2024-05-11
Algebra-I
00

Contents

§11 Short Exact Sequences and Semidirect Products
§11.1 Extensions—Short Exact Sequences
§11.2 Split Short Exact Sequences
§11.3 Semidirect Products

§11 Short Exact Sequences and Semidirect Products

§11.1 Extensions—Short Exact Sequences

Definition 11.1 A short exact sequence of groups is a sequence consisting of two homomorphisms

GHKG\to H\to K

satisfying the following conditions:

(1) GHG\to H is injective,

(2) HKH\to K is surjective, and

(3) the kernel of HKH\to K is equal to—not merely isomorphic to—the image of GHG\to H.

A short exact sequence is usually written as

1GHK1.1\to G\to H\to K\to 1.

Definition 11.2 We also say that HH is an extension of KK by GG.

Question: What do the 11's at the two ends mean?

The symbol 11 denotes the trivial group with one element.

The sequence above is “exact” because the image of each homomorphism is the kernel of the next homomorphism.

For example, the portion

1GH1\to G\to H

means that the image of

1G1\to G

is the kernel of

GH,G\to H,

which means that

GHG\to H

is injective.

For reasons that will become clearer later, short exact sequences are important because of the following idea:

we think of the group HH as being built out of the groups GG and KK.

Example 11.1 We have the following short exact sequences:

(1)

Z/2ZZ/4ZZ/2Z,\mathbb{Z}/2\mathbb{Z} \to \mathbb{Z}/4\mathbb{Z} \to \mathbb{Z}/2\mathbb{Z},

where the first homomorphism sends

12Z/4Z,1\mapsto2\in\mathbb{Z}/4\mathbb{Z},

and the second homomorphism sends

0,200,2\mapsto0

and

1,30Z/2Z.1,3\mapsto0\in\mathbb{Z}/2\mathbb{Z}.

(2)

Z/2ZZ/2Z×Z/2ZZ/2Z.\mathbb{Z}/2\mathbb{Z} \to \mathbb{Z}/2\mathbb{Z}\times\mathbb{Z}/2\mathbb{Z} \to \mathbb{Z}/2\mathbb{Z}.

We take

a(a,0)a\mapsto(a,0)

and

(a,b)b.(a,b)\mapsto b.

Thus we may think of both

Z/4Z\mathbb{Z}/4\mathbb{Z}

and the Klein four group as being built from two copies of

Z/2Z.\mathbb{Z}/2\mathbb{Z}.

But we see that different groups can be constructed from Z/2Z\mathbb{Z}/2\mathbb{Z} in different ways.

We also have the following short exact sequences:

(3)

Z/3ZS3Z/2Z,\mathbb{Z}/3\mathbb{Z} \to S_3 \to \mathbb{Z}/2\mathbb{Z},

where

A3Z/3Z,A_3\cong\mathbb{Z}/3\mathbb{Z},

and this is the short exact sequence associated with the inclusion

A3S3.A_3\to S_3.

and

(4)

Z/3ZZ/6ZZ/2Z.\mathbb{Z}/3\mathbb{Z} \to \mathbb{Z}/6\mathbb{Z} \to \mathbb{Z}/2\mathbb{Z}.

Thus we see that there are at least two different ways to construct a group of order 66 from

Z/3Z\mathbb{Z}/3\mathbb{Z}

and

Z/2Z.\mathbb{Z}/2\mathbb{Z}.

Remark The examples above show that:

(1) an extension need not be a direct product;

(2) an extension

GHKG\to H\to K

need not admit a map

HKH\leftarrow K

such that

KHK=idK;K\to H\to K = \mathrm{id}_K;

(3) an extension of Abelian groups can be non-Abelian.

§11.2 Split Short Exact Sequences

Definition 11.3 If there exists a homomorphism

KHK\to H

such that

KHK=idK,K\to H\to K = \mathrm{id}_K,

we say that the short exact sequence

GHKG\to H\to K

splits.

For readability, from now on we will write a short exact sequence

GHKG\to H\to K

as

LHR,L\to H\to R,

where LL stands for “left” and RR stands for “right”.

Since

LHL\to H

is injective, from now on we will identify LL with its image in HH in order to simplify notation.

Example 11.2 Among the short exact sequences above:

(1)

Z/2ZZ/4ZZ/2Z,\mathbb{Z}/2\mathbb{Z} \to \mathbb{Z}/4\mathbb{Z} \to \mathbb{Z}/2\mathbb{Z},

(2)

Z/2ZZ/2Z×Z/2ZZ/2Z,\mathbb{Z}/2\mathbb{Z} \to \mathbb{Z}/2\mathbb{Z}\times\mathbb{Z}/2\mathbb{Z} \to \mathbb{Z}/2\mathbb{Z},

(3)

Z/3ZS3Z/2Z,\mathbb{Z}/3\mathbb{Z} \to S_3 \to \mathbb{Z}/2\mathbb{Z},

(4)

Z/3ZZ/6ZZ/2Z.\mathbb{Z}/3\mathbb{Z} \to \mathbb{Z}/6\mathbb{Z} \to \mathbb{Z}/2\mathbb{Z}.

Only (1) does not split.

Notice that, a priori, there is no reason to regard RR as a subgroup of HH.

For example, in

Z/2ZZ/4ZZ/2Z,\mathbb{Z}/2\mathbb{Z} \to \mathbb{Z}/4\mathbb{Z} \to \mathbb{Z}/2\mathbb{Z},

the second copy of Z/2Z\mathbb{Z}/2\mathbb{Z} cannot naturally be “embedded” back into Z/4Z\mathbb{Z}/4\mathbb{Z}.

Proposition 11.1

Z/2ZZ/4ZZ/2Z\mathbb{Z}/2\mathbb{Z} \to \mathbb{Z}/4\mathbb{Z} \to \mathbb{Z}/2\mathbb{Z}

does not split.

Proof: Z/2Z\mathbb{Z}/2\mathbb{Z} contains only elements of order 11 and 22.

Therefore, no homomorphism

j:Z/2ZZ/4Zj:\mathbb{Z}/2\mathbb{Z} \to \mathbb{Z}/4\mathbb{Z}

can have an image containing an element of order 3\geqslant3.

Indeed, if

gn=1,g^n=1,

then we must also have

j(g)n=1.j(g)^n=1.

But in Z/4Z\mathbb{Z}/4\mathbb{Z}, both [1][1] and [3][3] have order 44:

[1]={[1],[2],[3],[0]},[3]={[3],[6]=[2],[5]=[1],[0]}.\langle[1]\rangle = \{[1],[2],[3],[0]\}, \quad \langle[3]\rangle = \{[3],[6]=[2],[5]=[1],[0]\}.

Therefore the image of any homomorphism

j:Z/2ZZ/4Zj:\mathbb{Z}/2\mathbb{Z} \to \mathbb{Z}/4\mathbb{Z}

must lie inside

{[0],[2]}Z/4Z.\{[0],[2]\} \subset \mathbb{Z}/4\mathbb{Z}.

But this is the kernel of the map

Z/4ZZ/2Z.\mathbb{Z}/4\mathbb{Z} \to \mathbb{Z}/2\mathbb{Z}.

Therefore no jj can split the identity map on

R=Z/2Z.R=\mathbb{Z}/2\mathbb{Z}.
 ~\tag*{$\square$}

Here is a more dramatic example.

Example 11.3 The short exact sequence

Z×nZZ/nZ\mathbb{Z} \xrightarrow{\times n} \mathbb{Z} \to \mathbb{Z}/n\mathbb{Z}

does not split for any

n1,0,1.n\neq-1,0,1.

Any homomorphism from Z/nZ\mathbb{Z}/n\mathbb{Z} to Z\mathbb{Z} would have to send an element of order nn to an element of finite order.

But Z\mathbb{Z} has no nonzero elements of finite order.

Therefore there is no injection

Z/nZZ.\mathbb{Z}/n\mathbb{Z} \to \mathbb{Z}.

Since this is our first attempt to understand short exact sequences, let us analyze the case in which RR can be regarded as a subgroup of HH.

If both LL and RR live inside HH, then perhaps it is easier to accept the idea that HH is “constructed” from LL and RR.

This leads us to the following definition.

Definition 11.4 If there exists a group homomorphism

j:RHj:R\to H

such that the composite

RjHRR\xrightarrow{j}H\to R

equals

idR,\mathrm{id}_R,

then we say that the short exact sequence splits.

A choice of

j:RHj:R\to H

is called a splitting.

Thus, if the short exact sequence is given by homomorphisms

ϕ:LH,ψ:HR,\phi:L\to H, \qquad \psi:H\to R,

then the definition means

ψj=idR.\psi\circ j = \mathrm{id}_R.

In particular, jj is injective.

§11.3 Semidirect Products

In the example above, it is clear that

Z/nZ\mathbb{Z}/n\mathbb{Z}

cannot be regarded as a subgroup of Z\mathbb{Z}.

So we arrive at a new idea.

We would like to be able to recognize semidirect products naturally, and we would also like to produce examples.

Let us analyze the situation.

As above, in a split short exact sequence, identify RR with j(R)j(R).

Then every element of RR defines an action on HH by conjugation:

hrhr1.h\mapsto rhr^{-1}.

Since LL is normal,

rLr1=L,rLr^{-1}=L,

so this defines an action on LL:

Cr:lrlr1.C_r:l\mapsto rlr^{-1}.

Moreover, this is a group isomorphism from LL to itself.

We can prove that this defines a group homomorphism

RAut(L),R\to\mathrm{Aut}(L),

given by

rCr.r\mapsto C_r.

Equivalently, it is enough to prove

CrCr=Crr.C_r\circ C_{r^{\prime}} = C_{rr^{\prime}}.

Indeed, for hRh\in R,

(CrCr)(h)=Cr(Cr(h))=Cr(rhr1)=r(rhr1)r1.(C_r\circ C_{r^{\prime}})(h) = C_r(C_{r^{\prime}}(h)) = C_r(r^{\prime}hr^{\prime-1}) = r(r^{\prime}hr^{\prime-1})r^{-1}.

Using associativity of the group operation,

(CrCr)(h)=(rr)h(rr)1=Crr(h).(C_r\circ C_{r^{\prime}})(h) = (rr^{\prime})h(rr^{\prime})^{-1} = C_{rr^{\prime}}(h).

This new map

RAut(L)R\to\mathrm{Aut}(L)

is called the conjugation action of RR on LL.

Thus every splitting gives rise to a group homomorphism

RAut(L).R\to\mathrm{Aut}(L).

Here Aut(L)\mathrm{Aut}(L) means the automorphism group of LL as a group, not merely as a set.

Question: Fix two groups RR and LL.

A natural question is:

does every group homomorphism

RAut(L)R\to\mathrm{Aut}(L)

produce a split exact sequence?

Another observation is that, given a splitting, both RR and LL become subgroups of HH.

Moreover, their intersection contains only

1H.1_H.

Indeed, if a nonidentity element

lLRl\in L\cap R

existed, then the map

RHRR\to H\to R

could not be injective.

Finally, since the orbits of the action of LL cover all of HH, we obtain

H=rRLr.H = \bigcup_{r\in R}Lr.

That is,

H=LR.H=LR.

Definition 11.5 Let LL and RR be subgroups of HH. We define

LR={g such that g=lr for some lL, rR}.LR = \{g~\text{such that}~g=lr ~\text{for some}~ l\in L,\ r\in R\}.

To prove that

H=LRH=LR

when the short exact sequence

LHRL\to H\to R

splits, we need the following lemma.

Lemma 11.2 Let

LHψRL\to H\xrightarrow{\psi}R

be a short exact sequence.

Let

q:HH/Lq:H\to H/L

be the quotient homomorphism

hLh.h\mapsto Lh.

Then there exists an isomorphism

z:H/LRz:H/L\to R

such that

zq=ψ.z\circ q=\psi.

That is, there exists zz such that the diagram

commutes.

Proof: We are given a short exact sequence

LHψR,L\to H\xrightarrow{\psi}R,

where

ψ:HR\psi:H\to R

is surjective and

ker(ψ)=L.\ker(\psi)=L.

We want to prove that there exists an isomorphism

z:H/LRz:H/L\to R

such that

zq=ψ,z\circ q=\psi,

where

q:HH/Lq:H\to H/L

is the quotient map.

Since ψ\psi is surjective and

ker(ψ)=L,\ker(\psi)=L,

the First Isomorphism Theorem gives

H/LR.H/L\cong R.

Define

z:H/LRz:H/L\to R

by

z(Lh)=ψ(h).z(Lh)=\psi(h).

This is well defined because if

Lh1=Lh2,Lh_1=Lh_2,

then

h1h21L=ker(ψ),h_1h_2^{-1}\in L=\ker(\psi),

so

ψ(h1)=ψ(h2).\psi(h_1)=\psi(h_2).

The map zz is an isomorphism because:

  • zz is a homomorphism since ψ\psi is a homomorphism;

  • zz is surjective since ψ\psi is surjective;

  • zz is injective because the kernel of zz is precisely LL, meaning that the only element mapped to the identity of RR is LL.

For every hHh\in H,

(zq)(h)=z(Lh)=ψ(h),(z\circ q)(h) = z(Lh) = \psi(h),

so

zq=ψ.z\circ q=\psi.

Therefore zz is the required isomorphism making the diagram commute.

By the First Isomorphism Theorem, there exists an isomorphism

z:H/LRz:H/L\to R

such that

zq=ψ,z\circ q=\psi,

which proves the lemma.

 ~\tag*{$\square$}

Once we have this lemma, we can prove the following corollary.

Corollary 11.3 If

j:RHj:R\to H

is a splitting of

LHR,L\to H\to R,

then

H=rRLj(r).H = \bigcup_{r\in R}Lj(r).

Proof: By the definition of a splitting,

ψj=idR.\psi\circ j = \mathrm{id}_R.

On the other hand, by the lemma,

ψ=zq.\psi=z\circ q.

Thus

zqj=idR.z\circ q\circ j = \mathrm{id}_R.

Since zz is a group isomorphism, its inverse is a homomorphism, and therefore

qj=z1.q\circ j = z^{-1}.

Now let us interpret

qj.q\circ j.

The homomorphism qq sends hh to

Lh.Lh.

Therefore,

qjq\circ j

sends rr to the coset

Lj(r)H/L.Lj(r)\in H/L.

Since

z1:RH/Lz^{-1}:R\to H/L

is a bijection, for every coset

LhH/LLh\in H/L

there exists a unique rRr\in R such that

Lh=Lj(r).Lh=Lj(r).

Since

H/LLh=H,\bigcup_{H/L}Lh = H,

this proves

rRLj(r)=H.\bigcup_{r\in R}Lj(r) = H.
 ~\tag*{$\square$}

In the discussion above, we used jj to identify rRr\in R with its image in HH, so we write

rRLr=H.\bigcup_{r\in R}Lr = H.

Question: Fix

L,RH.L,R\subset H.

If

LR={1},L\cap R=\{1\},

LHL\subset H is normal, and

LR=H,LR=H,

is HH the semidirect product of LL and RR?

This is exactly the right question, because the answer is yes.

Theorem 11.4 Fix a normal subgroup

LH,L\subset H,

and suppose

RH/L.R\cong H/L.

The following conditions are equivalent:

(1) A homomorphism

j:RHj:R\to H

splits the short exact sequence

LHR.L\to H\to R.

(2) An isomorphism

RRR\to R^{\prime}

onto a subgroup

RHR^{\prime}\subset H

such that

RL={1}R^{\prime}\cap L=\{1\}

and the map

L×RHL\times R^{\prime}\to H

is surjective.

(3) A group homomorphism

ϕ:RAut(L).\phi:R\to\mathrm{Aut}(L).

Among these descriptions, we like the last one best because it makes no reference to the group HH.

Once we construct a group homomorphism

ϕ:RAut(L),\phi:R\to\mathrm{Aut}(L),

we can construct a short exact sequence

LHR.L\to H\to R.

Question: What is the group operation on HH in terms of RR and LL?

Proposition 11.5 Any homomorphism

ϕ:RAut(L),rϕr\begin{aligned} \phi:R&\to\mathrm{Aut}(L),\\ r&\mapsto\phi_r \end{aligned}

defines a group HH and a split short exact sequence

such that:

(1) the following defines a group structure on the set

H=L×R:H=L\times R:
H×HH(l1,r1)(l2,r2):=(l1ϕr1(l2),r1r2).\begin{aligned} H\times H&\to H\\ (l_1,r_1)\cdot(l_2,r_2) &:= (l_1\cdot\phi_{r_1}(l_2),r_1r_2). \end{aligned}

The right-hand side is almost the usual group operation on L×RL\times R, except that before multiplying l1l_1 and l2l_2, we “twist” l2l_2 into another element of LL, namely

ϕr1(l2),\phi_{r_1}(l_2),

the value of r1r_1 under the homomorphism

ϕ:RAut(L).\phi:R\to\mathrm{Aut}(L).

Moreover,

(2) the set

{(l,1)}\{(l,1)\}

is a normal subgroup isomorphic to LL;

(3) the set

{(1,r)}\{(1,r)\}

is a subgroup isomorphic to RR.

Proof: Associativity:

(l1,r1)((l2,r2)(l3,r3))=(l1,r1)(l2ϕr2(l3),r2r3)=(l1ϕr1(l2ϕr2(l3)),r1(r2r3))=(l1ϕr1(l2)ϕr1(ϕr2(l3)),r1(r2r3))=(l1ϕr1(l2)ϕr1r2(l3),(r1r2)r3)=(l1ϕr1(l2),r1r2)(l3,r3)=((l1,r1)(l2,r2))(l3,r3).\begin{aligned} (l_1,r_1)\cdot((l_2,r_2)\cdot(l_3,r_3)) &= (l_1,r_1)\cdot (l_2\cdot\phi_{r_2}(l_3),r_2r_3)\\ &= (l_1\cdot \phi_{r_1}(l_2\cdot\phi_{r_2}(l_3)), r_1(r_2r_3))\\ &= (l_1\cdot \phi_{r_1}(l_2)\cdot \phi_{r_1}(\phi_{r_2}(l_3)), r_1(r_2r_3))\\ &= (l_1\cdot \phi_{r_1}(l_2)\cdot \phi_{r_1r_2}(l_3), (r_1r_2)r_3)\\ &= (l_1\cdot\phi_{r_1}(l_2),r_1r_2) \cdot(l_3,r_3)\\ &= ((l_1,r_1)\cdot(l_2,r_2)) \cdot(l_3,r_3). \end{aligned}

The third equality holds because

ϕr1\phi_{r_1}

is a homomorphism.

The fourth equality holds because

ϕ:RAut(L)\phi:R\to\mathrm{Aut}(L)

is a homomorphism.

Identity:

(1L,1R)(l,r)=(1Lϕ1R(l),1Rr)=(1Ll,1Rr)=(l,r).\begin{aligned} (1_L,1_R)\cdot(l,r) &= (1_L\cdot\phi_{1_R}(l),1_R\cdot r)\\ &= (1_L\cdot l,1_R\cdot r)\\ &= (l,r). \end{aligned}

The second equality follows because ϕ\phi is a homomorphism, so

ϕ1=1.\phi_1=1.

Inverse:

Proposition 11.6

(l,r)1=(ϕr1(l1),r1).(l,r)^{-1} = (\phi_{r^{-1}}(l^{-1}),r^{-1}).

Proof:

(l,r)(ϕr1(l1),r1)=(lϕr(ϕr1(l1)),rr1)=(lϕrr1(l1),rr1)=(ll1,rr1)=(1L,1R).\begin{aligned} (l,r)\cdot (\phi_{r^{-1}}(l^{-1}),r^{-1}) &= (l\cdot \phi_r(\phi_{r^{-1}}(l^{-1})), rr^{-1})\\ &= (l\cdot \phi_{rr^{-1}}(l^{-1}), rr^{-1})\\ &= (l\cdot l^{-1}, rr^{-1})\\ &= (1_L,1_R). \end{aligned}

The second equality follows because

ϕ:RAut(L)\phi:R\to\mathrm{Aut}(L)

is a homomorphism, so

ϕrϕr=ϕrr.\phi_r\circ\phi_{r^{\prime}} = \phi_{rr^{\prime}}.

The third equality follows because ϕ\phi is a group homomorphism, so

ϕ1=idL.\phi_1 = \mathrm{id}_L.

Also,

(ϕr1(l1),r1)(l,r)=(ϕr1(l1)ϕr1(l),r1r)=(ϕr1(l1l),r1r)=(ϕr1(1L),1R)=(1L,1R).\begin{aligned} (\phi_{r^{-1}}(l^{-1}),r^{-1}) \cdot(l,r) &= (\phi_{r^{-1}}(l^{-1}) \cdot \phi_{r^{-1}}(l), r^{-1}r)\\ &= (\phi_{r^{-1}}(l^{-1}l), r^{-1}r)\\ &= (\phi_{r^{-1}}(1_L), 1_R)\\ &= (1_L,1_R). \end{aligned}

The second equality holds because

ϕr1:LL\phi_{r^{-1}}:L\to L

is a group homomorphism.

The final equality holds because ϕr1\phi_{r^{-1}} is a group homomorphism.

 ~\tag*{$\square$}

Therefore, this really does define a group.

 ~\tag*{$\square$}

Definition 11.6 We denote this group by

LϕR.L\rtimes_{\phi}R.

When ϕ\phi is understood, we write

LR,L\rtimes R,

and call LRL\rtimes R a semidirect product of LL and RR.

Remark We say “a” semidirect product because different choices of ϕ\phi may produce different groups.

Although the word “and” is usually insensitive to order, LL and RR play very different roles here.

Remark Why do we use the symbol

?\rtimes?

Usually, when NN is a normal subgroup of GG, we write

NG.N\triangleleft G.

The symbol

\rtimes

is a hybrid of

\triangleleft

for “normal subgroup” and

×\times

for “product”.

Any diagram

gives rise to a map

RAut(L).R\to\mathrm{Aut}(L).

How?

By conjugation, since

LH is normal,Ch:LLlhlh1\begin{aligned} L\subset H &~\text{is normal},\\ C_h:L&\to L\\ l&\mapsto hlh^{-1} \end{aligned}

is a group automorphism of LL.

By essentially the same argument as in the exercise, we obtain a homomorphism

HAut(L)hCh.\begin{aligned} H&\to\mathrm{Aut}(L)\\ h&\mapsto C_h. \end{aligned}

The composite

is exactly the homomorphism ϕ\phi.

Question: How does RR act on

LLR?L\subset L\rtimes R?

Proposition 11.7

(1L,r)(l,1R)(1L,r1)=(ϕr(l),1R).(1_L,r)\cdot(l,1_R)\cdot(1_L,r^{-1}) = (\phi_r(l),1_R).

In other words, conjugation by rr in

LRL\rtimes R

recovers ϕr\phi_r.

Proof:

(1L,r)(l,1R)(1L,r1)=(1Lϕr(l),r1R)(1L,r1)=(ϕr(l)ϕ(1L),rr1)=(ϕr(l1L),1R)=(ϕr(l),1R).\begin{aligned} (1_L,r)\cdot(l,1_R)\cdot(1_L,r^{-1}) &= (1_L\cdot\phi_r(l),r\cdot1_R) \cdot(1_L,r^{-1})\\ &= (\phi_r(l)\cdot\phi(1_L),rr^{-1})\\ &= (\phi_r(l\cdot1_L),1_R)\\ &= (\phi_r(l),1_R). \end{aligned}
 ~\tag*{$\square$}

We now have enough ingredients to prove Theorem 11.4.

The key ideas are:

  • If
LHL\subset H

is normal, then

HLH\curvearrowright L

by conjugation.

  • Given a splitting

then RR also acts by conjugation:

RϕAut(L).R\xrightarrow{\phi}\mathrm{Aut}(L).
  • LRL\rtimes R is a group in which the conjugation action of RR on LL agrees with ϕ\phi.

Now let us prove

HLR.H\cong L\rtimes R.

To see why

LRH,L\rtimes R\cong H,

we need a lemma.

Lemma 11.8 For every

hH,h\in H,

there exist unique

lLl\in L

and

rRr\in R

such that

h=lj(r).h=l\cdot j(r).

Proof: We know that the diagram

commutes.

That is,

zq=ψ.z\circ q=\psi.

Given a splitting

HjR,H\xleftarrow{j}R,

we have

zqj=ψj=idR.z\circ q\circ j = \psi\circ j = \mathrm{id}_R.

Since zz is an isomorphism, it has an inverse

z1,z^{-1},

and therefore

qj=z1.q\circ j = z^{-1}.

The map z1z^{-1} is also a group isomorphism.

Since z1z^{-1} is a bijection, for every hHh\in H there exists a unique rr such that

qj(r)=[h]H/L.q\circ j(r) = [h] \in H/L.

Hence there exists a unique rRr\in R such that

j(r)Lh=[h].j(r)\in Lh=[h].

Equivalently, there exists a unique rRr\in R such that

[j(r)]=[h].[j(r)]=[h].

Thus there exists a unique rRr\in R such that

h=lj(r)h=l\cdot j(r)

for some lLl\in L.

Of course, once hh and j(r)j(r) are fixed, ll is uniquely determined:

l=hj(r)1.l = h\cdot j(r)^{-1}.

Therefore, for every hHh\in H, there exist unique l,rl,r such that

h=lj(r).h=l\cdot j(r).
 ~\tag*{$\square$}

Now we can prove the following.

Theorem 11.9 Let

be a split short exact sequence, and let

ϕ:RAut(L)\phi:R\to\mathrm{Aut}(L)

be the induced action.

Then

HLϕR.H\cong L\rtimes_{\phi}R.

Proof: Consider the map

LϕRαH(l,r)lj(r).\begin{aligned} L\rtimes_{\phi}R &\xrightarrow{\alpha} H\\ (l,r) &\mapsto l\cdot j(r). \end{aligned}

Then

(l1ϕr1(l2),r1r2)l1ϕr1(l2)j(r1)j(r2).(l_1\cdot\phi_{r_1}(l_2),r_1r_2) \mapsto l_1\phi_{r_1}(l_2)j(r_1)j(r_2).

But by definition,

ϕr1(l2)=j(r1)l2j(r1)1.\phi_{r_1}(l_2) = j(r_1)l_2j(r_1)^{-1}.

Therefore,

α((l1,r1)(l2,r2))=α((l1ϕr1(l2),r1r2))=l1ϕr1(l2)j(r1)j(r2)=l1j(r1)l2j(r1)1j(r1)j(r2)=l1j(r1)l2j(r2)=α((l1,r1))α((l1,r2)).\begin{aligned} \alpha((l_1,r_1)\cdot(l_2,r_2)) &= \alpha((l_1\phi_{r_1}(l_2),r_1r_2))\\ &= l_1\phi_{r_1}(l_2)j(r_1)j(r_2)\\ &= l_1j(r_1)l_2j(r_1)^{-1}j(r_1)j(r_2)\\ &= l_1j(r_1)l_2j(r_2)\\ &= \alpha((l_1,r_1)) \cdot \alpha((l_1,r_2)). \end{aligned}

Thus α\alpha is a homomorphism.

By Lemma 11.8, for every hHh\in H there exist unique lLl\in L and rRr\in R such that

h=lj(r).h=l\cdot j(r).

Therefore α\alpha is a bijection.

 ~\tag*{$\square$}

This is enough to show that a split short exact sequence and a semidirect product contain the same amount of data.

  • Given

we obtain

ϕ:RAut(L)\phi:R\to\mathrm{Aut}(L)

by conjugation.

  • By the theorem,
HαLRH\xleftarrow{\alpha}L\rtimes R

is an isomorphism.

  • Hence we have a surjection
LRαHψR.L\rtimes R \xrightarrow{\alpha} H \xrightarrow{\psi} R.
  • Since α\alpha is an isomorphism,
ker(ψα)=α1(kerψ)=α1(L)={(l,1R)}LR\begin{aligned} \ker(\psi\circ\alpha) &= \alpha^{-1}(\ker\psi)\\ &= \alpha^{-1}(L)\\ &= \{(l,1_R)\} \subset L\rtimes R \end{aligned}

by Lemma 11.8.

  • Therefore we obtain a short exact sequence
LLRψαRl(l,1R)\begin{aligned} L&\to L\rtimes R \xrightarrow{\psi\circ\alpha} R\\ l&\mapsto(l,1_R) \end{aligned}

with splitting

LRRL\rtimes R\leftarrow R

(1L,r)(1_{L},r)\mapsto r~r

The situation can be summarized by saying that the following diagram commutes:

That is, every sub-square that can be drawn is commutative.

Example 11.4 Recall that

SOn(R)On(R)SO_n(\mathbb{R}) \subset O_n(\mathbb{R})

is a subgroup of index 22.

By definition,

SOn(R)=ker(On(R)detR×).SO_n(\mathbb{R}) = \ker \left( O_n(\mathbb{R}) \xrightarrow{\det} \mathbb{R}^{\times} \right).

Every kernel is a normal subgroup, so

SOn(R)SO_n(\mathbb{R})

is normal.

Alternatively, one may use the fact that every subgroup of index 22 is normal.

Therefore we have a short exact sequence

Z/2Z\mathbb{Z}/2\mathbb{Z} \cong 1SOn(R)On(R)±11                                     1\to SO_{n}(\mathbb{R})\to O_{n}(\mathbb{R})\to{\pm 1}\to 1~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ \subset R×\mathbb{R}^{\times}

This sequence admits many different splittings.

For concreteness, take

n=2.n=2.

For example,

O2(R)Z/2Z:jO_2(\mathbb{R}) \leftarrow \mathbb{Z}/2\mathbb{Z}:j

(10 01)\begin{pmatrix} 1&0\ 0&1 \end{pmatrix}\mapsto[0][0]

(10 01)\begin{pmatrix} 1&0\ 0&-1 \end{pmatrix}\mapsto[1][1]

or

(10 01)\begin{pmatrix} -1&0\ 0&1 \end{pmatrix}\mapsto[1][1]

or

(01 10)\begin{pmatrix} 0&1\ 1&0 \end{pmatrix}\mapsto[1][1]