Generationg of H
Generationg of H
Generationg of H
Generators for Hn (S n )
Last time we calculated the homology groups of S n and the pair (Dn ,S n−1 ). To
make this calculation a bit more concrete, let us try to figure out the generators
of the infinite cyclic groups Hn (Dn , S n−1 ) and Hn (S n ):
• On the standard n-simplex, there is a special n-chain Sn (∆n ), called the fun-
damental n-simplex, given by the identity map ιn : ∆n → ∆n . We observed
in a previous lecture that ιn is not a cycle, since its boundary ∂(ιn ) ∈ Sn−1 (∆n )
is the alternating sum of the faces of the n-simplex each of which is a generator
in Sn−1 (∆n ).
X
∂(ιn ) = (−1)i φni (∆n−1 ) 6= 0.
i
Constructing ϕn
1
For each ∆n the point c = (t0 , . . . ,tn ) with ti = n+1 for all i is the barycen-
n
ter of ∆ .
For every point x ∈ ∆n which is not c, there is a unique ray from c to x.
We denote the unique point where this ray hits ∂∆n by f (x). In particular,
if x ∈ ∂∆n , then f (x) = x.
A concrete generator of H1 (S 1 )
We just learned that the class [∂(ϕ2 )] is a generator of H1 (S 1 ). We can
describe this class as follows:
By definition, ∂(ϕ2 ) is the 1-cycle
∂(ϕ2 ) = d0 ϕ2 − d1 ϕ2 + d2 ϕ2
= ϕ2 ◦ φ20 − ϕ2 ◦ φ21 + ϕ2 ◦ φ22 .
Recall that ϕ2 maps ∂∆2 homeomorphically to S 1 . With this in mind, the
summands look like
1 2
ϕ2 ◦ φ20 (1 − t, t) = eiπ(− 6 +t 3 )
7 2
ϕ2 ◦ φ21 (1 − t, t) = eiπ( 6 −t 3 )
7 2
ϕ2 ◦ φ22 (1 − t, t) = eiπ( 6 +t 3 ) .
First applications
But by our calculation of the Hq (S n−1 ), such an isomorphism can only exist if
n − 1 = q = m − 1. This contradicts the assumption n 6= m. QED
We can also give a short proof of Brouwer’s famous Fixed-Point Theorem:
Before we prove the theorem, let us have a look at dimension one, where the
result is very familiar:
If g(0) = 0 or g(1) = 1, we are done. But if g(0) > 0 and g(1) < 1, then the
Intermediate Value Theorem implies that there is an x0 ∈ (0,1) with
g(x0 ) = 0, i.e. f (x0 ) = x0 .
g : Dn → ∂Dn .
6
Let i : S n−1 = ∂Dn ,→ Dn denote the inclusion map. Note that if x ∈ ∂Dn ,
then g(x) = x. In other words,
g ◦ i = idS n−1 .
Hn (i) ) Hn (g)
n
Hn−1 (D ) = 0.
Before we start the proof, let us have a look at what happens for the reflection
map
r : D1 = [−1,1] → [−1,1] = D1 , t 7→ −t
and its restriction to S 0 . Recall that S 0 constists of just two points, x = 1 and
y = −1 (on the real line R). The effect of r on S 0 is to interchage x and y.
8
Z / Z⊕Z
/ Z / 0.
9
We know that the left hand and central squares commute, since the inclusion
and the reflection commute. We know that the horizontal maps are isomorphisms
from the calculation of these groups.
Thus, knowing H0 (r) = −1 on Ker (), we see that H1 (r) is also multiplica-
tion by −1 on H1 (S 1 ).
Now we can proceed by induction: For n ≥ 2, we have again a commutative
diagram from the calculation of Hn (S n ):
∼ ∼ ∼
Hn (S n )
= / n o
Hn (S n , D+ )
= n
Hn (D− , S n−1 )
= / Hn−1 (S n−1 )
Hn (r) Hn (r) Hn (r) Hn−1 (r)
∼ ∼ ∼
Hn (S n )
= / n o
Hn (S n , D+ )
= n
Hn (D− , S n−1 )
= / Hn−1 (S n−1 ).
The right most square commutes by an exercise from last week. The left hand
and central squares commute, since the inclusion and the reflection commute.
Assuming the assertion for n − 1, i.e., Hn−1 (r) is multiplication by −1, we see
that Hn (r) is also multiplication by −1. QED