0

I have the following Problem:

let $G$ be a topological group. We denote by $G_0$ the path connected component of the identity and by $G_0'$ the connected component of the identity. I now want to show that $G_0$ is a normal subgroup of $G$

I wanted to proceed as follows:

Claim 1 $G_0$ is a subgroup of $G$. We need to show the following three points:

  1. (Contains the identity element) Indeed we see that by definition $id\in G_0$.
  2. (stable under composition) Now I wanted to take two path connected components $A,B\in G_0$. Then clearly $id\in A\cap B$ and thus the intersection is not empty. But then from the lecture we know that $A\cup B$ is again path connected and thus in $G_0$
  3. (Stable under inverses) here I have no idea because I don't know how an inverse of a connected subset looks like

Since my approach was wrong I add here a correct one, which was discussed below:

enter image description here

Afterwards I would check that it is normal but for the moment I would be happy if you could take a look at my approach and correct me because I'm really unsure about topological groups since we have never disscused them in class, the only appeared in one exercise sheet.

Thanks for your help.

user123234
  • 2,885
  • 1
    The elements of $G_0$ are not the path-connected components if $G$, they are the points in the path-connected component of the identity element $e_G\in G$! – Jonas Linssen Dec 31 '21 at 12:38
  • So but we had the definition: The union $\Pi(p)$ of all path connected subset of $M$ which contains $p$ is the path connected component of $p$. Is this wrong? – user123234 Dec 31 '21 at 12:42
  • Could you maybe help me to correct my thinking errors because I'm really lost in this part. I don't see where to start and with wich objects I need do work – user123234 Dec 31 '21 at 12:47
  • 1
    Okay, but is a union of subsets again a subset (yes!) or a set of subsets (no!). I hope my answer clarifies things a bit. – Jonas Linssen Dec 31 '21 at 12:57

2 Answers2

1

If $a,b\in G_0$ and $\gamma\colon [0,1]\to G$ is a path from $id$ to $a$, then $[0,1]\to G$, $t\mapsto \gamma(t)b$ is a path from $b$ to $ab$. Append this to a path from $id$ to $b$ to find a path from $id$ to $ab$. Also note that $[0,1]\to G$, $t\mapsto a^{-1}\gamma(1-t)$ is a path from $id $ to $a^{-1}$. Conclude that $G_0$ is a subgroup.

Use a similar simple argument for normality (or perhaps exhibit a homomorphism to a suitable group that has precisely $G_0$ as kernel - this could kill subgroup and normality in a single strike).

  • Thank you for your help. But could you maybe explain once again how $G_0$ looks like? Because somehow I missunderstood my definition I comment above. – user123234 Dec 31 '21 at 12:50
  • So Is $G_0$ the largest path connected subset of $G$ containing $id$ in our case? Thus elements in $G_0$ can always been connected with $id$ by a continuous path? – user123234 Dec 31 '21 at 12:54
  • @Wave Yes, that's what they said: $G_0$ is the path connected component of $G$ that contains $id$. E.g., in the multiplicative group of the non-zero real numbers, this would be the positive reals. – Hagen von Eitzen Dec 31 '21 at 12:57
  • ah okey thank you – user123234 Dec 31 '21 at 12:58
1

The path-connected component of a point $x$ in some space $X$ consists of all points $y\in X$ such that there exists a $xy$-path in $X$. All those $y$ define a subset and thus with the subspace topology a subspace.

A topological group is a topological space $G$ together with a chosen identity element $e_G \in G$ and two continuous maps $\cdot:G \times G \rightarrow G$ and $(-)^{-1}:G \cong G$ satisfying the usual group axioms (associativity of multiplication, unitality of $e_G$ and having inverses given by $(-)^{-1}$).

You are supposed to show that the connected component subspace of the element $e_G$ of a topological group $G$ again forms a topological group. $e_G$ is contained in its own connected component, because there is a constant $e_Ge_G$-path. It is not so obvious that given $x,y$ in the component $x\cdot y$ is in the component. There we need to recall that a $xy$-path is given by a continuous map $[0,1] \rightarrow X$ sending $0$ to $x$ and $1$ to $y$. Since both $x$ and $y$ are in the component of $e_G$ there are $xe_G$- and $ye_G$-paths $\alpha,\beta$. Now the multiplication is a continuous map $G\times G \rightarrow G$ and we can use the paths $\alpha,\beta$ to define a continuous map $$\begin{array}{ccccc} [0,1] & \overset{(\alpha,\beta)}{\longrightarrow} & G \times G &\overset{\cdot}{\longrightarrow} & G\\ t & \mapsto & (\alpha(t),\beta(t)) & \mapsto & \alpha(t)\cdot \beta(t)\\ 0 & \mapsto & (x,y) & \mapsto & x\cdot y\\ 1 & \mapsto & (e_G,e_G) & \mapsto & e_G \cdot e_G = e_G \end{array}$$ This defines a $(x\cdot y)e_G$-path proving that $x\cdot y$ is contained in the component of $e_G$.

I leave the proof of $x^{-1}$ being in the component up to you. The idea is precisely the same: use an $xe_G$-path and continuity of $(-)^{-1}$ to define an $x^{-1}e_G$-path.

Jonas Linssen
  • 11,016
  • Thanks a lot, so but to check that it is a topological subgroup I only need to check the three points in my original task (containing the identity, stable under composition and stable under inverses). And did I understood it correctly that the composition is the concatination of paths? – user123234 Dec 31 '21 at 12:58
  • Yes and no. To have a topological subgroup you only have to check the subgroup axioms since is understood that you want to have the subspace topology. In my argument no path composition is involved. I will edit. – Jonas Linssen Dec 31 '21 at 13:00
  • Could I send you here my proof of $x^{-1}$ so that you can have a look and I can be sure that it works? – user123234 Dec 31 '21 at 13:00
  • So but in algebra we had this three subgroup axioms, why are there less here in topology? – user123234 Dec 31 '21 at 13:01
  • Where do you see less? 1. $e_G \in G_0$ 2. $x\cdot y \in G_0$ 3. $x^{-1} \in G_0$… – Jonas Linssen Dec 31 '21 at 13:05
  • ah o sorry i didn't see the last part. Is there a explanation why we use the multiplicative structure here or is it just "normal" to write it in the multiplicative way unless we know more? – user123234 Dec 31 '21 at 13:07
  • It was the convention I am used to. Potentially nonabelian groups get a multiplication, abelian groups get an addition. One also often uses $m$ for multiplication, but I figured $\cdot$ would be less confusing to you. – Jonas Linssen Dec 31 '21 at 13:10
  • ah okey perfect. thanks a lot. So i will try to recap the whole thing and I will upload a hopefully correect solution for all the other members who are interested in this question. Maybe you could also tell me if there are still some mistakes, if you have time – user123234 Dec 31 '21 at 13:12
  • Sorry can I ask again somethin. While resolving it I don't get why your paths go from $[0,1]$ to $G$ and not to $G_0$. Is it because we don't know if they land in $G_0$? – user123234 Dec 31 '21 at 13:30
  • This is a good question and the answer is: it doesn’t matter. We want to show that some elements lie in the same component $G_0$ of the space $G$, so we shall consider paths in $G$. But on the other hand the component is maximal with respect to paths meaning that every path with some point in $G_0$ already lies completely in $G_0$. So you might as well use $G_0$ everywhere. – Jonas Linssen Dec 31 '21 at 13:36
  • Ah okey sure makes sense so we use here that $G_0$ is the largest subset of $G$ that is path connected and contains $id$. So all the paths need to live in $G_0$ because otherwise we don't have a path to an element in $G\setminus G_0$ right? – user123234 Dec 31 '21 at 13:40
  • 1
    Yup………………………….. – Jonas Linssen Dec 31 '21 at 13:41
  • In addition I uploaded my solution is this correct now? – user123234 Dec 31 '21 at 13:41
  • It looks good to me :) – Jonas Linssen Dec 31 '21 at 14:02
  • Perfect thank you very much for your huge help I appreciate that! – user123234 Dec 31 '21 at 14:07