0

In most textbooks a subgroup $H$ of group $G$ is defined as normal if $$i) \quad gHg^{-1} \subset H \quad \forall g \in G.$$ I understand that they don't write $gHg^{-1} = H$ as the definition of normality because we can deduce from condition "i" that $H \subset gHg^{-1}$. My question is if $H$ is a subgroup of $G$ is it true in general that $H \subset gHg^{-1}$?

  • 2
    Your question is unclear. State clearly what you are assuming, what you are not assuming, and what you are trying to prove, please. – Arturo Magidin Dec 07 '23 at 04:26
  • No, of course it is not true in general. $H$ is not normal in $G$ if and only if there exist a $g\in G$ such that $H\not\subset gHg^{-1}$. Equivalently, $H$ is normal if and only if for all $g\in G$, $H\subseteq gHg^{-1}$, if and only if for all $g\in G$, $gHg^{-1} \subseteq H$, if and only if for all $g\in G$, $gHg^{-1}=H$. – Arturo Magidin Dec 07 '23 at 04:38
  • This is an answer to the question, not just a comment. Please post your answer as an answer. This brings extra visibility to the answer und puts the question off the unanswered list (otherwise it will be bumped up periodically). Also notice that comments are not indexed by the full text search, don't have any revision history, can only be edited for 5 minutes and allow only limited markup. Comments should only be used to clarify, not answer the problem. See How do comments work for more information. – Martin Brandenburg Dec 07 '23 at 04:54
  • In case it is a duplicate, answers are not wanted, and the rule doesn't change when the answer is hidden inside of a comment. – Martin Brandenburg Dec 07 '23 at 04:56

3 Answers3

1

No, but before I provide a counterexample, note that the map $\gamma_g=a\mapsto gag^{-1}$ is a bijection at least, since it has an inverse in $\gamma_{g^{-1}}=a\mapsto g^{-1}ag$. If $H$ is finite, this would trivially imply that $H=gHg^{-1}$ and $H$ would therefore be normal. I do not know if you have considered that not every subgroup of a finite group is normal. But here is an example, in case not.

In $S_4$, we may consider the cyclic subgroup of order $2$, $\langle(12)(34)\rangle$. The only non-identity element is $(12)(34)$ and if we conjugate this element by $(123)$ we get $$(123)*(12)(34)*(132)=(13)(24)$$ $\langle(13)(24)\rangle$ also has order $2$, and in particular, does not contain $(12)(34)$ as an element, so $$\langle(12)(34)\rangle\not\subseteq(123)\langle(12)(34)\rangle(123)^{-1}$$

ZKe
  • 800
1

Let $G$ be a group, $H$ a subgroup.

  1. For a fixed $g\in G$, $gHg^{-1}\subseteq H\iff H\subseteq g^{-1}Hg$.

Indeed, from $gHg^{-1}\subseteq H$, multiply by $g^{-1}$ on the left and $g$ on the right to get $H\subseteq g^{-1}Hg$; from $H\subseteq g^{-1}Hg$, multiply on the left by $g$ and the right by $g^{-1}$ to get $gHg^{-1}\subseteq H$.

  1. For all $g\in G$, $gHg^{-1}\subseteq H \iff$ for all $g\in G$, $H\subseteq gHg^{-1} \iff$ for all $g\in G$, $gHg^{-1}=H$.

If $gHg^{-1}\subseteq H$ for all $g\in G$, then it holds for $g^{-1}$: $g^{-1}H(g^{-1})^{-1}=g^{-1}Hg\subseteq H$. That implies $H\subseteq gHg^{-1}$, by 1. Similarly, if $H\subseteq gHg^{-1}$ holds for all $g$, then applying 1 to $H\subseteq g^{-1}Hg$ gives $gHg^{-1}\subseteq H$. The third equivalence is then clear.

  1. If $H$ is finite, then for a fixed $g$, $gHg^{-1}\subseteq H \iff H\subseteq gHg^{-1}\iff gHg^{-1}=H$. For infinite $H$, this need not hold.

From $gHg^{-1}\subseteq H$, by 1 we get $H\subseteq g^{-1}Hg$. If $H$ is finite, then since $g^{-1}Hg$ has the same number of elements as $H$, we therefore have $H=g^{-1}Hg$. This gives $g^{-1}Hg\subseteq H$, and applying 1 again we get $H\subseteq gHg^{-1}$. The converse follows similarly.

For $H$ infinite, an example where $gHg^{-1}\subseteq H$ but $H\not\subseteq gHg^{-1}$ can be found here.

If we define "a subgroup $H$ is normal in $G$" as "for all $g\in G$, $gHg^{-1}=H$, we get the following equivalences, which are left as an exercise for the reader:

Theorem. Let $G$ be a group, and let $H$ be a subgroup of $G$. Assume $X\subseteq G$ and $Y\subseteq H$ are subgroups such that $\langle X\rangle = G$ and $\langle Y\rangle = H$. The following are equivalent

  1. $H\triangleleft G$.
  2. For all $g\in G$, $gHg^{-1}\subseteq H$.
  3. For all $g\in G$, $H\subseteq gHg^{-1}$.
  4. For all $g\in G$, for all $h\in H$ $ghg^{-1}\in H$.
  5. For all $g\in G$, for all $y\in Y$, $gyg^{-1}\in H$.
  6. For all $x\in X$, $xHx^{-1}=H$.
  7. For all $x\in X$, $xHx^{-1}\subseteq H$ and $x^{-1}Hx\subseteq H$.
  8. For all $x\in X$, for all $y\in Y$, $xyx^{-1}\in H$ and $x^{-1}yx\in H$.
Arturo Magidin
  • 398,050
0

We have the following two conditions:

  1. $g H g^{-1} ⊆ H$ for every $g ∈ G$,

  2. $H ⊆ g H g^{-1}$ for every $g ∈ G$.

You have already stated that condition 2 can be deduced from condition 1. But in the same way, condition 1 can also be deduced from condition 2! In other words, conditions 1 and 2 are actually equivalent. You are therefore asking if every subgroup is normal, which is not the case.