11
$\begingroup$

The problem is to prove or disprove that there is a noncyclic abelian group of order $51$.

I don't think such a group exists. Here is a brief outline of my proof:

Assume for a contradiction that there exists a noncyclic abelian group of order $51$.

We know that every element (except the identity) has order $3$ or $17$. Assume that $|a|=3$ and $|b|=17$. Then I managed to prove that the subgroups generated by $a$ and $b$ only intersect at the identity element, from which we can show that $ab$ is a generator of the whole group, so it is cyclic. Contradiction.

So every element (except the identity) has the same order $p$, where $p$ is either $3$ or $17$.

If $p=17$, take $a$ not equal to the identity, and take $b$ not in the subgroup generated by $a$. Then we can prove that $a^kb^l$ where $k,l$ are integers between $0$ and $16$ inclusive are distinct, hence the group has more than $51$ elements, contradiction.

If $p=3$, take $a$ not equal to the identity and take $b$ not in the subgroup generated by $a$. Then we can prove that $a^kb^l$ where $k,l$ are integers betwen $0$ and $2$ inclusive are distinct. This subgroup has $9$ elements so we can find $c$ that's not of the form $a^kb^l$. Then we can prove that $a^kb^lc^m$ where $k,l,m$ are integers betwen $0$ and $2$ inclusive are distinct. Then this subgroup has $27$ elements so we can find $d$ that's not of the form $a^kb^lc^m$. Then we prove that $a^kb^lc^md^n$ where $k,l,m,n$ are integers between $0$ and $2$ inclusive are distinct, this being $81$ elements. Contradiction.

$\endgroup$
  • $\begingroup$ Have you learned the Sylow theorems? $\endgroup$ – coldnumber Jul 8 '15 at 5:31
  • $\begingroup$ I haven't learned the Sylow Theorems. $\endgroup$ – Cay Jul 8 '15 at 5:31
  • $\begingroup$ If you've learned Lagrange's theorem, then your last paragraph can be shortened. $\;$ $\endgroup$ – user57159 Jul 8 '15 at 6:59
  • $\begingroup$ I have learned Lagrange's Theorem, but I can't see how it can help us shorten the last paragraph. $\endgroup$ – Cay Jul 8 '15 at 7:53
  • $\begingroup$ Your problem is a particular case of the following problem from Herstein's Book:Let $G$ be a group of order $pq$ ($p,q$ are primes and $p <q$) and $p$ does not divide $q-1$ then $G$ is a cyclic group.Moreover if p divides $q-1$ then there is exactly one nonabelian group of order $pq$.As far as I know proof of this requires a Use of Sylow's Theorem. $\endgroup$ – Arpit Kansal Jul 8 '15 at 10:05
9
$\begingroup$

Using the Sylow theorems would shorten your proof considerably, because from the first Sylow theorem it follows that if a prime divides the order of a group, then the group contains an element of that order. (This eliminates the need to check cases where all elements have order 3 or all have order 17.)

Hence, if $\left|G\right|=51$, then $G$ has an element $a$ of order 3 and an element $b$ of order 17.

Then, as you said, $ab$ has order 51; you can show this directly, without using the proposition for product groups:

Since $\left|G\right|=51$, by Lagrange's theorem, the order of $ab$ divides 51.

Now, since $G$ is abelian, $(ab)^3=a^3b^3=b^3 \neq 1$ because $b$ has order 17, and $(ab)^{17}=a^{17}b^{17}=a^2\neq 1$, because $a$ has order 3.

Hence, $ab$ generates the group, so $G$ is cyclic.

$\endgroup$
  • $\begingroup$ Thanks! I like your solution! $\endgroup$ – Cay Jul 8 '15 at 5:54
  • $\begingroup$ Thanks! I edited because I realized invoking Lagrange is easier than what I was doing to show that the order of $ab$ divides 51. $\endgroup$ – coldnumber Jul 8 '15 at 6:04
  • $\begingroup$ There is no need for the Sylow theorems here! The result you mention is Cauchy's theorem, much simpler than Sylow's theorems. $\endgroup$ – Matthew Towers Jul 8 '15 at 10:27
  • $\begingroup$ @mt_ Interesting! I hadn't heard of it before; Artin has the existence of the elements as a corollary to the first Sylow Theorem. $\endgroup$ – coldnumber Jul 8 '15 at 10:38
  • 1
    $\begingroup$ @ColdNumber, the proof of the Sylow theorems for abelian groups is also very simple! $\endgroup$ – Mariano Suárez-Álvarez Jul 8 '15 at 22:43
5
$\begingroup$

You're right.

If you know the theorem that classifies finite abelian groups, then the only possible abelian groups of order $51$ are $\mathbb Z/51 \mathbb Z$ which is cyclic and $\mathbb Z/3 \mathbb Z \times \mathbb Z/17 \mathbb Z$ which is also cyclic because $3$ and $17$ are prime and $\gcd(3,17)=1$ so $\mathbb Z/3 \mathbb Z \times \mathbb Z/17 \mathbb Z \simeq \mathbb Z/51 \mathbb Z$.

$\endgroup$
2
$\begingroup$

Any finite abelian group is isomorphic to a group of the form $$\Bbb Z_{p_1^{a_1}}\times \Bbb Z_{p_2^{a_2}}\times\cdots\times \Bbb Z_{p_n^{a_n}}$$ where $p_i$ are (not necessarily distinct) primes. The order of such a group is $p_1^{a_1}\cdots p_n^{a_n}$. How many ways can this be done for $51$?

$\endgroup$
  • 1
    $\begingroup$ Thanks! So it must be $\mathbb{Z}_{51}$ or $\mathbb{Z}_3 \times \mathbb{Z}_{17}$, both of which are cyclic. I did see that theorem later in my textbook, but I thought I wasn't allowed to use it as this question appeared in an earlier chapter. Does my proof work? $\endgroup$ – Cay Jul 8 '15 at 5:30
  • 1
    $\begingroup$ @Cay Your proof looks very good, except for the parts where you just say "it can be proven that they are all distinct". I believe you, but you would need to actually prove it to get full score in my book, had I been your teacher. $\endgroup$ – Arthur Jul 8 '15 at 5:40
  • 1
    $\begingroup$ @Cay actually, $\mathbb{Z}_{51}$ is isomorphic to $\mathbb{Z}_3\times \mathbb{Z}_{17}$ because gcd(3,17)=1. This is true in general for any integers $m$ and $n$. Maybe you could mess around and try to prove this with a strategy similar to the proof you used in this question. $\endgroup$ – Alex Mathers Jul 8 '15 at 5:45
  • $\begingroup$ @mathers101 Yes you're right and I see why this is true now. $\endgroup$ – Cay Jul 8 '15 at 5:52
  • $\begingroup$ @Arthur Yes if it had been an exam I would have fully written out the reasoning, just wanted to get out the main steps here :) $\endgroup$ – Cay Jul 8 '15 at 5:52

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.