Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

et $G$ be an infinite group and $H,K \leq G$ with $[G:H] = m$ and $[G:K] = n$. I want to show that $[G:H \cap K] < \infty$.

My idea:

Let $\{h_1,\cdots,h_m\}$ be representants of $\{gH \mid g \in G\}$ and similar $\{k_1,\cdots,k_n\}$. Then each $g \in G$ is in $h_iH \cap k_jK$ for some $1 \leq i \leq m$ and $1 \leq j \leq n$. Then I want to show that if $g \in h_iH \cap k_jK$ then $g( H \cap K) = h_iH \cap k_jK$.

Does this show that $[G:H \cap K] <\infty$ ? I think that this shows that $[G: H \cap K] \leq mn$.

share|cite|improve this question
Show that $\,\{h_ik_j\}\,$ is a complete set representative of the cosets of $\,H\cap K\,$ in $\,G\,$ (though, most probably, not a minimal one!) – DonAntonio Jan 5 '13 at 14:32
You mean $\{h_i,k_j \mid 1 \leq i \leq n , 1 \leq j \leq m \}$ ? – Epsilon Jan 5 '13 at 14:35
No @Andre, I meant what I wrote. Read my answer below. – DonAntonio Jan 5 '13 at 16:47
up vote 3 down vote accepted

Let $H$ act on the set of left cosets of $K$ in $G$ by left multiplication. (This action is not necessarily transitive) The stabilizer of the coset $K$ is $H \cap K$. By the orbit-stabilizer theorem, $[H: H \cap K]$ is finite since $[G:K]$ is finite. Since $[G:H]$ is also finite the result follows.

share|cite|improve this answer
Why is the action transitive ? – Epsilon Jan 5 '13 at 14:56
Straightforward approach. I favorited it. +1 – Babak S. Jan 5 '13 at 15:02
Just to clarify it for myself: Let $A:= \{gK \mid g \in G\}$. Then $|A| = [G:K]< \infty$ and $[H : H \cap K] = [H: Stab(K)] = |Orb(K)| < \infty$ because $Orb(K) \subseteq A$. How to proceed then ? – Epsilon Jan 5 '13 at 15:35
Sure. Thanks ! Nice approach . – Epsilon Jan 5 '13 at 15:38

Let $A$ be the set of all right cosets of $H\cap K$ in $H$ and $B$ be the set of all right cosets of $K$ in $G$. Then the map $\varphi:A\to B$ given by $(H\cap K)h\mapsto Kh$ for $h\in H$ is well defined since $(H\cap K)h'=(H\cap K)h$ implies $h'h^{-1}\in H\cap K\subseteq K$, hence $Kh'=Kh$. To see that $\varphi$ is injective, suppose $Kh_1=Kh_2$. Then $h_1h_2^{-1}\in H\cap K$, hence $h_1\in (H\cap K)h_2$ and we have $(H\cap K)h_1=(H\cap K)h_2$. Then $[H:H\cap K]=|A|\leq |B|=[G:K]$. Then notice that $H\cap K$ is a subgroup of $H$, and in turn $H$ is a subgroup of $G$, thus we have $$ [G:H\cap K]=[G:H][H:H\cap K]\leq [G:H][G:K], $$ therefore, $H\cap K$ has finite index in $G$.

share|cite|improve this answer
Thanks. Why is $[G: H \cap K] = [G:H][H:H \cap K]$? Is this a general resut if $H,K \leq G$ ? – Epsilon Jan 5 '13 at 14:59
It follows from the third isomorphism theorem. Since $(G/H\cap K) \big/ (H/H\cap K)\cong G/H$, we multiply the index over and obtain the equality. Hope this helps! – Clayton Jan 5 '13 at 15:02
Yes this helps :) Bute $K,N$ are maybe not normal in $G$ ? – Epsilon Jan 5 '13 at 15:04
Yes, since we know $H\cap K\leq H$ and $H\leq G$. That is all the idea is, really. – Clayton Jan 5 '13 at 15:09
I have chosen the answer of Serkan because its faster. But yours is nice,too because its very precise. – Epsilon Jan 5 '13 at 15:39

Let $(H\cap K)a$ be the rigth coset of subgroup $H\cap K$ in your group $G$. It can be shown that $(H\cap K)a=Ha\cap Ka$. In fact as you noted, every right(or left) coset of $H\cap K$ is an intersection of a right coset of $H$ and a right coset of $K$. It means as you also noted the right cosets of $H\cap K$ resemble $Hx\cap Ky$. If we enumerate the distinct right cosets of $H\cap K$, we will find the number of them less than all above compositions, $Hx\cap Ky$, in $G$. So the index of $H\cap K$ in the group is finite and less than $mn$. Moreover if $(m,n)=1$ then $[G:H\cap K]=mn$.

share|cite|improve this answer
+1 I think I saw this question asked again this month! Maybe you came to the rescue for that newer question, too. – amWhy Feb 23 '13 at 3:26

Addressing Andre's question:

$$\forall\,x\in G\;\;\exists\, !\,\,h_i\in\{h_1,...,h_m\}\,\,s.t.\,\,x=h_i\alpha\,\,,\,\,\alpha\in H$$

But for this $$\alpha\,\,\,\exists\,!\,k_j\in\{k_1,...,k_n\}\,\,s.t.\,\,\alpha=k_j\beta\Longrightarrow \left\{h_ik_j\right\}_{1\leq i\leq m\,,\,1\leq j\leq n}$$

is a complete, though perhaps not minimal, set of representatives of the cosets of $\,H\cap K\,$ in $\,G\,$

share|cite|improve this answer
This is what you commented Andre correctly. – Babak S. Jan 5 '13 at 17:20

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.