w
c+1,k
) Area
cent
(w
c+1,k
) constant l(w
k+1,c
)
c+1
In the Heisenburg group example this is
Area([a
k
, b
k
]) Area
cent
([a
k
, b
k
]) constant (Lk)
c+1
11 Mar 5, 2013
11.1 Some Open Problems Relating to Dehn Functions
Recall, nitely generated nilpotent class c groups have Dehn function (i.e. area) n
c+1
and FL(n) n.
Free nilpotent class c groups on 2 generators have Dehn function n
c+1
.
Theorem. (Gromov, Allcock, Olshanskii-Sapir) The (2k + 1)-dimensional Heisenberg group for
k 2 has Dehn function n
2
.
1
5
=
_
_
_
_
1 Z Z Z
1 0 Z
1 Z
1
_
_
_
_
1
7
=
_
_
_
_
_
_
1 Z Z Z Z
1 0 0 Z
1 0 Z
1 Z
1
_
_
_
_
_
_
of class 2.
Theorem. (R. Young) For all c there exists nilpotent class c group with Dehn function n
2
, n
3
, . . . , n
c+1
.
Theorem. (S. Wenger) There exists a nilpotent group with Dehn function ~ n
for any .
Sapirs Example a
1
, b
1
, . . . , a
s
, b
s
[ class 2, [a
1
, b
1
] [a
s
, b
s
] = 1. n
2
(n) Area(n) n
2
log n where
(n) as n .
Open Question.
1. Classify nitely generated nilpotent groups up to quasi-isometry
2. Classify them by Dehn function
3. Do there exist uniform upper bounds on the Dehn function and lling length functions on other classes
of groups, like:
(a) polycyclic groups (NB always nitely presented)
(b) nitely presentable metabelian groups
(c) nitely presentable linear groups
There is a problem:
Theorem. (Kharlampovich) There exist nitely presented 3-step solvable groups with undeciable word
problem.
11.2 HNN-Extensions
Denition. G a group, A, B subgroups and : A B an isomorphism. Then
G
:= G, t[t
1
at = (a) a A
Brittons Lemma Suppose w = g
0
t
e
1
g
1
t
e
2
t
e
n
g
n
where e
i
= 1, n 1, g
i
G, and w = 1 G. Then
it contains a subword of the form
1. t
1
gt for some g A, or
2. tgt
1
for some g B
This subword is called a pinch.
Proof. Draw a van Kampen diagram for w. Consider the cell attached to t
e
1
. This cell represents a
relation in G
, and so must be of the form t

1
at(a)
1
reading either clockwise or anticlockwise depending
on e
1
.
This leaves an open t-edge on the cell (possibly), and by the same argument, this must have a cell of the
same form attached. If we repeat this process we see that we form a corridor across the van Kampen
diagram ending in another t edge on the boundary, with elements in A along one side, and elements of B
along the other.
All ts along the boundary must pair up in this way. Consider now the outermost t-corridor. Suppose that
the t is oriented as shown in the diagram:
then we see that along the right side of the corridor we get w word in a. Now the green cell gives a relation
g
2
= a
1
a
2
a
3
a
4
. More generally we get that g
2
is a word in either A or B, depending on the orientation of t
(i.e. e
i
= 1), and this gives us the desired subword of one form or the other.
Proposition. (Normal Form) Every g G
can be represented uniquely as

w = g
0
t
e
1
g
1
t
e
2
t
e
n
g
n
where e
i
= 1, n 1, g
i
G, and such that w contains no pinches.
Proof. Existence: easy - just represent the word in that form and remove pinches.
Uniqueness: Suppose two such w
1
, w
2
both representing g. Then by Brittons lemma w
1
w
1
2
contains a
pinch as it represents the identity in G. If one of w
i
does not contain a t then nor does the other. If both
contain no t then the result is immediate. So suppose that at lease one of the w
i
s contains a t. Now w
1
w
1
2
contains a pinch and both contain a t.
They must pair the last t
1
with w
1
with the last t
1
in w
2
. By induction on the max of the number of t
1
in w
1
and w
2
, u and v are the same. It follows that w
1
and w
2
are the same.
11.3 Baumslag-Solitar Groups
These were introduced in 1963, as
BS(m, n) = a, t[t
1
a
m
t = a
n
= Z
This is an HNN-extension where : mZ nZ by a

m
a
n
, where a is a generator. Trivial examples that
were known prior to Baumsalg and Solitars paper are:
1. BS(1, 1) = Z Z
2. BS(1, 1) = Z Z
Denition. A group G is Hopan if it is not isomorphic to any one of its proper quotients i.e. every
epimorphsim G G is an isomorphism.
For example: Z
, consider innite sequences of integers. f : (a

1
, a
2
, a
3
, . . .) (a
2
, a
3
, a
4
. . . , ) is a surjective
homomorphism with innite kernel. Baumslag-Solitar groups were constructed to give an easy example of
non-Hopan groups.
Theorem. BS(2, 3) is non-Hopan.
Proof. We have
BS(2, 3) = a, t[t
1
a
2
t = a
3
Now a a
2
, t t denes an epimorphism G G with non-trivial kernel.
Well dened: t
1
a
2
t t
1
a
4
t = (t
1
a
2
t)
2
= (a
3
)
2
= a
6
so a
3
a
6
.
Surjective: a
1
t
1
at a
2
t
1
a
2
t = a
1
a
3
= a
Kernel: a
1
t
1
ata
1
t
1
ata
1
a
1
t
1
a
2
ta
2
t
1
a
2
ta
2
= a
2
a
3
a
2
a
3
a
2
= 1
Denition. G is residually nite when for all g G1 there exists a nite group H and a homomor-
phism : G H with (g) ,= 1.
Remark. Linear Residually Finite
Proposition. BS(m, n) is residually nite if and only if [m[ = [n[ or [m[ = 1 or [n[ = 1 (if and only if
linear)
Proof. BS(m, n) Z[
1
mn
] Z by a (1, 1), t (m/n, 0)
And Z[
1
mn
] Z

=
_
(m/n)
Z
Z[
1
mn
]
0 1
_
GL
2
(). Not injective when m ,= 1, n ,= 1.
Proposition. BS(m,n) is
1. Hopan if and only if it is residually nite, or m and n have the same prime divisors.
2. Solvable if and only if it is metabelian if and only if [m[ = 1 or [n[ = 1.
12 Mar 7, 2013
12.1 The Geometry of Baumslag-Solitar Groups
12.1.1 Normal Form for BS(1, n)
BS(1, n) = a, t[t
1
at = a
n
.
1. HNN-normal form: Use the HNN-normal form to give a normal form for BS(1, n),
t
i
a
k
t
j
: i, j 0, n [ k
This is not an (a)synchronous k-fellow travelling combing for n 2: For example in BS(1, 2),
t
j
a = t
j
at
j
t
j
= a
2
j
t
j
Now t
j
and a
2
j
t
j
are both in normal form, and dier by one generator. Consider a van Kampen
diagram: here each cell represents the relation t
1
at = a
2
, and we get
Now consider the paths for t
j
(in purple) and for t
j
a = a
2
j
t
j
(in green) dened by this
HNN-normal form:
Claim: d(p, ) j. This is only a claim because I have not got drawn the Cayley graph, so this
diagram only gives an idea of whats going on. We will see the Cayley graph later this lecture.
2. Asynchronous k-fellow travelling normal form: Use the rewrite rules
a
2
t
1
t
1
a
a
1
t
1
at
1
a
1
to convert it to an asynchronous fellow travelling combings. For example consider the words t
i
a
k
t
j
and t
i
a
k
t
j
a written in the new normal forms. Well t
i
a
k
t
j
a = t
i
a
k+2
j
t
j
by the rst rewrite rule.
Now compare their normal forms with the rewrite rules above (not these are in HNN normal form,
not our new normal form) (exersize) to see asynchronous combing of length 2
n
.
12.1.2 Dehn Function of BS(1, m)
Proposition. BS(1, m) has Dehn function 2
n
Proof. Prove this for BS(1, 2), but the idea is the same for m 2.
Upper bound: Area(n) nL(n) for any asynchronous combing, and we have just given an example where
L(n) 2
n
. (Alternatively, think of Brittons lemma: identifying a pinch in a word equal to the identity.)
Lower bound: Let w
n
= [a, t
n
at
n
]. Now l(w
n
) = 4n + 4. w
n
= 1 BS(1, 2) because t
nat
n
= a
2
n
.
Consider the van Kampen diagram for w
n
:
Consider the t-corridor starting as shown from the rightmost t (corridor starting in green). We know this
must end at a t on the boundary. Simple orientation considerations prevent the grey paths from being
valid corridors, leaving only the purple and the red as potential corridors. I claim it cannot be the red
corridor. Indeed if it were the red corridor shown, then the only option for any t-corridor contained within
this corridor (e.g. the blue t-corridor in the diagram) can only start and end in the red highlighted area.
But this is impossible as the orientation would agree. So the only possible choice for the red t-corridor is
to end up in the leftmost t, as shown below:
The blue cell here gives a dening relation in the a, namely that a = a
2
k for some k. This is not a valid
relation, showing that this is not a valid t-corridor. This leaves only the purple t-corridor as an option:
I claim that this t-corridor must end up in the rightmost position (as shown): indeed, all n copies of t we
have forming the top edge and the bottom edge must line up one-to-one, for if they didnt we would get
that one t-corridor had to loop back on itself (contradicting orientation considerations), or would have to
pass through another t-corridor (illegal). And so the black word highlighed is freely equal to a
r
for some r.
But reading around the diagram we get
a
r
= t
n
at
n
so r = 2
n
and so Area(w
n
) 2
n1
.
Corollary. BS(1, m) admits no combing with length function L(n) n, and in particular is not automatic.
Proof. Since automatic implies at most quadratic Dehn function.
12.1.3 Cayley Graph of BS(1, 2)

The Cayley graph is the 1-skeleton of the Cayley 2-complex, which is the universal cover of
the space obtained from gluing the cell representing the relation t
1
at = a
2
to the rose R
2
. The universal
cover of the resulting space is
Indeed, consider the bottom most row of as. We glue in the relation a
2
= t
1
at everywhere we can, we
start with:
Notice that if we start one further along the bottom row we can glue in another relation, and thus get
another strip:
We can now repeat this process along each line of as, at each point we get two folds emanating, and the
result is:
the Cayley diagram for BS(1, 2). Notice:
1. Let T be the innite binary tree. Then we get that the Cayley graph is T 1 - a Bass-Serre tree
(vertices correspond to cosets of a):
2. This is the universal cover of the aforementioned space: indeed it is simply connected as is takes the
form (Tree)1
12.1.4 Treebolic Space
Fix q > 0. Height function h : H 1 by z log
q
(Im z) (this gives our heights in red):
We have a tree T
p
= innite rooted tree with braching number p, e.g. the innite binary tree has p = 2.
We can embed T
p
in this picture according to the height function. Now the height function restricts to
h : T
p
1 (note this function depends on q):
Denition. Treebolic space is the horocyclic product HT(q, p) := (u, v) HT
p
: h(u) +h(v) = 0
(to be clear - this depends on q as it is used to dene the function h).
Example: Let v be the root of the tree, having height 1. Then we have that pairs (u, v) HT
p
: Im(u) =
p H(p, p), so we get that point v T
p
crossed with the line u H : h(u) = 1.
Similarly for all points v T
p
we get a line homeomorphic to 1 in HT(p, p)
Theorem. BS(1, p) acts properly discontinuously and cocompactly on HT(p, p).
Part II
Student Talks
13 Geometry of Sol (Chenxi Wu)
Sol = 1
2
1 where : 1 GL
2
1 by t
_
e
t
0
0 e
t
_
Discrete subgroup in Sol: Z
2
Z where : 1 A =
_
0
0
1
_
(Note: the mapping torus of T
2
glued with an Anasov map will also have a Sol structure)
13.1 Dehn Function
Denition. Area(l) = sup
length()=l
inf
D,=D
Area(D)
We get that Z
2
Z is quasi-isometric to Sol. Hence we get

Area
Z
2
Z
Area
Sol
Proposition. Area
Sol
exp
13.2 Quasi-Rigidity of Sol
is quasi-isometric to Sol then there exists K , [K[ < such that Z
2
A
Z
f.i
/K where here
A : Z Aut(Z
2
) by 1 A.
13.3 Area function becomes quadratic
Sol
3
=
_
_
e
t
x
e
t
y
1
_
_
= 1
2
1 = x, y t
Sol
5
=
_
_
_
_
e
t
1
x
e
t
2
y
e
t
3
z
1
_
_
_
_
= 1
3
1
2
= x, y, z t
1
t
2
, t
1
t
3

14 Geometry of the Lamplighter Group (Margarita Amchislavska)
Denition. The Lamplighter has numerous equivalent denitions:
1.
1
= Z[x, x
1
] Z, generator of Z factor acts on Z[x, x
1
] by multiplication by x
2.
1
=
iZ
Z Z = Z wr Z
3.
1
=
__
x
k
f
0 1
_
: k Z, f Z[x, x
1
_
4.
1
= a, t[[a, a
t
k
] = 1, k Z nitely generated, not nitely presented.
5. Picture: Lamps at integer locations, each lamp has an integer worth of possible brightness or values.
Lamplighter moves around adjusting the brightness (as we saw in the lectures)
14.1 Dead-end Depth
Let G be a group with generating set X. g G is a dead-end element if no geodesic ray from the identity
to g in the Cayley graph Cay(G, X) can be extended past g. That is, x X, d(1, gx) d(1, g). The
depth of a dead-end element g with d(1, g) = n is the length of the shortest path from g to an element
in B
c
n
(in other words, a dead-end element g means that the geodesic from 1 to g if extended by 1 (i.e.
multiplication by any other element x X results in moving back inside the ball of radius d(1, g). It still
could leave this ball if you extend far enough, and the minimum distance for which this happens is called
the dead-end depth).
Examples
1. G = Z, X = 2, 3. 1 has dead end depth 2.
2.
1,2
= a, t[a
2
= 1, [a, a
t
k
] = 1, k Z
Theorem (Clearly, Taback 2003)
1,2
has elements of unbounded dead-end depth.
d(1, g
n
) = 4n + 2n + 1 = 6n + 1, but d(g
n
, h) n + 1 h B
c
6n+1
since the lamplighter will
have to go to at least one lamp at n + 1 to get out of the ball.
14.2 Horocyclic Product of Two Z-branching Trees
There is a bijection between vertex set of innite Z-branching trees and cosets of Laurent polynomials in
some variable.
Remark
1. Each level partitions the polynomials.
2. Given a polynomial and its height there is a unique vertex corresponding to it.
3. Coecients of the polynomials give sequences of edge labels when following successive downward
edges.
Consider two such trees: one using Z[x] cosets (as exemplied above), the other using Z[x
1
] cosets.
Dene the height functions
1. H(f +x
k
Z[x]) = k
2. H(f +x
k
Z[x
1
]) = (k + 1)
The horocyclic product of these two trees T
1
, T
2
has vertex set (v
1
, v
2
) T
1
T
2
: H(v
1
) +H(v
2
) = 0 and
edges (v
1
, v
2
) (w
1
, w
2
) are connected if and only if v
1
w
1
and v
2
w
2
. Let 1
1
denote this horocyclic
product of the two trees with Z[x] and Z[x
1
]. Note: (g +bx
n1
+x
n
Z[x], g +dx
n
+x
n1
Z[x
1
]) V (1
1
).
We will show that 1
1
Cay(
1
, Y ).
Lets look at
1
again: g
1
is given by k Z and f =
f
i
x
i
in the matrix description say. OR:
1. k Z and a pair of sequences
(a) f
k1
, f
k2
, . . .
(b) f
k
, f
k+1
, . . .
OR:
2. A point (f +x
k
Z[x], f +x
k+1
Z[x
1
]
Another presentation of
1
is
1
=
i
, i Z[
m
i

m
j
=
m
j
m
i
, i, j, m Z
(check these are equivalent via
i
= a
i
t). In trees
i
corresponds to go down in the second tree T
2
(without
choice by f
k
), and go up in the rst tree T
1
(by f
k
+ i). This gives a graph isomorphism between the
Cayley graph Cay(
1
, Y ) and 1
1
(where Y is the
i
s).
15 Generalisations of the Lamplighter Group (Margarita Amchislavska)
2
= Z[x, x
1
, (1 +x)
1
] Z
2
=
__
x
k
(1 +x)
l
f
0 1
_
: k, l Z, f Z[x, x
1
, (1 +x)
1
]
_
= a, s, t[[a, a
t
] = 1, [s, t] = 1, a
s
= aa
t
2
was constructed by Baumslag in 72 as the rst example of a nitely presented group with an abelian
normal subgroup of innite rank.
15.1 Model
2 dimensional rhombic grid. Lamplighter travels on the lattice points. There are lamps on the integer
positions of the t-axis, and on the negative integer positions of the s-axis. The instructions are then: a ips
a switch, t says move one unit to the right, and s says move one unit up.
Elements in
2
correspond to integers on the grid points (only nitely many non-zero) and the position
of the lamplighter. Call a nite collection of integers on the lattice points a conguration. The identity
corresponds to zeros everywhere and the lamplighter at (0, 0). g
2
is given by (k, l, f) (k, l) is nal
position of lamplighter and f =
f
ij
x
i
(1 +x)
j
where f
ij
is the integer at position (i, j).
Two congurations are equivalent if they dier in the following way: if there is a triangle of integers a, b, c
with a sitting above b and c, then it can be replaced with a 1, b + 1, c + 1. If one follows this all the way
down to the t-axis we get a row from pascals triangle sitting on the lampstand (the t-axis together with the
negative 2-axis).
Fact For each conguration there exists a unique conguration equivalent to it supported only on the
lampstand.
15.2 Some Results
Get a basis for Z[x, x
1
, (1 +x)
1
] taking x
i
, (1 +x)
j
: i Z, j Z
<0
. One can use this to give a normal
form for
2
.
Let
2,r
be the group
2
with the additional relation that a
r
= 1.
Theorem (Cleary, Riley)
2,2
has elements of unbounded dead-end depth.
This was the rst example of a nitely presented group with this property.
Theorem (Grigorchuk, Linnell, Schick, Zuk) Gave
2,2
as a counter example to Atiyahs conjecture
(by constructing a 7-dimensional manifold with fundamental group
2,2
and the third L
2
-Betti number
1
3
not an integer).
Theorem (de Cornulier, Tessene)
2,2
has quadratic Dehn function.
Theorem (Kassabov, Riley)
2
has exponential Dehn function
Remark
1. Filling length function unknown.
2.
2
= , , c, d[[, ] = 1,
1
c
2
= c,
1
d
2
= d
15.3 Horocyclic Product
Claim Cay(
2
, Y ) is to the 1-skeleton of the horocyclic product of three innite Z-branching trees.
Use another presentation of
2
:
2
=
i
,
i
,
i
(i Z)[
i
=
i
mu
i
,
i+j
=
i
j
(i, j Z)
with
i
= a
i
t,
i
= a
i
s,
i
= a
i
ts
1
a
i
Recall
1. Vertices of an innite Z-branching trees correspond to cosets of Laurent polynomials in some variable
Z[x], Z[1 +x], Z[x
1
].
2. Given a Laurent polynomial and height, there exists a unique vertex corresponding to it.
3. Coecients of polynomials in the given variables give sequences of edge labels when following successive
downward edges.
Combinatorially Use the model. Given g
2
want to see how to get a vertex in 1
2
. Well, g
2

(k, l) and a conguration of integers on the grid.
1. Use the equivalence relation on triangles to propagate the conguration to the lampstand based at
(k, l).
2. Project portions to the axis: Split the lampstand at (k, l) into three portions, split horizontal portion
into two, left of (k, l) (not inclusively) and to the right of (k, l) (inclusively), and then the third
piece is the vertical portion. Now for the left portion project this down the the t-axis using the
triangle relation (rows of pascal triangle). Only consider the image on the t-axis to the left of k (not
inclusively). Similarly for the right portion project down and consider only piece to the right of k +l
(inclusively). Finally project the vertical portion to the s-axis and consider only the image up to l
(not inclusively).
3. Read of three heights and three sequences from these three projections: In the left projection h
1
=
k, a
1
, in the right most projection h
3
= (k +l), a
3
in the vertical protection h
2
= l, a
2
Claim this gives an element in the horocyclic product of three trees. This is actually a bijection.
16 Geometry of the Magnus Embedding (Andrew Sale)
What is a free solvable group? F
r
= x
1
, . . . , x
r
is the free group in the variety of r-generated groups.
Therefore any r-generated group is a quotient of F
r
. Similarly, Z
r
is the free group in the variety of r-
generated free abelian groups, so any r-generated free abelian group is a quotient of Z
r
. S
r,d
is going to be
the free group in the variety of r-generated solvable groups of derived length d, and therefore
any r-generated solvable group (of derived length d) is a quotient of S
r,d
. So,
S
r,d
= F
r
/F
(d)
r
where F
r
= F
1
r
= [F
r
, F
r
] and F
(i+1)
r
= [F
(i)
r
, F
(i)
r
].
16.1 Magnus Embedding
N F
r
. Let : F
r
F
r
/N, N
= [N, N]. The aim is to embed,

: F/N
Z
r
/ F
r
/N
If N = F
(d)
r
then S
r,d+1
Z
r
/ S
r,d
16.2 Applications of the Magnus Embedding
1. Various algorithm results for F/N e.g. (Kargapolov-Remeslennikov, 66) Conjugacy Problem is solv-
able in S
r,d
2. (Remeslennikov-Sokolov, 70) S
r,d
is conjugacy separable
3. Baumslags embedding theorem (that every nitely generated metabelian group embeds in a nitely
presented metabelian group) (see Amins talk)
4. Random walks on S
r,d
(see Tianyis talk)
Classical Denition We will realise Z/F/N as a matrix group. Let 1 be the free abelian Z[F/N]-module
generated by t
1
, . . . , t
r
.
M(F/N) =
__
g p
0 1
_
: g F/N, P = 1
_
16.3 Magnus Representation
: F M(F/N) by w
_
(w)

w
x
i
t
i
0 1
_
which is involving something called the Fox derivative.
(Magnus 30s) ker = N
the Magnus embedding.

16.4 Fox Calculus
A Fox derivative is a derivation
T : Z[F] Z[F]
where (recall) a derivation must satisfy, for a, b Z[F]
1. T(a +b) = T(a) +T(b)
2. T(ab) = T(a)(b) +aT(b)
where : Z[F] Z by sending g F 1.
F = x
1
, . . . , x
r
. Dene
x
i
(x
j
) =
ij
extend to F by (2), extend to Z[F] by (1). Heres a concrete example:
x
1
(x
1
x
2
x
1
1
+x
1
1
) =

x
1
(x
1
x
2
x
1
1
) +

x
1
(x
1
1
)
=

x
1
(x
1
x
2
)(x
1
1
) +x
1
x
2
x
1
(x
1
1
) x
1
1
Lemma. (Fox) Let T : Z[F] Z[F] by any derivation. Then K
i
Z[F] such that T =
r
i=1
K
i
x
i
Z[F]
D
$$
D
//
Z[F]
Z[F/N]
16.5 Wreath Products
A = Y , B = X, [X[, [Y [ < . Let = A/ B =
B
AB.
An element is a pair (f, b) where f : B A with nite support, and b B.
Multiplication: (f, b)(g, c) = (fg
(b)
, bc) where g
(b)
(x) = g(b
1
x)
16.5.1 Geometric Interpretation
Generating set for A/ B, S = (1, x), (f
y
, e
b
) : x X, y Y where f
y
(g) =
_
y g e
b
e
A
else
Let w be a word on S S
1
representing (f, b) A/ B. We can group together the letters of w which come
from X X
1
and those from Y Y
1
.
w = (f
0
, e
B
)(1, u
1
)(f
1
, e
B
) (f
m
, e
B
). Cay(B, X):
Supp(f
i
) = e
B
(except possibly for for i = 0, m where it may be empty).
Notice (f, b) = w = (f
0
f
(u
1
)
1
f
(u
1
u
2
)
2
, u
1
u
2
) Supp(f) e
B
, u
1
, u
1
u
2
, . . .
So a word w for (f, b) describes a path in Cay(B, X) from e
B
to b via Supp(f). Notice, this is just a
generalisation of the lamplighter model. Let K(Supp(f), b) be the length of the shortest such path, then:
Lemma. [(f, b)[
S
= K(Supp(f, )b) +
xB
[f(x)[
A
Lemma. M(F/N)
= Z
r
/ F/N
Proof. Take
_
b p
0 1
_
(f
P
, b)
where f
P
: F/N Z
r
, P =
r
i=1
(
gF/N
a
(i)
g
)t
i
with a
(i)
g
Z, then f
P
(g) = (a
(1)
g
, . . . , a
(r)
g
)
16.6 Geometric Denition of the Magnus Embedding

Dene
geo
: F Z
r
/ F/N by w (f
w
, (w)) (where : F F/N is the quotient map). What is f
w
?
The word w will describe a path
w
in Cay(F/N, (X)):
Dente a function
w
on the edges by setting
w
(g, gx
i
) to count the net number of times
w
crosses (g, gx
i
)
i.e. count +1 when
w
goes from g to gx
i
and count 1 when it goes from gx
i
to g:
f
w
: F/N Z
r
g (
w
(g, gx
1
), . . . ,
w
(g, gx
r
))
The function
w
is a ow on Cay(F/N, (X)) i.e. it determines a family of paths from e to (w).
(Drons-Lewin-Servatis 95)
w
=
u
if and only if uN
= wN
This result implies that ker

geo
= N
.
So
geo
: F/N Z
r
/ F/N is an embedding.
Proposition.
geo
= , that is, the two embeddings are equivalent.
Proof. NTS f
w
= f
P
. P =
r
i=1
w
x
i
t
i
. That is, NTS
w
x
i
=
gF/N
w
(g, gx
i
)g
Induction on l(w). If w = x
i
therefore
w
(g, gx
j
) =
_
1 g = e, i = j
0 else
and

w
x
j
=
ij
.
The inductive step considers w = w
x
i
:
Hence
w
(g, gx
j
) =
_

w
(g, gx
i
) + 1 g = (w), i = j
w
(g, gx
j
) else
, and
x
j
(w
x
i
) =
w
(g, gx
j
)g +w
ij

17 Finitely Presented Metabelian Groups (Amin Saied)
Here is an outline of the talk:
1. Introduction: Higmans Embedding Theorem, Metabelian groups, Baumslag-Remeslennikovs Theo-
rem, Baumslags sketch of proof
2. Baumslag/Margaritas Example: embedding
1
into
2
using an HNN-extension
3. Sketch of Proof:
(a) Magnus Embedding Theorem: reduce proof to case where G = H A where A, H are abelian.
(b) Special Polynomial Lemma: construct injective endomorphisms of A - will be used to construct
HNN-extensions
(c) Construct ascending HNN-extensions of G = H A where H = h
1
h
n
culminating in
G = QA
(d) Show G is metabelian
(e) Show G is nitely presented
17.1 Introduction
Motivating Question There are uncountably many nitely generated groups, but there are only count-
ably many nitely presented ones. Which countably many nitely generated groups then appear as sub-
groups of nitely presented ones?
Higmans Embedding Theorem (60s) A nitely generated group is recursively presentable if and
only if it is isomorphic to a subgroup of a nitely presented group.
Recursively Presentable: G = a
1
, . . . , a
n
. .
nite
[ w
1
, w
2
, w
3
, w
4
, w
5
. . .
. .
recursively enumerable
where a set is recursively enumerable if there exists an algorithm (a Turing machine) that can produce
a complete list of its elements.
Remark.
1. Higmans embedding theorem provides a complete answer to the motivating question. Its proof how-
ever would take us too far aeld from the theory of solvable groups.
2. There is, somewhat astonishingly, a related theorem for solvable groups, or more specically, for
metabelian groups.
Denition. A group G is metabelian if it has derived length 2, that is it has an abelian series,
1 H G
The following theorem was proved independently by Baumslag and Remeslennikov,
Baumslag-Remeslennikovs Thorem (73) Every nitely generated metabelian group embeds into a
nitely presented metabelian group.
Remark. In a paper titled Finitely Presented Metabelian Groups by Baumslag he sketches a proof of
this theorem in three steps:
(Andrew) Reduce the proof with the Magnus embedding: embed G into W/N where W = A / H where A, H
abelian
(Margarita) Show that W is embedded into a nitely presented metabelian group in the case
W = Z / Z =
1
(Amin) Complete the proof...
I am going to use Margaritas step as an example both of the theorem and of how the proof works. Then
I will use Andrews step to reduce the problem, demonstrate Margaritas step in full generality (not just
when A
= H

= Z, as Baumslag did) and sketch out the remainder of the proof.
17.2 Baumslag/Margaritas Example: Embedding
1
= Z / Z into
2
In Margaritas rst talk we saw the following presentation for the lamplighter group,
1
= a, s[[a, a
s
k
] = 1, k = 1, 2, . . .
This is clearly an innitely presented group. I claim it is metabelian:
Claim.
1
is metabelian
Proof. Surjective homomorphism
:
1
Z
a 1
s s
Then ker()
= Z
is abelian. Thus we have the following abelian series for

1
,
1 ker
1
So
1
is an innitely presented metabelian group. Baumslag-Remeslennikovs theorem says that we should
be able to embed this into a nitely presented metabelian group.
Construct G
1. Dene an injective endomorphism of
1
, :
1

1
by (a) = aa
s
, (s) = s.
2. Form HNN extensions of
1
:
G = t,
1
[g
t
= (g), g
1
Then G has the presentation,

G = a, s, t[a
t
= aa
s
, s
t
= s, [a, a
s
k
] = 1, k = 1, 2, . . .
Proposition. G is metabelian
Proof. Claim: G
= a
s
i
: i Z
() aa
s
= a
t
a
s
= a
1
t
1
at = [a, t] G
. Simple induction to complete (e.g. s

1
a
s
s = s
1
(a
1
t
1
at)s =
(as)
1
t
1
(as)t since [s, t] = 1)
() Basic fact: G/H abelain G
H, since G/G
is the largest abelain quotient of G.

Consider G modulo a
s
i
: i Z. Get
G/a
s
i
: i Z = s, t[[s, t] = 1 = Z
2
Proving the claim.
Now G/G
is abelian (by proceeding proof), and G
is free abelian of innite rank. Therefore

1 G
G
is an abelian series for G of derived length 2, that is, G is metabelian.
Proposition. G is nitely presented. In fact, G =

2
= a, s, t[a
t
= aa
s
, [a, a
s
] = 1 = [s, t]
Proof. The relations [a, a
s
i
] = 1 are redundant for i > 1. Indeed, suppose [a, a
s
j
] = 1 for j = 2, . . . , i
follows from [a, a
s
] = 1. Then
1 = [a, a
s
i
]
t
= [a
t
, (a
s
i
)
t
] = [a
t
, (a
t
)
s
i
] = [aa
s
, (aa
s
)
s
i
] = [aa
s
, a
s
i
a
s
i+1
] = [a, a
s
i+1
]
Where the equalities follow (respectively) from:
1. By hypothesis that [a, a
s
i
] = 1
2. Conjugation is a homomorphism, so [x, y]
t
= [x
t
, y
t
]
3. [s, t] = 1 commute
4. a
t
= aa
s
5. Again conjugation is a homomorphism
6. By the inductive hypothesis. (Exersize)
[aa
s
, a
s
i
a
s
i+1
] = aa
s
a
s
i
a
s
i+1
(a
s
)
1
a
1
(a
s
i
)
1
(a
s
i+1
)
1
The underlined terms give [a, a
s
i+1
]. Use the inductive hypothesis to show that the other stu com-
mutes and cancels.
Remark. So we have embedded a nitely generated, innitely presented metabelian group
1
into a
nitely presented metabelian group
2
by constructing an HNN-extension and showing that it was nitely
presented. This is the underlying idea behind the proof Baumslag-Remeslennikov theorem. Perhaps knowing
now, as we do, that
1
is the horocyclic product of two trees, and
2
is the horocyclic product of three
trees, this embedding is not so surprising. It is then perhaps more surprising that this works in general.
17.3 Proof of Baumslag-Remeslennikov Theorem
Magnus Embedding Theorem F a free group on x
i
[ I, R F. Given an isomorphism F/R H
by x
i
R h
i
. Let A be the free abelian group on a
i
[i I. Then the assignment x
i
R
h
i
a
i
determines
an embedding of F/R
into the wreath product A/ H.

Lemma 1. G a nitely generated metabelian group. Then G can be embedded into G
ab
A where A is
a nitely generated ZG
ab
-module.
Proof. G = g
1
, . . . , g
n
, F = Fx
1
, . . . , x
n
, : F G, K := ker
G metabelian F
K (for f
, (f
) G
= 1)
R :=
1
(G
) R = F
K
R
K (indeed, f F
, k K, [f, f] = [f, k] = [k, k] = 1)

Let A
0
be the free abelian group on a
1
, . . . , a
n
. We can apply the Magnus embedding with H = G
ab
since
F/R
= G/G
= G
ab
,
: F/R
A
0
/ G
ab
= A
(G
ab
)
0
G
ab
= W B := A
(G
ab
)
0
base group
x
i
R

_
x
i
R
0 1
_
where B, the base group.
Claim 1 W is nitely generated metabelian group
B and G
ab
both abelian.
Dene N = (K/R
)
Claim 2 N B
(wR
)
_
wR
0 1
_
. But R = F
K contains K. So if wR
N then w K R, thus upper left entry

is trivial:
_
1R
0 1
_

=
B
Claim 3 N Im()
Third isomorphism theorem: K/R
F/R
. Hence N = (K/R
) (F/R
) = Im().
Claim 4 N W
Im() contains all x
i
R. W is generated by B (abelian) and by (x
i
R).
Finally, by the third isomorphism theorem again,
G
= (F/R
)/(K/R
)
so induces an embedding of G into W/N = G
ab
(B/N) where B/N is a nitely generated ZG
ab
-module.
Remark The key point here is that we embedded G into a nitely generated metabelian group of the
form HA where H, A were abelian. So if we can prove Baumslag-Remeslennikov theorem in that case we
are done.
The idea of the proof is to embed G into ascending HNN-extensions and show that eventually we are left
with a nitely presented metabelian group. We need a way to construct these HNN-extensions, namely, we
need an endomorphism:
Lemma 2. H a nitely generated abelian group, A a nitely generated ZH-module. For each h H
there exists a polynomial
p = 1 +c
1
x + +c
r1
x
r1
+x
r
Z[x]
such that a ap(h) is an injective ZH-endomorphism of A.
Proof. A is a ZH-module, so multiplication by p(h) is clearly an endomorphism. Need to show it is
injective. Call polynomials of the form above special polynomials. Dene
A
0
= a A : ap(h) = 0 some special p
Then A
0
is a ZH-submodule (indeed multiplication of two special polynomials is again special).
Fact: G virtually polycyclic then R = ZG is a Noetherian R-module.
Since H is abelian, and since A is nitely generated, then A
0
is nitely generated. Say it is generated by
b
1
, . . . , b
s
Then there exist special polynomials p
i
such that b
i
p
i
(h) = 0. Dene
p = xp
1
p
s
+ 1
This is clearly a special polynomial. Suppose ap(h) = 0 some a A. Then a A
0
. Therefore
a = b
1
f
1
+ +b
s
f
s
some f
i
ZH. Notice that
b
i
p(h) = b
i
(hp
1
(h) p
s
(h) + 1) = hp
1
(h) b
i
p
i
(h)
. .
=0
p
s
(h) +b
i
= b
i
Therefore
0 = ap(h) = (b
1
f
1
+ +b
s
f
s
)p(h) = b
1
f
1
+ +b
s
f
s
= a
Thus a = 0 and we see that the map is injective.
Proof of Baumslag-Remeslennikov Theorem

Assume G = H A where A, H abelian, is a nitely generated metabelian group. If H were nite then G
would be polycyclic and therefore nitely presented. So assume H innite,
H = h
1
h
r
h
n
where h
1
, . . . , h
r
have innite order, and h
q
i
i
= 0, i = r+1, . . . , n. By lemma 2 there exist special polynomials,
p
1
, . . . , p
r
such that a ap
i
(h
i
) determines an injective ZH-endomorphism of A, say
i
.
Constructing the HNN-extensions
1. G
0
= G = H A
Extend
1
to an injective endomorphism of G
0
acting as the identity on H (H abelian). Dene,
G
1
= t, G
0
[g
t
1
0
= g
1
0
, g
0
G
0
2. Extend
2
to G
1
by requiring it to act as identity on the abelian subgroup H, t
1
. Dene,
G
2
= t
2
, G
1
[g
t
2
1
= g
2
1
, g
1
G
1
3. Repeat this r times resulting in G

r
.
Claim 1 G
r
= G = QA, where Q = H t
1
t
r
and A = A
t
1
,...,t
r
is the normal closure of A

in A, t
1
, . . . , t
r
.
Pf: Starting with G
0
= HA. Adding t
i
s and forcing them to commute with H and with a dened action
of t
i
on A. So we expect a semi-direct product of this form. The slight question is, why A. To see this
consider conjugation by a negative power of t
i
. For example, if g = (h, a) G
1
t
1
gt
1
1
=
1
1
(g)
i
are not necessarily surjective. Hence we take the normal closure. This is easier to visualise with the
simpler example BS(1, 2) = a, t[a
t
= a
2
.
Claim 2 G embeds in G and G is metabelian.

Pf: G is generated by the elements
h
1
, . . . , h
n
t
1
, . . . , t
r
a
1
, . . . , a
m
. .
generators for A
Similarly to before, 1 QG is an abelian normal series for G because both Q and A are abelian.
Constructing G
What relations do we have?

1. [h
i
, h
j
] = [t
i
, t
j
] = [h
i
, t
j
] = [a
i
, a
j
] = 1
2. h
q
i
i
= 1 for i = r + 1, . . . , n
3. a
t
j
i
= a
i
p
j
(h
j
) for i = 1, . . . , m, j = 1, . . . , r
Fact: ZH Noetherian A (a ZH-module) is nitely presented as nitely generated is equivalent to nitely
presented for modules over Noetherian rings. Put these relations in,
4. a
r
i1
1
a
r
i2
2
a
r
im
m
= 1 for i = 1, . . . , k with r
ij
ZH
Finally need relations to ensure the normal closure of a
1
, . . . , a
m
in G is abelian,
5. [a
i
, a
j
] = 1 for , of the form h
u
1
1
h
u
n
n
where 0 u
i
d
i
for d
i
the degree of p
i
, when 1 i r,
and 0 u
i
< q
i
for r + 1 i n
Dene
G
= h
1
, . . . , h
n
, t
1
, . . . , t
r
, a
1
, . . . , a
m
[1, 2, 3, 4, 5
The Conclusion There is a surjective homomorphism
G
G
Hall (54) Finitely generated metabelian groups satisfy max-n, the maximum condition on normal sub-
groups. That is, every normal subgroup is nitely generated.
1. Prove G
metabelian
2. Then G
satises max-n
3. Hence G is nitely presented: Indeed, G = G
/N some normal subgroup N G
. Now since G
satises max-n N is nitely generated. Therefore G is too.

Lemma. G
is metabelian.
Sketch of Proof:
1. If we show that A
= a
1
, . . . , a
m
is abelian, then
G
= h
1
, . . . , h
n
, t
1
, . . . , t
r
, a
1
, . . . , a
m
[1, 2, 3, 4, 5
G
/A
= h
1
, . . . , h
n
, t
1
, . . . , t
r
[[h
i
, h
j
] = [h
i
, t
j
] = [t
i
, t
j
] = 1 is abelain
hence 1 A
is an abelian series of derived length 2 and therefore G
is metabelian.
2. To prove A
is abelian one uses the special polynomials p

i
. It involves bashing out even larger
commutators then in the example of embedding
1
into
2
. The idea is similar enough.

18 Sigma Invarients (Teddy Einstein)
18.1 -invariants
Denition. A character of a group G is a homomorphism : G 1. Observe
G
//
1
G
ab
==
where G
ab
= Z
k
torsion. Hom(G, 1)
= 1
k
Denition. (Character Sphere) , characters on G. = c some c > 0. [] is the
equivalence class of . The character sphere is,
S(G) = [] : ,= 0
Remark. Thinking of this equivalence relation as identifying rays in the space of characters provides
motivation for the name character sphere.
Denition. For each : G 1 dene
G
= g G : (g) 0
Clearly if then G
= G
.
Denition. (
0
-invariant) Let G be a nitely generated group, and let A be nitely generated over
ZG. Dene
0
(G; A) = [] : A f.g. as a ZG
-module
Let M = G
/G
. We consider this as a module over ZG where scalar multiplication is by conjugation:

g G, m = [g
1
, g
2
]G
M then
g m = m
g
= ([g
1
, g
2
]G
)
g
= [g
1
, g
2
]
g
G
= [g
g
1
, g
g
2
]G
M
Theorem. M = G
/G
a module over ZG by conjugation. If G is nitely presented and solvable then
0
(G/G
; M)
0
(G/G
; M) = S(G)
(where we are taking the union with the antipode on the sphere).
Remark. The converse holds if G is metabelian.
We can generalise this,
G
= [] S(G) : G
f.g over a f.g. submonoid of G
Say G is nitely generated by X. (G, X) the Caley graph wrt X.

Denition. (
1
-invariant) Let
be the subgraph of (G, X) with vertices g such that (g) 0, and

with edges: if g, h
connect an edge if there was one in (G, X).
1
(G) = [] :
connected
Examples.
1. Z = t. The character is completely determined by the choice (t). If we choose (t) = 1 then []
represents all characters where (t) > 0. Then [] represents all characters where (t) < 0.
S(Z) = [] where (t) = 1.
2. Z
2
a, b[[a, b] = 1. Character is determined by (a), (b). Eg (a) = 1 = (b). So vanishes on
y = x.
is connected. Here,
S(Z
2
) =
1
(Z
2
)
3. G = BS(1, 2) = a, u[a
u
= a
2
= Z[
1
2
] Z where ((x, 2
m
)(y, 2
m
) = (x + 2
m
y, 2
m+m
), u = (0, 2), a =
(1, 2
0
))
Consider the Cayley graph drawn on a plane. We build it inductively: We need to represent the
relation u
1
au = a
2
. Fill out the plane with this relation starting at a chosen point, say, the origin.
Now, if we chose a dierent we still need to see this relation, so add it in. Think of this as projecting
dierent leaves into the same plane. At height h we project 2
h
leaves into the plane.
The character is determined by (a) and (u). But,
(a) = (u
1
au) = (a
2
) = 2(a)
(a) = 0
So we get that the character is determined by (u), and we have the same as in example 1,
S(G) = [
0
] where
0
(u) = 1
Claim. [
0
] / S(G) but [
0
] is.
(a)
0
not connected:
_
1
2
+z, 2
m
_
m 0
check that
_
1
2
+z, 2
m
_
_
0, 2
1
_
=
_
1
2
+z, 2
m1
_
check
_
1
2
+z, 2
m
_
_
1, 2
0
_
=
_
1
2
+z 2
m
, 2
m
_
m 0 z 2
m
Z. First coordinate is an integer. No path from (
1
2
, 2
m
) to the identity in
0
.
0
is not connected.
(b)
0
is connected in Z[
1
2
]: Arbitrary point (t, 2
m
) where m 0. There exists n N, z Z
such that 2
n
z = t. Point (2
n
z, 2
m
). Walk to (2
n
z, 2
n
) by a z times,
(2
n
z, z
n
)(1, 2
0
) = (2
n
(z 1), 2
n
)
Get to (0, 2
n
). Walk by u n-times to get to the identity. Hence
0
is connected, therefore
[
0
]
(G).
Theorem.
1
is invariant under choice of generators.
Proof. X
1
, X
2
nite generating sets. WLOG X
2
= X
1
z. Write = (G, X
1
) and
= (G, X
2
). If
is connected then
is a subgraph of
, hence
is connected.
Suppose
is connected.
Denition. p a path. Let v
(p) = min(g) : g vertex on p

So v
0 p lies on
. Let z = w
1
w
n
be a word over X
1
. There exists t X
1
with (t) > 0. Want
k such that (t
k
) = k. (t) [v
(w
1
, . . . , w
n
)[. Take g, h
. Then there exists a path p
from t
k
gt
k
to
t
k
ht
k
in
.
Take
p
= x
1
x
2
x
m
x
i
X
2
to
q = y
1
y
2
y
l
by replacing all z
1
by (w
1
w
2
w
n
)
1
. Then v
(g) [v
(w
1
w
2
w
n
)[. Shift a by t
k
. Path t
k
q from gt
k
to ht
k
gives v
(t
k
q) = v
(q) +k(t) 0.

18.2 Results
Theorem. G a f.g. group with no non-abelian free subgroups. G nitely presented then
(G)
(G) = S(G)
Example. [
0
] BS(1, 2) S(BS(1, 2))
Corollary. If G is nitely presented as above, and G/G
has rank 2 over Z, then there exists a normal

subgroup N G such that G/N is cyclic and N = ker for some character .
Theorem. is rank 1, t G with (t) = 1. Then []
1
(G) if and only if H N = ker such that
t
1
Ht H
and
N =
_
lN
t
l
Ht
(Here rank of is the rank of (G) in 1 over Z)
19 Sigma Invarients (Yash Lodha)
There are two kinds of sigma invarients:
1. Homotopical: input the group G (usually nitely generated)
2. Homological: input (G, A) for a group G and a ZG-module A
Consider the vector space Hom(G, 1) consisting of characters : G 1. This is a real vector space
isomorphic to 1
k
where k is the torsion free rank of G. If X is a path connected topological space we have
1
(X) = G, Hom(
1
(X), 1)
= Hom(H
1
(X), 1)
= H
1
(X, 1).
Recall from last time,
G
= g G : (g) 0
and recall that G
= G
c
for all c > 0. So this set of mono ids is parametrised by the set of rays in
Hom(G, 1). Identify this with the unit sphere, call it S(G).
Now consider the pair (G, A) where G is a nitely generated group and A is a ZG-module. Then
A
= [v] S(G) : A nitely generated as a ZGmodule
Lets try and clarify this somewhat opaque denition. Let C(A) = y ZG : a A, g a = a be the
centraliser (suppose action of g is by conjugation). Let I(A) = C(A) 1 = g 1 : g C(A). Now for all
g I(A) and for all a A, g a = 0.
Proposition. [v]
A
if and only if there is a d C(A) such that v(d) > 0.
Proof. Assume there exists d C(A) such that v(d) > 0. For all g G there exists m > 0 such that
v(g) < v(d
m
) = mv(d). Hence gd
m
a = g a and v(gd
m
) > 0 gd
m
ZG
v
.
If a
1
, . . . , a
k
is a nite generating set for A as a ZG-module.
i
g
i
a
i
=
i
g
i
d
m
i
a
i
for g
i
d
m
i
G
v
.
Conversely, assume [v]
A
. Let a
1
, , a
k
be a generating set for A as a ZG
v
-module. Take d ZG
v
such that v(d) > 0. Say
d
1
a
i
=
i
i,j
a
i
Hence
j
(
ij
d
i,j
)a
j
= 0
Therefore det(
ij
d
ij
) I(A). But
ij
d
i,j
= 1 d for ZG
v
. Therefore d C(A). v(d) =
v(d) +v() > 0.
Denition.
A
=
C(A)
[v] S(G) : v(d) > 0
Proposition.
1.
A
is an open subset of S(G)
2.
A
=
ZG/I
where I = I(A).
3. A
A A
of ZG-modules then
A
=
A

A

Proof.
1. d : S(G) 1 by d(v) = v(d). So 1 follows since the sets in the denition are open sets.
2. ZG/I as a ZG-module has annihilation I = I(A). Therefore C(A) = C(ZG/I), and
A
=
ZG/I
.
3.
A

A
follows since if A is nitely generated as a ZG
v
-module then so is A
. Similarly,
A

A

A
, since if A
, A
nitely generated wrt ZG

v
then so is A. We need
A

A
. I(A) I(A
) therefore
C(A) C(A
). So this follows by denition.

19.1 Finiteness Properties
If A is a ZG-module, a projective resolution for A is an exact sequence
P
i
P
i1
P
0
A 0
where each P
i
is a projective ZG-module.
We are interested in the case A = Z. Examples:
1. One always has the free resolution.
2. Take X a K(G, 1), that is, a connected CW complex such that
1
(X) = G and
i
(X) = 0 for i > 1.
This is unique up to homotopy type and

X is contractable.An example of a free resolution of Z over
ZG is the chain complex of

X.
Denition. (FP
n
) A group G is of type FP
n
if Z has a projective resolution
P
i
P
0
Z 0
Remark.
1. FP
1
nitely generated
2. FP
2
nitely presented. There is an example (by Bestvina-Feign) of an FP
2
which is not nitely
presentable.
Denition. (F
n
) A group G is of type F
n
if it admits a K(G, 1) with a nite n-skeleton.
Remark.
1. F
1
FP
1
2. F
2
nitely presentable i.e FP
2
.
3. In general F
n
FP
n
Theorem. If G is an abelian group and A is a ZG-module then
A
= S(G) if and only if A is nitely
generatable as a group.
Recall that a group G is metabelian if and only if there is an exact sequence
1 A G

Q 1
where A and Q are abelian. A can be viewed as a ZQ-module as follows: qa = l
1
al where l
1
(Q), and
where q Q, a A.
Theorem. (Bieri-Strebel) G is nitely presentable if and only if
A

A
= S(Q)
A is 2-tame as a ZQ-module if and only if
A

A
= S(Q). In general, for A a ZQ-module, then A is
m-tame if any m point subset of
C
A
is contained in an open hemisphere.
Conjecture. (Bieri-Groves) Given the short exact sequence
1 A G Q 1
G is of type FP
m
if and only if A is m-tame as a ZQ-module.
Remark. This conjecture is open. There are some special cases in which it has been proved:
1. If G is of type FP
m
then A
Z
k is m-tame as a KQ-module for any eld k
2. The full conjecture holds for groups of nite profer rank
3. The only if part holds true if the additive group of A is torsion or if the extension is split.
Will end with a criterion for FP
n
:
Denition. Let G be a group. Then X is a G-Space if X is a connected CW-complex with G acting on
X by cell preserving homeomorphisms.
Denition. A G-space X is m-good if
1.

H
i
(X) = 0 for i < r
2. Stabilisers of each cell in G is of type F
np
for 1 p n 1
Let D be a collection of G-invariant subcomplexes such that D = X
A
where A is a poset and X
/G has
a nite n-skeleton.
Theorem. (Brown) G is of type FP
n
if and only if

H
1
(X
) is essentially trivial.
D here is a directed system of subcomplexes: that is A if X
. For all , such that

X
and X
H
i
(X
) is essentially trivial if for all there exists such that

H
i
(X
)

H
i
(X
) is trivial.
20 Automatic Groups (Scott Messick)
Here is an example of a nite automata: this is a word acceptor for Z:
Any double circle is an accepting state. The idea is input your word on the left (hence the arrow coming
in), follow the arrows depending on your word and if you are a reduced word in Z you will end up in an
accepting state. e.g. aaaa is a reduced word, aaa
1
a
1
aa
1
is not a reduced word.
Given words
we have
(, ) (
where
= where represents the the empty word. Consider = abs, = cccdc. Then,
(, ) =
_
a b a
c c c d c
_
Thus we can say a relation on strings is FA-recognisable.
_
a a a
a a a a
_
Denition. (CGA) A Cayley graph automatic (CGA or graph automatic) structure for a group G with
semigroup generators X consists of
1. A regular (FA-recognisable) language 1 over some alphabet
2. A surjection : 1 G
3. An FA-recognising equality i.e. (, ) : () = ()
4. For each x X an FA-recognising right multiplication by x i.e. (, ) : ()x = ()
We get an automatic structure if X = and interprets strings as group words.
Lemma. In reverse binary representation, integer addition is FA-recognisable.
Proof. Here is an automaton for non-negative integers: rst consider an example of reverse binary addi-
tion:
Starting from the left, we add 1 and 0 to get 1. This is represented by
_
_
1
0
1
_
_
. Then we add 1 and 1, we get
a 0 in this column, but we carry the 1, this is represented by
_
_
1
1
0
_
_
, and so on... We get the automaton:
Proposition. 1
3
(Z) is CGA
Proof. Represent
_
_
1 a c
0 1 b
0 0 1
_
_
as (a, b, c) Multiplication is given by,
(a, b, c)
A
(1 +a, b, c)
(a, b, c)
B
(a, 1 +b, c)
(a, b, c)
C
(a, a +b, 1 +c)

Remark. Any regular subgroup of the full monoid of integer matrices, M
n
(Z) is CGA.
Lemma. Suppose f : D
n
D is FA-recognisable. Then the value f(x
1
, . . . , x
n
) can be computed in
linear time.
Proof. Fix an FA M recognising f. Calculate the set of states
S = s : y such that [y[ max
i
[x
i
[ and M running (x
1
, . . . , x
n
, y) nishes in state s
To do this recursively calculate
S
j
= s : y [y[ = j and M runs (x[
j
, . . . , x
n
[
j
, y) nishes in s
Getting S
j+1
from S
j
required O(1) time. Actually, also track witnesses (at least one for each s S
i
).
If S has an accepting state, were done. If not, continue nding S
j
for j > max
i
[x
i
[ (pad x
i
s with s). For
some j < max
i
[x
i
[ + (number of states), S
j
will have an accepting states.
Theorem. If G is a CGA then the word problem for G is solvable in O(n

2
) time.
Remark. The proof follows from the above lemma and the fact that nite automata can be done in linear
time.
20.1 Closure Properties
The class of graph automatic groups is closed under:
1. Finite extensions
2. Free products
3. Amalgamated free produce over nite groups
4. Direct products
5. Semi-direct products G
H where for all g G, (g) Aut(H) is automatic.

6. Regular nitely generated subgroups.
Remark. If G is nite, G/ Z is CGA.
Theorem. For n 1, BS(1, n) is CGA
Proof. Recall BS(1, 2)
= Z Z[
1
2
] where (n, p) (m, q) = (n +m, 2
m
p +q).
Write the elements as (n,
m
2
k
) where m is odd. Represent this as (n, m, 2
k
), which we will represent in
reverse binary again (note that this works nicely for 2
k
).
_
n,
m
2
k
_
(0, 1) =
_
n,
m+ 2
k
2
k
_
_
n,
m
2
k
_
(1, 0) =
_
n + 1,
m
2
k1
_
Theorem. If G is nilpotent of class 2 then G is CGA.

20.2 Biautomatic Groups
Denition. A (graph) biautomatic structure for G is a (graph) automatic structure for left and right
multiplication over the same language.
Theorem. Graph biautomatic groups have solvable conjugacy problem.
Proof. (Sketch) Regular languages are closed under existential quantiers. So for p G
(, ) : () p = ()
is regular. Because its biautomatic we get,
(, ) : () = p ()
is regular. Fix g, h G.
(, ) : () = () and g () = () h
then that is regular.
20.3 Open Problems

1. How many (solvable) groups are graph automatic?
2. Is there are geometric relationship, i.e. a necessary and/or sucient condition?
3. Is every (graph) automatic group (graph) biautomatic?
4. Is the automatic group isomorphism problem decidable?
5. Is Thompsons group F (graph) automatic?
21 Quasi-Isometric Rigidity (Iian Smythe)
Assume all groups nitely generated throughout.
Basic Quasi-Isometry Facts
1. If S, S
are nite generating sets for G then Cay(G, S) is quasi-isometric to Cay(G, S
)
2. Quasi-isometry is an equivalence relation on the class of metric spaces
3. IF G
= H then G is quasi-isometric to H
4. G is quasi-isometric to a point if and only if G is nite.
Fundamental Lemma Suppose G is countable, discrete and acts geometrically (i.e. properly discon-
tinuously, cocompactly by isometries) on a proper, geodesic metroc space (X, d), then G is nitely generated
and quasi-isometric to (X, d). In fact, for any xed x
0
X the map g g x
0
is a quasi-isometry.
() Suppose G and H are such that G
G, H
H are of nite index and F G
, K H
are nite with

G
/F

= H
/K
Claim. G is quasi-isometric to H
Proof. Let G
act on Cay(G, S), this is geometric. Let G
act on Cay(G
/F, T), this is geometric. Similarly

for H, H
, H
/K
Denition. If G and H are as in (), we say G and H are virtually isomorphic.

The claim says that virtually isomorphic groups are quasi-isometric.
Fact. Virtually isomorphic is an equivalence relation on the class of all groups (transitivity takes a bit of
work).
We should note that quasi-isometric does not in general imply virtually isomorphic. Heres an example of
this.
Example. Let Hyp(2, Z) be all matrices A SL
2
Z such that A is diagnosable over 1 and A
2
,= I. So if
,
are eigenvalues of A, neither has absolute value 1.

Let G
A
= Z
2
A
Z where a = Z, and a acts on Z
2
by multiplication by A. By diagonalising A, we get an
embedding
G
A
Sol = 1
2
D
1
where
D(t) =
_
e
t
0
0 e
t
_
t 1
It follows that G
A
is torsion-free (since Sol is). This embedding makes G
A
a lattice in Sol, and its action
on Sol is geometric. So G
A
is quasi-isometric to Sol.
Pick A, B Hyp(2, Z) such that A
m
is not conjugate to B
m
for any n, m Z0
A =
_
2 1
1 1
_
B =
_
3 2
1 1
_
1
2
(3
5) 2
3
Let H
A
be a nite index subgroup of G
A
. H
A
will intersect the maximal rank 2 abelian subgroup Z
2
of G
A
in a rank 2 abelian subgroup, L
A
. The image of H
A
under the quotient map G
A
G
A
/Z
2
= Z must be
innite cyclic, say generated by n. It follows that
H
A

= Z
2
A
n Z
and similarly
H
B

= Z
2
B
m Z
But any isomorphism H
A
H
B
will carry L
A
L
B
and induces a conjugation of A
n
to B
m
. Therefore
we have shown that these quasi-isometric groups are note virtually isomorphic.
Quasi-isometric rigidity is when quasi-isometry does imply virtual isomorphism.
Denition.
1. A group G is quasi-isometrically rigid if whenever H is quasi-isometric to G, H is virtually iso-
morphic to G
2. A class of groups ( is quasi-isometrically rigid if whenever H is quasi-isometric to some C (
then there exists C
( such that H is virtually isomorphic to G
.
Remark We get increasingly coarse partitions by the equivalence relations
= virtually isomorphic quasi-isometric

and we want to understand when the last implication goes both ways.
In general, if a class of groups is dened by a quasi-isometric invariant, then the class is quasi-isometrically
rigid.
Examples.
1. The class of nite groups
2. Any class dened by its growth rate. For example, nilpotent groups (by Gromovs theorem - i.e.
polynomial growth gives you nilpotent)
3. Any class dened by its Dehn function/ isoperimetric inequality. For example solvable word problem
(linear), hyperbolic groups (linear isperimetric inequality).
4. Amenable groups (G with nite generating set A is amenable if and inly if for every > 0 F G
nite such that [
S
F[ [F[)
5. Finitely presented groups. Sketch: Let S be a nite generating set for G, let L(k) denote the set of
all loops in Cay(G, S) of the form
1
where is a path e x and is a loop at x of length
k. Fact: G is nitely presented if and only if L(k) generates
1
(Cay(G, S) for some k
. .
q.i invariant
.
Theorem. (Gromov) Finitely generated abelian groups is a quasi-isometrically rigid class, in fact, Z
n
is quasi-isometrically rigid.
Theorem. (Stallings) The class of non-abelian groups is quasi-isometrically rigid.
Theorem. (Farb-Mosher) Each solvable Baumslag-Solitar groups BS(1, q) is quasi-isometrically rigid.
Theorem. (Eskin-Fisher-Whyte)
1. Being a lattice in a Sol is quasi-isometrically rigid
2. F / Z and F
/ Z are quasi-isometric if and only if there exists d, r, s such that [F[ = d

s
and [F
[ = d
r
Non-Examples
1. Theorem. (Burger-Moses) The class of simple groups is not quasi-isometrically rigid
2. Theorem. (Dyubina) The class of nitely generated solvable groups is not quasi-isometrically rigid.
Pf: We start with a lemma,
Lemma. Let A, B be bi-Lipschitz equivalent. Then for any group C, C / A and C / B are bi-Lipschitz
equivalent. (where C / A means C is acting on A)
Let D be a nite non-solvable group (e.g. A
5
). Let G = Z / Z, H = Z / (Z D). So G and H are
bi-Lipschitz. G is solvable, but H contains
Z
D, so any nite index subgroup of H will contain a
copy of D (it will actually contain
C
D where [C[ = SS
0
). So H is not virtually isomorphic to any
solvable group.
Remark. By taking A and B to be a solvable and a non-solvable nite group (resp.) of the same
order, one can get Z/ (ZA) and Z/ (ZB) to be isometric, but the latter not virtually isomorphic
to a solvable group. (So solvability is not a geometric property)
Open Questions.
1. Is the class of polycyclic groups quasi-isometrically rigid?
2. Is the class of nitely presented solvable groups quasi-isometrically rigid?
22 Quasi Isometric Rigidity (Kristen Pueschel)
Positive results for QI rigidity
Mostow Rigidity Finite volume hyperbolic manifolds of dimension at least 3 are determined up to
isometry by their fundamental groups
Eskin, Fischer, Whyte Let be quasi-isometric to the discrete cocompact lattice Sol. Then is
virtually a discrete cocompact subgroup of Sol.
Farb-Mosher
1. BS(1, n) QIBS(1, m) if and only if n = r
, m = r
if and only if commensurable

2. BS(1, 2) is QI rigid
22.1 Proof Form for QI Rigidity
Step 1 Work out QI(X) = f : X X : f quasi-isometry/f g if they dier within a constant
Example (EFW) Bulk of the paper spent doing this.
Any (K, C)-quasi-isometry of Sol, is within bounded distance of a height-respecting quasi-isometry
.
Step 2 If f : X is a quasi-isometry, then there is an induced homomorphism QI(X) by g
f L
g
f
1
. Try to show this map has nite kernel.
Step 3 Show QI(X) acts on the boundary of X. Study the dynamics of this action.
22.2 X
n
and its Boundary
BS(1, n) has a Cayley 2-complex, X
n
.
Facts
1. X
n
has embedded hyperbolic planes - describe these as lines in the tree T
n
(the cross section)
2. There are trees in X
n
isometric to T
n
- describe these as sections of X
n
to H
2
(think of collapsing
everything up to a given height). These correspond to vertical lines in H
2
. So there is a tree for each
element of 1. Code these as elements of
n
- the n-adic rationals,
iZ
i
n
i
for i 0, . . . , n 1.
(Note: Eventually zero in the negative direction)
Denition
1. The upper boundary
u
(X
n
) = set of hyperbolic planes
2. The lower boundary
l
(X
n
) = set of trees 1
BS(1, n) acts on these boundaries: BS(1, n) = a, t : tat
1
= a
n
, action is dened as:
l
(X
n
) : a x = x + 1
t x = nx
u
(X
n
) : a x = x + 1
t x = nx
Proposition. If f : X
n
X
m
is a quasi-isometry then for each hyperbolic plane Q X
n
(reps. tree T)
there is a unique hyperbolic plane Q
X
m
(resp. T
) such that
d
H
(f(Q), Q
) C
This induces a map between boundaries f
u
:
l
(X
n
)
l
(X
m
). f
u
and f
l
are bi-Lipschitz maps.
Step 1 Identify QI(BS(1, n))
Theorem. QI(BS(1, n)) Bilip(1) Bilip(
n
)
Step 2 QI rigidity condition: We have an induced homomorphism g f L
g
f
1
This action is properly
discontinuous and cocompact (note: this is actually a quasi-action, and as such we have things close
to properly discontinuous and close to cocompact, but in QI(X
n
) this is ne).
Denition. Suppose X, Y are two topological spaces. T(X, Y, Y ) = (x, , ) : x X, , Y, ,= .
Similarly T(X, X, Y ) = (x, y, ) : x, y X, Y, x ,= y. Then G acts uniformly biconvergently on
(X, Y ) if the diagonal actions T(X, X, Y ) and T(X, Y, Y ) are properly discontinuous and cocompact.
Proposition. BS(1, n) acts uniformly biconvergently on (
l
(X
n
),
u
(X
n
)).
Pf (Sketch). We can send T(X, Y, Y )

X
n
continuous, proper isometry.
d( (f
l
(x), f
l
(y), f
l
())) f((x, y, )) c
Theorem. (Contraction Property) Let G be a nitely generated group quasi-isometric to X

n
. Let
(
l
,
u
) : G (Bilip(1), Bilip(
n
)). For every compact K
n
, and every open subset U
n
there
exists g G such that g K U.
l
: G Bilip(1) projection. Result: there is a Bilip(1) such that for all g G (g) =
l
(g)
1
Aff(1), (g)x = m
g
x +bx
G
1
//
A
//

//
Stretch()
//
1
1
//
Isom(1)
//
Aff(1)
//
SL(1, 1)
//
1
Construction principle:
1. has nite kernel
2. Stretch()
= Z. = AZ ...some work... = B, t : t
1
Bt = B
1
where B
1
B A.
23 Random Walks on Solvable Groups (Johannes Cuno)
23.1 Random Walks on Graphs and Groups
On a Graph Require
1. locally nite
2. connected
3. a probability function that determines the probability of moving to a given adjacent vertex:
p : V V [0, 1]
v V :
wV
p(v, w) = 1.
4. concentrated on neighbours. That is, supp(p(v, )) = neighbours of v
k-step transition function:
p
(k)
(v, w) =
wV
p
(k1))
(v, w
) p(w
, w)
Denition. The random walk is irreducible if and only if v, w V k N
0
: p
(k)
(v, w) > 0
A random walk X = (X
0
, X
1
, . . .) takes values in V
= (v
0
, v
1
, . . . , ) : v
i
V . Dene a -algebra: a
cylinder set
P
v
= (v
1
, v
2
, . . . , v
n
, , , )
=
v,v
0
p(v
0
, v
1
) p(v
1
, v
2
) p(v
n1
, v
n
)
where we have the rst, say, n values determined, and afterwards anything can happen. Dene
= (cylinder sets) = smallest -algebra containing the cylinder sets
This makes (V
, , P
v
) a measure space.
On a Groups G a nitely generated group, S G a nite generating set, a measure on G with support
S S
1
(tells you the probability of going to an adjacent edge). Here,
P
e
= g
0
, g
1
, . . . , g
n
, , , . . .)
=
e,g
0
(g
1
0
g
1
)(g
1
1
g
2
) (g
1
n1
g
n
)
23.2 Examples
Rie Shue
G = S
52
according to Gilbert-Shannon-Reeds.
After one Rips shue we cant get too far away from the starting arrangement. Indeed, the relative order
of each of the cut decks remains the same. After repeatedly shuing we would like to know how long it
takes to get close to a uniform distribution. Need a notion of distance between probability distributions:
d
TV
(p
(n)
(e, ), unif
S
5
2
()) =
1
2
gG
[p
(n)
(e, g) unif(g)[
Here are some values of this function:
n 1 2 3 4 5 6 7 8 9 10
d
TV
1.000 1.000 1.000 1.000 0.924 0.624 0.312 0.161 0.083 0.041
So we get exponential decay after a certain cut-o time. That is, we see that after 7 or 8 rie shues we
get very close to the uniform distribution.
Simple Random Walk on Z
d
G = Z
d
, S = standard generating set.
(g) =
_
1
2d
g S S
1
0 else
Denition. w V is recurrent if one of the following equivalent conditions holds:
1. P
v
many n N
0
: X
n
= w = 1
2. P
v
n N
0
: X
n
= w = 1
3. E
v
[[n N
0
: X
n
= w[] =
n=0
P
v
X
n
= w =
Remark Intuitively recurrence says that the random walker will return to the given point innitely many
times. Recurrence does not depend on w V . If not recurrent then say you are transient. We see
transience then as saying that after a point the walker never returns to a given point.
Theorem. (Polya) The simple random walk X = (X
0
, X
1
, X
2
, . . .) on Z
d
is recurrent for d 1, 2 and
transient for d 3.
23.3 Flows and Energy
Denition. A ow is a function from the oriented edges
f : E 1
We think of the ow as follows: suppose we pump water through our vertex, and imagine the oriented
edges connected to the vertex as pipes - then the ow tells us how mush water ows through each pipe.
Deniiton. The energy is
eE
f(e)
2
.
Theorem (Lyons) X reversible random walk (e.g. simple random walk on a graph). Then transient if
and only if f with nite energy.
We can use this to prove Polyas theorem.
Proof of Polyas Theorem
d = 1: Consider an arbitrary ow f : E 1. There must a constant non-zero ow in at least one
direction (left or right), indeed there is just one choice at the beginning (how much ows left, how
much ows right) and then everything is xed. Therefore the energy must be innite and thus the
random walk is recurrent.
d = 2: Consider the spheres of size n and n + 1. One observes that there are 8n + 4 edges connected
the n-sphere to the (n + 1)-sphere. Say [E
n
[ = 8n + 4[.
eE
f(e)
2
=
n=0
eE
f(e)
2
n=0
1
[E
n
[
_
eE
f(e)
2
_
. .
=1
=
n=0
1
8n + 4
=

Remark. Using the ow criterion from Lyons theorem one sees that recurrence/transience is independent
of choice of nite generating set S.
23.4 Growth of Groups and Return Probability
Another question one can ask is, what are the return probabilities. That is, dene the return probability
(n) = P
e
X
2n
= e
Motivating the forcing of parity above (need n even) is the example of simple random walks on Z
d
. Indeed,
it is impossible to return to the origin in Z
d
after an odd number of steps. In this example one obtains the
result that,
(n) n
d/2
Where a(n) b(n) means : a(n) _ b(n) (that is c > 0 : a(n) cb(n)), and vice versa.
How does (n) behave?
1. V (n) = [B
e
(n)[ n
d
(n) n
d/2
2. V (n) exp(n
) for some [0, 1] (n) exp(n
+2
)
In the case of exponential growth
(n) exp(n
1/3
)
23.5 Advertisement for Next Talks
Dene a function
h : V 1
Think of this as saying how much money you get at a given vertex. Want to understand how the expected
payo varies as you random walk. Dene the function
Ph(g) =
G
p(g, g
)h(g
)
which tells you the expected earning of the next step. We think of dening a discrete analogue of the
Laplacian as Ph h, and then a harmonic function is one where this is 0, that is the set
1 = h : Ph = h
and also superharmonic functions
o = h : h Ph
Remark. Notice rst that any constant function h is harmonic (and therefore superharmonic). What
about non-constant harmonic functions?
Theorem A random walk on (G, , P
v
) is recurrent if and only if o
0
= constants.
Questions.
1. Martin boundary: non-constant h 1
0
2. Poisson boundary non-constant bounded h 1
Tail -algebra
= (cylinder sets). (V
, , P
e
). Consider
n
= elements of being determined by (g
n
, g
n+1
, )
Dene a new -algebra
=
nN
n
and the triple
(G
, P
e
)
24 Boundaries of Random Walks (Mathav Murugan)
Recall from last time, (G
, P)
1. Martin Boundary:
- Topological space
- Potential Theory
- Integral Representation for non-negative harmonics
2. Poisson Boundary:
- Measure Space
- Ergodic/Measure theory
- Representation for bounded harmonic function
Denition. p
1
(x, y) = p(x, y) the probability measure, p(x, y) = (x
1
y) We say is non-degenerate
if supp() generates G as a semi-group.
n 2 then
p
n
(x, y) =
zG
p
n1
(x, z)p(z, y)
Denition. A function f : G 1 is said to be harmonic on (G, ) if
f(x) =
gG
(g)f(xg)
24.1 Martin Boundary
Denition. Greens function G(x, y) =
n=0
p
n
(x, y)
If G(x, y) < the random walk is said to be transient. Fix o G.
K
0
(x, y) =
G(x, y)
G(o, y)
x, y
G(x, y) = F(x, y))G(y, y) where F(x, y) = probability that a random walk starting at x ever visits y.
1 F(x, y) > 0
F(x, y) F(x, z)F(z, y)
F(o, x) K
0
(x, y) =
F(x, y)
F(o, y)

1
F(o, x)
K
0
(, y) : y G is relatively compact in topology of point wise convergence.
Denition. Martin compactication is the pointwise closure of K
0
(, y) : y G
Denition. The Martin boundary,
M
G is the topological boundary of K
0
(, y) : y G Denote

M
G as function K
0
(, ).
Denition. A positive harmonic function is said to be minimal if for all harmonic functions f
such that
0 f
f then f
= tf for t 1.
Facts
1. All minimal harmonic functions belong to the Martin boundary (with normalisation)
2. Minimal harmonic functions form a Borel subset of
M
G called the minimal Martin boundary,
m
G.
3. Every harmonic function can be uniquely represented as an integral of minimal harmonic functions
f(x) =
_
m
G
K
0
(x, y)d
o,f
()
The measure corresponding to the constant function
o
=
0,1
is the harmonic measure.
4. (
m
, G,
0
) is isomorphic to the Poisson boundary (to be dened).
Examples
Theorem. (Choquet-Deny) All minimal harmonic functions in an abelian group are 1
+
-valued multi-
plicative characters
Proof. Let f be a minimal harmonic function.
T
x
f(g) = f(x, g) = f(g, x)
Suppose (x) > 0.
f(g) (x)f(gx)
Now
T
x
f(g) = f(xg) =
(y)f(xgy) =
(y)T
x
f(gy)
using that f is a minimal harmonic function.
f(gx) = f(g) g G x : (x) > 0
hence f is an 1
+
-valued multiplicative character.
1 is a minimal harmonic for abelian groups. It is true for nilpotent groups a result of (Dynkin-Maljutov).
Example. Free group generated by a
1
, . . . , a
n
write a
1
i
= a
i
f(a
i
1
a
i
n
) =
_

n1
k=0
1
(2n1)
k
i
1
> 0
n1
k=0
1
(2n1)
k
i
1
< 0
Martin boundary is space of ends is innite irreducible words. Let = a
i
1
a
i
2
. Call the rst n terms in
this expression
n
.
lim
n
K
0
(x,
n
) = K
0
(x, ) = K
0
(, )
24.2 Poisson Boundary
T : G
by (x
n
) (x
n+1
). x, x
are equivalent (write x x
) if
T
n
(x) = T
n
(x
)
for some n, n
N. Let S be the measurable union of the -equivalence classes. S forms a -algebra (, ).

There exists a map b : G
such that coincides with pre images of measurable sets in .

Denition. Entropy of the probability measure p = (p
i
) is dened as
H(p) =
p
i
log p
i
If n is the size of the support then H(p) log n if
n
i=1
p
i
= 1.
Dene measure
n
(g) =
_
(g) n = 1
zG
n1
(z)(z
1
g) n > 1
Dene
h(G, ) = lim
n
1
n
H(
n
)
H( ) H() +H()
Theorem. (Kaimanovich-Vershik) (r, ) is trivial if and only if h(G, ) = 0
Corollary. For any group of sub exponential growth and any nitely supported measure (r, ) is trivial.
Proof.
1
n
H(
n
)
1
n
log [supp(
n
)[
Theorem. (Furstenbery) Let (G, ) be a non-degenerate markov chain. Let G

0
be a subgroup of G
which is a recurrent set. Dene a measure
0
on G
0
as the distribution of the point where the walk starting
from the identity hits G
0
for the rst time. Then r(G, )
= r(G
0
,
0
)
Proof. L
(
..
f
, )
= H
( G
..
f
, ). f(g) =
f, g. f = 0 in U, f = g in U.
Alternatively, f(g) = E
g
f(X
). How does (G, )
= (G
0
,
0
) work?
f f
0
by restriction
f
0
f f(g) = E
g
(f(x
i
))
Theorem. G polycyclic, nitely supported symmetric measure. Then the Poisson boundary is trivial.
Theorem. (Kaimanovich-Vershik) If Gis non-amenable then (, ) is non-trivial every non-degenerate
(G, )
If G is amenable then there exists a symmetric non-degenerate probability measure such that (, ) is trivial.
25 Rate of Escape of Random Walks on Groups (Tianyi Zheng)
2 diusive lower bound
Eischler, James and Peres
E[S
n
[ C
_
n
2[S[
where C is the universal constant and G is amenable. (G nitely generated amenable then there exists a
non-constant equivalent harmonic map : G 1 where 1 is some hilbert space.
Question. (Vershik-Peres) For what [
1
2
, 1] does there exist G, such that
E[S
n
[ n
+o(1)
where f(n) g(n) if there exists c, C > 0 such that
cg(cn) f(n) C(GCn)
Erschler G
1
= (Z/2Z) / Z, G
2
= G
1
/ Z, G
i+1
= G
i
/ Z. Denote by L
G,
(n) = E[S
n
[
Theorem. L
G
i
,switch-walk-switch
(n) n
1
1
2
i
The range of the simple random walk on Z at time n n
1
2+o(1)
. On Z
2
the range is like
n
log n
.
Now repeat this construction with:
H
1
= (Z/2Z) / Z
2
, H
i+1
= H
i
/ Z
2
Theorem. L
H
i
,switch-walk-switch
(n)
n
log
i
n
where log
i
n = log log
. .
i
n
Amir-Virag Given any function f such that
a
3
4
f(n) f(an) af(n)
for example f(n) = n
with [
3
4
, 1]. Then there exists G = Z /
"
m
/
m
such that
E[X
n
[ f(n)
(where /
m
is the piecewise-linear mother group)
Questions.
1. Are there examples with (
1
2
,
3
4
)?
2. What are the possible behaviours of rate of escape on nitely presented groups? (Polycyclic groups
have diuse behaviour)
3. What are the possible behaviours in solvable groups? (Free solvable groups, = 1, e.g. free metabelian
group F
2
/F
2
)
Proof of Diuse Lower Bound E[S
n
[
2
. : G Ane Isometries of 1
: G 1
with harmonic equivalriant.
(g) =
sS
1
(gs)(s)
(gh) = (g)(h)
(S
n
) is a martingale E[(S
n+1
)[S
1
. . . S
n
] = [S
n
E[[(S
n
)[[
2
H
= E([[(S
n
) (S
n1
+ + (1) (e)[[
2
H
)
= E[[(S
n
) (S
n1
)[[
2
H
+ +E[[(S
1
) (e)[[
2
H
= nE[[(X
1
) (e)[[
2
H
normalise such that E[[(X
1
) (e)[[
2
H
= 1. is
1
2|S|
Lipshitz.
E[S
n
[
2
1
2[S[
E[[(S
n
)[[
2
=
1
2[S[
n
E[S
n
[ c
_
n
2[S[
In the upper bound direction E[S

n
[ _ n
. One can use equivalent embedding into a Banach space (Naor-

Peres)
#
L
p

1
p (G)
Tessera establishes a connection between /
2
-isoperimetry prole and Hilbert compression.
Corollary. Class L containing polycyclic groups F / Z, BS(1, n) diuse speed exponent n
1
2
+o(1)
.
Question. Characterise groups with diuse behaviour.
Revelle Z
d
/ Z. Switch-walk-switch
X
1
, . . . , X
n
, on the lamps
Z
1
, . . . , Z
n
, . . . on the base .
W
n
= (f
n
, V
n
) ( )
n
Writing S
n
= X
1
+ +X
n
V
n
= Z
1
Z
n
on base. f
n
(i)
(X
j
+X
j+1
)1
Z
j
=i
S
2(i,n)
where (1, n) =
n
i=1
1
Z
j
=i
[W
n
[

iZ
[F
n
(i)[ + size of rangeV
1
, . . . , V
n
Method of Revelle: Exponential tightness of random walk on the lamps gives tightness normal of W
n
P([S
n
[ xn
2
) c
1
exp(c
2
x
) valid for all x.

Example. Simple random walk on Z
d
, P([S
n
[ xn
1/2
c
1
exp(c
2
x
2
) satises a Gaussian tail.
iZ
[f
n
(i)[ =
iZ
[f
n
(i)[
i,n

i,n
iZ
([f
n
(i)[
i,n
)
1
1
_
1
. .
iZ
i,n
1
. .
=n
Lemma. X
1
, . . . , X
n
conditionally independent, each of themP([X
i
[ > x) c
1
exp(x
) then P(
n
i=1
X
i
>
n) c
1
exp(c
2
). In the end gives n

/(1)
Example. Z
2
A
Z where A =
_
2 1
1 1
_
. Character polynomial
2
3 + 1 = 0
Z / Z
W
n
=
iZ
A
i
S
(i)
2(n,i)
Index
0
-invariant, 52
1
-invariant, 52
-ultralimit, 10
-algebra, 9
-algebra generated by S, 10
Asymptotic Cone, 11
Baumslag-Solitar Groups, 4
Biautomatic Structure, 62
Cayley 2-Complex, 14
Cayley Graph Automatic Structure, 60
Centralised Area, 24
Centralised Isoperimetric Functions, 24
Character, 52
Character Sphere, 52
Commutator Subgroup, 1
Compression Words, 21
Dead-end Element, 37
Depth, 37
Derivation, 41
Derived Length, 2
Derived Subgroup, 1
Energy, 69
Filling Length, 14
Flow, 69
Fox Derivative, 41
Free Nilpotent Group, 24
G-Space, 58
Geodesic Words, 5
Growth Function, 6
Harmonic Functions, 70
Irreducible Random Walk, 68
K(G, 1), 57
Lamplighter, 4
Lower Central Series, 2
m-Tame, 58
Max Condition, 2
Nilpotent Group, 2
Non-principle Ultralter, 10
Null-sequence, 19
Poisson Boundary, 70
Polycyclic Group, 2
Presentation 2-Complex, 14
Probability Measure, 10
Projective Resolution, 57
Quasi-isometry, 8
Quasi-Isometry Rigid, 14
Recurrent Random Walk, 69
Solvable Group, 1
Superharmonic Functions, 70
Transient, 69
Treebolic Space, 33
Type F
n
, 57
Type FP
n
, 57
Upper Central Series, 2
van Kampen Diagram, 14
Area, 14
Diameter, 14
Filling Length, 15
van Kampens Lemma, 18
Virtually Nilpotent, 9
Word Metric, 5
Wreath Product, 4
77

The Geometry of Solvable Groups-Amin Saied

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Geometry of Solvable Groups-Amin Saied

Uploaded by

Copyright:

Available Formats

Spring 2013

= [G, G]. The

d(a, b) d((a), (b)) d(a, b) +

(X, e, s). Then

(Y, (e), s). As is a quasi-isometry (and in particular we are

Proposition. For a nitely generated group , the asymptotic cone,

(Cay(), e, s). So I can just prove it for the latter space.

(X, e, s) is a geodesic space.

(X, e, s). Then there exists a geodesic path

and the rst letter of u is not b

denes an asynchronously fellow-travelling combing. The re-write rules:

represents 1 in G. Then w represents a loop in the Cayley graph.

9 Feb 26, 2013

by carrying all the zs to the right end, and all the z

10 Feb 29, 2013

= a, b : [a, b] the Area(n) n

Centralised Isoperimetric Functions

Denition. Let F = F(a, b). Write F

, and so must be of the form t

Proposition. (Normal Form) Every g G

can be represented uniquely as

This is an HNN-extension where : mZ nZ by a

, consider innite sequences of integers. f : (a

12.1.3 Cayley Graph of BS(1, 2)

Z is quasi-isometric to Sol. Hence we get

Tim Riley The Geometry of Solvable Groups 36

= [N, N]. The aim is to embed,

the Magnus embedding.

16.6 Geometric Denition of the Magnus Embedding

This result implies that ker

Tim Riley The Geometry of Solvable Groups 44

is abelian. Thus we have the following abelian series for

Then G has the presentation,

. Simple induction to complete (e.g. s

is the largest abelain quotient of G.

is abelian (by proceeding proof), and G

is free abelian of innite rank. Therefore

Proposition. G is nitely presented. In fact, G =

into the wreath product A/ H.

, k K, [f, f] = [f, k] = [k, k] = 1)

N then w K R, thus upper left entry

Proof of Baumslag-Remeslennikov Theorem

3. Repeat this r times resulting in G

is the normal closure of A

Claim 2 G embeds in G and G is metabelian.

What relations do we have?

/N some normal subgroup N G

satises max-n N is nitely generated. Therefore G is too.

is an abelian series of derived length 2 and therefore G

is abelian one uses the special polynomials p

Tim Riley The Geometry of Solvable Groups 51

. We consider this as a module over ZG where scalar multiplication is by conjugation:

a module over ZG by conjugation. If G is nitely presented and solvable then

f.g over a f.g. submonoid of G

Say G is nitely generated by X. (G, X) the Caley graph wrt X.

be the subgraph of (G, X) with vertices g such that (g) 0, and

connect an edge if there was one in (G, X).

(p) = min(g) : g vertex on p

. Then there exists a path p

Tim Riley The Geometry of Solvable Groups 54

has rank 2 over Z, then there exists a normal

Tim Riley The Geometry of Solvable Groups 56

nitely generated wrt ZG

). So this follows by denition.