So I'm proving that a group $G$ with order $112=2^4 \cdot 7$ is not simple. And I'm trying to do this in extreme detail :)

So, assume simple and reach contradiction. I've reached the point where I can conclude that $n_7=8$ and $n_2=7$.

I let $P, Q\in \mathrm{Syl}_2(G)$ and now dealing with cases that $|P\cap Q|=1, 2^2, 2^3$ or $2^4$.

I easily find contradiction when $|P\cap Q|=2^4$ and $2$.

Um, got stuck REAL bad on the case $|P\cap Q|=2^3$ and $2^2$.

If $|P \cap Q |=2^3= 8$ and $|P|=|Q|=16$, is there any relationship between $P,Q$ and their intersection that can help me?


If $G$ is a simple group, it must have exactly $7$ Sylow $2$-subgroups. Thus $G$ embeds into $S_7$, and in particular into $A_7$ since $G$ does not have a subgroup of index $2$. But the order $A_7$ is not divisible by $112$.

If you want to go along the lines of your original idea, you can rule out the case $|P \cap Q| = 2^3$ by noticing that then $P \cap Q$ is normal in $P$ and $Q$ (as a subgroup of index $2$), so $N_G(P \cap Q)$ contains $P$ and $Q$, which implies that $N_G(P \cap Q) = G$.

ADDED: I'm not sure if there is an easy way to deal with rest of the cases. However, there is a nice argument which also works for proving that every group of order $p^n q$ ($p$, $q$ distinct primes) is nonsimple. I believe the idea of the proof goes back to G. A. Miller (around 1900-1910). Here's an illustration of it in this case.

Suppose that $G$ is a simple group of order $112$. Then $G$ has exactly $7$ Sylow $2$-subgroups. Let $P, Q \in Syl_2(G)$ be such that $P \neq Q$ and that $D = P \cap Q$ has largest possible order. Steps for the proof:

  1. Using the fact that $D < N_P(D)$ and $D < N_Q(D)$ (proper inclusion), prove that $N_G(D)$ cannot be a $2$-group.

  2. Thus $D$ is normalized by an element $g \in G$ of order $7$. Prove that $P, gPg^{-1}, \ldots, g^6Pg^{-6}$ are distinct. Conclude that $D$ is contained in every Sylow $2$-subgroup.

  3. Since the intersection of all Sylow $2$-subgroups is normal, $D$ is trivial.

  4. By counting elements in Sylow $2$-subgroups, prove that $G$ contains exactly one Sylow $7$-subgroup.

This same argument works for proving the statement for groups of order $p^n q$.

  • $\begingroup$ Could you expand on your statement "Thus G embeds into S7". Thanks! $\endgroup$
    – Ziggy
    Apr 24 '13 at 0:50
  • 3
    $\begingroup$ @Ziggy: If $G$ is a simple group and $[G:H] = n$, then the left coset action gives us an injective homomorphism $f: G \rightarrow S_n$ (in this case $H = N_G(P)$ for $P$ some Sylow $2$-subgroup). With Sylow subgroups you could apply the conjugation action as well. $\endgroup$ Apr 24 '13 at 8:07
  • 1
    $\begingroup$ @ErginSuer: Let $H \leq S_n$. If $H$ has order $> 2$, then you can see that $H$ must contain an even permutation $\neq 1$. Thus $H \cap A_n$ is a nontrivial normal subgroup of $H$. In the case where $H$ is simple this implies $H \cap A_n = H$, ie. $H \leq A_n$. $\endgroup$ Sep 3 '14 at 6:09
  • 1
    $\begingroup$ @Explorer1234: You could just use Lagrange's theorem. You know $N_G(P \cap Q)$ has order a multiple of $2^4$ since it contains $P$, so..? Also I don't remember anymore if this is the original reference, but in the old book by Miller/Blichfeldt/Dickson, Miller gives the argument in §73 p. 185 there. $\endgroup$ Jun 21 at 15:19
  • 1
    $\begingroup$ @Explorer1234: For 1., if $N_G(D)$ is a $2$-group, it is contained in some $2$-Sylow $R \neq P$, and then $R \cap P > D$, contradicting maximality of $D$. In 2., you have $D$ contained in $g^i P g^{-i}$ for all $i$. $\endgroup$ Jun 23 at 7:14

I wanted to write out Mikko Korhonen's first idea for proof in detail as a separate answer, since it is not trivial at all, and provoked some questions in the comments.

As mentioned in the original question, we assumed $n_2=7$. From Sylow's second theorem, we know that all the 2-Sylow subgroups are conjugate, and we can look at $G$'s action on them by conjugation. This action implies a homomorphism: $$f:G\rightarrow S_7$$ $\ker(f)$ cannot be $G$ since the action is a transitive action. Then, if $\ker(f)$ is non-trivial (meaning $ker(f)\neq\{e\}$) then it is a non-trivial normal subgroup of $G$, and therefore $G$ isn't simple.

Otherwise, we get $ker(f)=\{e\}$ and therefore $f$ is injective, meaning that $G$ is isomorphic to a subgroup of $S_7$. For convenience purposes, we'll write $G\leq S_7$. $G$ cannot be contained in $A_7$ because 112 doesn't divide $|A_7|$. In that case $GA_7=S_7$ and using the second isomorphism theorem we get: $$G/(G\cap A_7)\cong GA_7/A_7\cong S_7/A_7\cong\mathbb{Z}_2$$ and therefore $[G:(G\cap A_7)]=2$ meaning that $G\cap A_7$ is a normal subgroup of $G$.


Sylow's theorems require that $n_2=1$ or 7, $n_7=1$ or 8, so $G$ has a chance to be simple only if $n_2=7$ and $n_7=8$. Note that the Sylow 7-subgroups can only intersect at the identity, any two Sylow 2-subgroups can share a subgroup of order at most 8, and a Sylow 7-subgroup and a Sylow 2-subgroup can share only the identity.

Hence the union of Sylow 7-subgroups has $1+8\cdot(7-1)=49$ elements, and the union of the Sylow 2-subgroups has at least $8+7\cdot(16-8)=64$ elements, which happens precisely when they all share a subgroup $H$ of order 8. In this case, the union of all Sylow 7-subgroups and 2-subgroups has $64+49-1=112$ elements, so we learn that no other scenario (one with a greater union of the Sylow 2-subgroups) is allowed.

Now notice that $H$ is a normal subgroup of $G$ since conjugation by $g\in G$ permutes the Sylow 2-subgroups and so preserves their intersection. Hence $G$ is not simple.

Edit: This answer is wrong because the union of the Sylow 2-sbgps can be smaller than 112, see the comments below.

  • $\begingroup$ how exactly do you know that the intersection of 2 sylow 2 subgp cannot be 8 though? $\endgroup$
    – Akaichan
    Apr 5 '13 at 3:14
  • 1
    $\begingroup$ If $n_2=7$, $n_7=8$, the intersection of all Sylow 2-subgroups (and hence any two) MUST have order 8, otherwise the elements wouldn't fit in $G$. The point is that when this happens, the intersection is a normal subgroup of $G$. $\endgroup$ Apr 5 '13 at 3:19
  • $\begingroup$ I'm sorry that I'm having a hard time with this but how is the intersection is a normal subgp of $G$? $\endgroup$
    – Akaichan
    Apr 5 '13 at 3:33
  • 1
    $\begingroup$ Let $H_j$, $j=1,\ldots,7$ be the Sylow 2-sbgps and let $H$ be their intersection. Let $g\in G$, then $gH_jg^{-1}=H_{\sigma(j)}$, where $\sigma\in S_7$. If $h\in H_j$, then $g h g^{-1}\in H_{\sigma(j)}$ and so if $h\in H$, then $ghg^{-1}\in H$ since $\sigma$ is a bijection. $\endgroup$ Apr 5 '13 at 3:47
  • 1
    $\begingroup$ I'm not sure I follow your argument in "the union of the Sylow $2$-subgroups has at least..". For example, let $G = S_3 \times S_3$. Then $G$ has exactly $9$ Sylow $2$-subgroups. Intersections of distinct Sylow $2$-subgroups have order at most $2$. So if I understand your claim, you're saying that $G$ would have at least $2 + 9 \cdot (4 - 2) = 20$ elements in the union of Sylow $2$-subgroups. But this isn't true, there are only $16$ elements in the union of Sylow $2$-subgroups of $G$. $\endgroup$ Apr 5 '13 at 7:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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