Topology (H) Lecture 24

Lecturer: Zuoqin Wang

Time: June 03, 2021

APPLICATIONS OF FUNDAMENTAL GROUPS

1. Applications of π1 (S 1 ) ' Z
π1 (S 1 ) is not only the first non-trivial fundamental group, but also the most im-
portant fundamental group. We list a number of applications:

¶ Compute the fundamental groups of many simple spaces.

• π1 (S 1 × Rn ) ' Z, ∀n; π1 (Rn \ Rn−2 ) ' Z, ∀n ≥ 2.
1
• π1 (S
| × ·{z· · × S 1}) ' Z
| ⊕ ·{z
· · ⊕ Z}; 1
· · ∨ S 1}) ' Z
| ∨ ·{z
π1 (S | ∗ ·{z
· · ∗ Z}.
r r r r
• π1 (Möbius strip) ' Z, π1 (S 2 ∨ S 1 ) ' Z.
• π1 (S 2 ∪ {(0, 0, z) | |z| ≤ 1}) ' Z, π1 (R3 \ {(x, y, 0)|x2 + y 2 = 1}) ' Z.
We have seen more in PSets and we will see more below and in today’s pset.

¶ Distinguishing R2 with Rn .
Spectral invariants are widely used to distinguish topological spaces. For example,
Proposition 1.1. For n > 3, Rn and R2 are not homeomorphic.

Proof. Suppose Rn ' R2 , then Rn \ {0} ' R2 \ {0}. But

Rn \ {0} ∼ S n−1 , Rn \ {0} ∼ S 1
and π1 (S n−1 ) = {e}, π1 (S 1 ) ' Z, a contradiction! 
Remark 1.2. Note that
• Using path connectivity (i.e. π0 ), R is not homeomorphic to Rn for n > 2.
• Using π1 , we just saw R2 is not homeomorphic to Rn for n > 3.
• In general, using πk one can prove Rk is not homeomorphic to Rn for n > k + 1.

¶ Non-retraction.
One can use π1 to prove that a subspaces is not a retract of the whole space:
Proposition 1.3. The circle S 1 is not a retract of the closed unit disk D.

Proof. By PSet 10-1-3(a), if r : D → S 1 is a retraction, then r∗ : π1 (D) → π1 (S 1 ) is

surjective, which is impossible sine π1 (D) ' {e}. 

1
2 APPLICATIONS OF FUNDAMENTAL GROUPS

Similarly one can prove: S 1 (as the equator) is not a retract of S 2 , S 1 ∨ S 1 is not
a retract of S 1 × S 1 etc.
One might think that this method works only if the fundamental group of the
subspace is larger than the fundamental group of the whole space. This is not the case:
Proposition 1.4. The boundary circle is not a retract of the Möbius strip.
Proof. Let M be a Möbius strip, B its boundary circle and C its center circle. Then
along the deformation retraction of M to C, B gets mapped onto C with two rotations.
It follows that the induced map ι∗ : π1 (C) ' Z → π1 (M ) ' Z is given by n 7→ 2n. If
there is a retraction r : M → B, then r ◦ i = Id implies r∗ (2n) = n, which can’t be a
group homomorphism, a contradiction. 
¶ Brouwer’s fixed point theorem (n = 2).
Theorem 1.5 (Brouwer’s fixed point theorem, n = 2). For any continuous map f :
D → D, there exists p ∈ D s.t. f (p) = p.
Proof. Suppose f (p) 6= p for all p ∈ D. Then we can define g : D → S 1 by
g(p) = the intersection point of S 1 with the ray f (p)p.
One can check that the map g is continuous, and g|S 1 = Id|S 1 . So S 1 is a retract of D,
which is a contradiction! 
¶ The fundamental theorem of algebra.
Let p(z) = z n + an−1 z n−1 + · · · + a1 z + a0 be a polynomial with complex coefficients.
Theorem 1.6 (The fundamental theorem of algebra). ∃z0 ∈ C s.t. p(z0 ) = 0.
Proof. Suppose p has no root. We define a continuous map
f : S 1 → S 1, z 7→ p(z)/|p(z)|.
a0
Then f is null homotopic (to the constant map f0 (z) ≡ |a0 |
) via the path homotopy
F : [0, 1] × S 1 → S 1 , (t, z) 7→ p(tz)/|p(tz)|.
By Lemma 3.6 in Lecture 19, we conclude that the induced group homomorphism f∗ :
π1 (S 1 ) → π1 (S 1 ) is the “zero” homomorphism, i.e. f∗ (m) = 0 for all m ∈ Z ' π1 (S 1 ).
On the other hand, f is homotopic to f1 : S 1 → S 1 , z 7→ z n via1
z n + an−1 tz n−1 + · · · + a1 tn−1 z + a0 tn
G : [0, 1] × S 1 → S 1 , (t, z) 7→ n .
|z + an−1 tz n−1 + · · · + a1 tn−1 z + a0 tn |
However, sine f1 (γ1 ) = γn , the group homomorphism (f1 )∗ : π1 (S 1 ) → π1 (S 1 ) is given
by (f1 )∗ (1) = n and thus f∗ is not the zero homomorphism, a contradiction. 
1Note: the numerator is continuous, is obviously nonzero for t = 0, and is nonzero for t 6= 0 since
in this case it equals tn p(z/t).
 APPLICATIONS OF FUNDAMENTAL GROUPS 3

¶ Borsuk-Ulam Theorem (n = 2).

In Lecture 15 we used the connectedness (i.e. π0 ) to prove the following
Theorem 1.7 (Borsuk-Ulam Theorem for n = 1). For any continuous map f : S 1 →
R, there exists x0 ∈ S 1 such that f (x0 ) = f (−x0 ).

Now we use π1 to prove

Theorem 1.8 (Borsuk-Ulam Theorem for n = 2). For any continuous map f : S 2 →
R2 , there exists x0 ∈ S 2 such that f (x0 ) = f (−x0 ).

Proof. Suppose to the contrary that there exists a continuous map f : S 2 → R2 such
that f (x) 6= f (−x) holds for all x. Define a map g : S 2 → S 1 by
f (x) − f (−x)
g(x) = .
|f (x) − f (−x)|
Then g is continuous and antipodal-preserving, i.e. g(−x) = −g(x).
Let ι : S 1 → S 2 be the inclusion of equator to the sphere and let h := g◦ι : S 1 → S 1 .
Then for any [γ]p ∈ π1 (S 1 ),
h∗ ([γ]p ) = [h ◦ γ]p = [g ◦ ι ◦ γ]p = g∗ ([ι ◦ γ]p ) = g∗ (e) = e.
It follows that h can be lifted to a continuous map h̃ : S 1 → R. Note that h is
antipodal-preserving, i.e. h(−x) = −h(x). So for any x ∈ S 1 ,
p ◦ h̃(−x) = h(−x) = −h(x) = −p ◦ h̃(x).
It follows that h̃(−x) − h̃(x) ∈ Z + 21 , and thus by connectedness, there exists m ∈ Z
such that h̃(−x) − h̃(x) = m + 12 for all x ∈ S 1 . Now let F (x) = h̃ ◦ p, where p : R → S 1
is the canonical covering map. Then for any x ∈ R,
1 1 1 1
F (x + ) = h̃ ◦ p(x + ) = h̃(−p(x)) = h̃(p(x)) + m + = F (x) + m + .
2 2 2 2
As a consequence, we get
1 1
F (1) = F ( ) + m + = F (0) + 2m + 1 6= F (0),
2 2
which is a contradiction since p(0) = p(1). 

By using higher invariants like homotopy groups or homology groups, one can prove
Theorem 1.9 (Borsuk-Ulam Theorem). For any continuous map f : S n → Rn , there
exists x0 ∈ S 2 such that f (x0 ) = f (−x0 ).
Corollary 1.10. No subset in Rn is homeomorphic to S n .

Proof. By Borsuk-Ulam, there exists no injective continuous map f : S n → Rn . 

4 APPLICATIONS OF FUNDAMENTAL GROUPS

¶ Borsuk-Ulam and Lusternik-Schnirelmann for all n.

We say a map defined on S n is antipodal-preserving if f (−x) = −f (x) for all x ∈ S n .
It turns out that Borsuk-Ulam theorem has many equivalent forms. (The proof will be
left as an exercise.)
Proposition 1.11. The following statements are equivalent (and thus are all true):
(1) For any continuous f : S n → Rn , there exists x0 ∈ S n s.t. f (x0 ) = f (−x0 ).
(2) There does not exist an antipodal-preserving continuous map f : S n → S n−1 .
(3) For any antipodal-preserving continuous map f : S n → Rn , there exists x0 ∈ S n
s.t. f (x0 ) = 0.
(4) There does not exist continuous map f : B n → S n−1 s.t. the restriction to the
boundary of f , f |S n−1 : S n−1 → S n−1 , is antipodal-preserving.
(5) Let F1 , F2 , · · · , Fn+1 be a covering of S n by closed sets, then there exists 1 6
i 6 n + 1 s.t. Fi ∩ (−Fi ) 6= ∅.
(6) Let U1 , U2 , · · · , Un+1 be a covering of S n by open sets, then there exists 1 6 i 6
n + 1 s.t. Ui ∩ (−Ui ) 6= ∅.
Remark 1.12. The theorem was proved by Borsuk in 1933. The statement (1), i.e. the
Borsuk-Ulam theorem, was first conjectured by Ulam. The assertion (5)-(6) was first
proved by Lusternik and Schnirelmann using combinatorial method, and is frequently
called the Lusternik-Schnirelmann-Borsuk theorem.

¶ Ham-Sandwich Theorem.
Finally we prove the Ham-Sandwich Theorem that we mentioned in Lecture 1:
[But unfortunately I can only prove the version for n = 2, since I only established the
Borsuk-Ulam theorem for n = 2. You can prove it for general n in a similar way as
long as you assume Borsuk-Ulam for general n.]
Corollary 1.13 (Ham-Sandwich Theorem, n = 2). Given two (nice, at least measur-
able) bounded sets in R2 , there exists a line that cut each of the two sets into halves.

Sketch of proof. For any u = (u0 , u1 , u2 ) ∈ S 2 , if u1 or u2 6= 0, then denote

Lu = {(x1 , x2 )|u1 x1 + u2 x2 < u0 };
if u = (±1, 0, 0), then denote
L(1,0,0) = R2 , L(−1,0,0) = ∅.
Denote the two sets by A and B. One can check the map
f : S 2 → R2 , u 7→ (area(A ∩ Lu ), area(B ∩ Lu ))
is continuous. So there exists u0 = (u00 , u01 , u02 ) ∈ S 2 s.t. f (u) = f (−u), which implies
that the line u01 x1 + u02 x2 = u00 bisect each of the two sets into two halves. 
 APPLICATIONS OF FUNDAMENTAL GROUPS 5

2. Other applications of fundamental groups

¶ Null-homotopic.
By definition any continuous map to a contractible space is null-homotopic.
Proposition 2.1. We have
(1) Any continuous map f : S 2 → S 1 is null-homotopic.
(2) Any continuous map f : RP2 → S 1 is null-homotopic.
Proof. (1) Since Im(f∗ ) = {e}, f can be lifted to a map f : S 2 → R which is null-
homotopic. Thus f = p ◦ f˜ is null-homotopic.
(2) Although π1 (RP2 ) ' Z2 6= {e}, the only group homomorphism from Z2 to Z is the
trivial homomorphism. Thus we still have Im(f∗ ) = {e}, and the same argument as
above shows that f is null-homotopic. 

¶ Find all covering spaces.

By the classification theorem of covering spaces, the covering spaces of a (nice)
topological spaces are in one-to-one correspondence with the subgroups of its funda-
mental group. Using this one may be able to find all covering spaces of a topological
space. For example, in PSet11-2-3 you were asked to find all covering spaces of S 1 ∨ S 2
and S 1 × S 1 . Here is another example: Since
π1 (RPn × RPm ) ' Z2 ⊕ Z2 = {e, a} ⊕ {e, b}
with subgroups
{e}, {e, a}, {e, b}, {e, ab}, Z2 ⊕ Z2 .
we conclude that the covering spaces of RPn × RPm are:
S n × S m, RPn × S m , S n × RPm , RPn × RPm , S n × S m /(v, w) ∼ (−v, −w).

¶ Use of van Kampen theorem: The fundamental groups of graphs.

Let’s explain by an example.
Let G0 = a maximal tree in G (which is
simply-connected and contains all vertices
G=
of G), as marked by red in the picture.
Label the remaining edges by e1 , e2 , e3 , · · ·
To compute the fundamental group of G, we start with G0 . Since G0 is a tree
which is contractible,
π1 (G0 ) ' {e}.
Now we add the edge e1 and compute the fundamental group of G1 = G0 ∪ e1 . The
decomposition of G1 into open subsets is shown as the following picture: Since U1 ∩ V1
is contractible, while V1 is homotopy equivalent to S 1 , we get
π1 (G1 ) ' π1 (G0 ) ∗{e} Z = Z.
6 APPLICATIONS OF FUNDAMENTAL GROUPS

Then we repeat the same procedure by adding e2 to get

π1 (G2 ) = π1 (G1 ∪ e2 ) ' π1 (G1 ) ∗{e} Z ' Z ∗ Z
and finally
π1 (G) ' Z ∗ Z ∗ Z.
In general, given any connected graph G,
• there always exists a maximal tree G0 (simply-connected and contains all ver-
tices of G). [Proof: Induction for finite tree. Use Zorn lemma for general case.]
• If we let {eα } be the set of all edges in G that are not in G0 , Then
π1 (G) ' ∗eα Z.
• For the case of finite graph G, if we denote
|V (G)| = the number of vertices in G
and
|E(G)| = the number of edges in G,
then the number of edges in G0 is |V (G)| − 1. It follows
π1 (G) ' Z
| ∗ ·{z
· · ∗ Z},
k

where k = |E(G)| − |V (G)| + 1.

¶ Application to algebra: Nielsen-Schreier Theorem.

Not only algebra can be used to study topology, but also topology can be applied
to give a quick proof of the following theorem in algebra:
Theorem 2.2 (Nielsen-Schreier). Any subgroup of a free group is still a free group.

Sketch of proof. Let F be a free group, and H a subgroup. Let G be a linear graph
s.t. π1 (G) ' F. One can check G is path-connected, locally path connected, and semi-
locally simply connected. Thus there exists covering space p : G Ü → G s.t. π (G,
Ü ee) ' H.
1
GÜ is a linear graph and π (G)
Ü is free. For details, see §83 and §84 of Munkres’ book. 
1
 APPLICATIONS OF FUNDAMENTAL GROUPS 7

¶ Use of van Kampen theorem: The fundamental groups of T2 again.

Next we compute the fundamental group of Σ1 = T2 . Although we already know
π1 (T2 ) = π1 (S 1 × S 1 ) ' Z2 , we want to re-compute its fundamental group via van
Kampen, because the computation shed a light on how to compute the fundamental
group of Σg (the closed oriented surface with g holes).
We first write T2 to a union of

U1 = T2 \ D f
and U2 = D,
f is a small disc which contains D. We first compute the
where D is a small disc and D
2
fundamental group π1 (T \ D). According to the following picture,

we have
π1 (U1 ) ' π1 (S 1 ∨ S 1 ) ' Z ∗ Z = ha, bi.
On the other hand, obvious U2 is contractible, and U1 ∩ U2 is a “thin annulus” which
is homotopy equivalent to a circle, so

π1 (U2 ) ' π1 (“pt”) = {e} and π1 (U1 ∩ U2 ) ' π(S 1 ) ' Z.

Unfortunately the above information is still not enough

to determine π1 (T2 ), since we still need the map
ϕ = ι∗ : π1 (U1 ∩ U2 ) → π1 (U1 )
This group homotopy is induced by the inclusion map.
According to the picture, the generator of π1 (U1 ∩ U2 ) is
a circle which can be deformed inside U1 to the boundary
loop aba−1 b−1 . In other words,
ϕ(1) = aba−1 b−1
So we conclude

π1 (T2 ) ' (Z ∗ Z) ∗Z {e} = ha, b |aba−1 b−1 = 1i = ha, b | ab = bai ' Z ⊕ Z.

¶ Use of van Kampen theorem: The fundamental groups of Σg = T2 # · · · #T2 .

We start with Σ2 = T2 #T2 , which can be decomposed as below:
8 APPLICATIONS OF FUNDAMENTAL GROUPS

Note that in computing π1 (T2 ), we have already computed the fundamental group of
U1 and U2 . So we get
π1 (U1 ) ' ha1 , b1 i, π1 (U2 ) ' ha2 , b2 i, π1 (U1 ∩ U2 ) ' Z.
Moreover, we also explained just now that
ϕ(1) = a1 b1 a−1 −1
1 b1 , ψ(1) = b2 a2 b−1 −1
2 a2 .

It follows from van Kampen’s theorem that

π1 (T2 #T2 ) ' ha1 , b1 , a2 , b2 | a1 b1 a−1 −1 −1 −1
1 b1 a2 b2 a2 b2 = 1i
| {z }
ϕ(1)ψ(1)−1

Note that this is no longer abelian.

In general by induction one can compute the fundamental group of Σg = T2 # · · · #T2 ,
| {z }
g

and the result is

π1 (T2 # · · · #T2 ) ' ha1 , b1 , · · · , ag , bg | a1 b1 a−1 −1 −1 −1
1 b1 · · · ag bg ag bg = 1i
| {z }
g

¶ Application to recreational mathematics: Picture hanging.

Here is a picture hanging puzzle proposed by A. Spivak in 1997:
How to hang a picture with several nails, such that none of the mails is
removable, i.e. such that the picture will fall if any nail is removed?
Here are solutions for two or three nails:
 APPLICATIONS OF FUNDAMENTAL GROUPS 9

In general the solution is not unique. So, what kind of twist is a solution? We may
think of this problem via the fundamental group: For simplicity consider the two nail
problem. Let γ : [0, 1] → R2 \ {p, q} be a loop representing a solution. Then
[γ]p 6= e ∈ π1 (R2 \ {p, q})
since the picture is hanged on the two nails without falling down. Since removing either
nail will resulting a falling of picture, we get
[(ip )∗ γ]p = e ∈ π1 (R2 \ {q}), [(iq )∗ γ]p = e ∈ π1 (R2 \ {p}),
where ip : R2 \ {p, q} ,→ R2 \ {p} and iq : R2 \ {p, q} ,→ R2 \ {q} are the inclusion maps.
If we denote π1 (R2 \ {q}) ' Z = hai, π1 (R2 \ {p}) ' Z = hbi, then a solution to the
two nail picture hanging puzzle is
[γ]p = aba−1 b−1 ,
which is realized by the first picture. you can construct a loop [γ]p = aba−1 b−1 aba−1 b−1 .
You can also generalize this to more nails, e.g. for the four nail problem you may take
the loop to be
aba−1 b−1 cbab−1 a−1 c−1 dcaba−1 b−1 c−1 bab−1 a−1 d−1 = [[[a, b], c], d]
or
aba−1 b−1 cdc−1 d−1 bab−1 a−1 dcd−1 c−1 = [[a, b], [c, d]].
For more discussions and generalizations of this puzzle, c.f.
[Link]

¶ Application to recreational mathematics: the plate trick.

Here is the description of the plate trick:
Resting a small plate flat on the palm, it is possible to perform two
rotations of one’s hand while keeping the plate upright. After the first
rotation of the hand, the arm will be twisted, but after the second
rotation it will end in the original position.
The mathematics behind the plate trick: π1 (SO(3)) has an order 2 element!
To explain, let’s attach the standard orthogonal basis for R3 to each point along
the arm. Now construct two different loops in SO(3):
• the loop Q traced out by the fingertips over one rotation,
• the loop P traced out along the arm at the end of the first rotation.
We put them in the following picture, where the t parameter represents “time”, so
that at t = 1 you finished one rotation (and your arm is twisted), and at t = 2
10 APPLICATIONS OF FUNDAMENTAL GROUPS

you arm ends with the original position; and the s parameter represents your ar-
m, so that s = 0 represents your shoulder and s = 1 represents your fingertips.

From the first square (from bottome+right to left+top in the square A in the
picture) we see
[P ]p = [Q]p .
From the two squares together (from top to bottom) we get
[Q]p + [Q]p = e.
Since at the end of one rotation, your arm is twisted and you can’t untwist your arm
while fixing your fingertips, we have
[P ]p 6= e.
So we conclude that the path homotopy class [P ]p (= [Q]p ) is an order two element in
π1 (SO(3))!