1

If $G$ is a group such that $x^2 = e$ for all elements $x$ in $G$, then show that $G$ is abelian.

I wanted to propose a method, that is based on properties of order $2$ elements, that I observed is useful in doing exercises on filling partially filled tables.
The property is:
If any element $a$ has order $2$, then get two sets of equations, where $\exists x,y\in G$ :
1. $ax = y$
2. $ay = x$

So, $aay = y \mid aax = x \implies a^2 = 1$

Similarly, for the set of equations:
3. $xa =y$
4. $ya =x$

I thought that this approach could be used to construct an intuitive explanation, but facing problems as shown below:

Let there be a group table for $G$: \begin{array}{|c|c|c|c|c|} \hline * & e & a & b & c & d \\ \hline e & e & a & b & c & d \\ \hline a & a & e & x & y & z \\ \hline b & b & x' & e & m & n \\ \hline c & c & y' & m' & e & p \\ \hline d & d & z' & n' & p' & e \\ \hline \end{array}

To show it is an abelian group, need diagonal elements as equal.
For $cb = m'$, need show $bc = m = m'$.

Want to use the property stated earlier for order $2$ elements, and for that the last statement means:
$cb = m'$,
$m'b = c$

There is only one choice for $m'=d$, so:
$cb = d$,
$db = c$

\begin{array}{|c|c|c|c|c|} \hline * & e & a & b & c & d \\ \hline e & e & a & b & c & d \\ \hline a & a & e & x & y & z \\ \hline b & b & x' & e & m & n \\ \hline c & c & y' & d & e & p \\ \hline d & d & z' & c & p' & e \\ \hline \end{array}

but this means that $ab =a\implies b =e$.

So, where lies the mistake?

Shaun
  • 44,997
jiten
  • 4,524
  • @ArturoMagidin $x,y,z,m,n,p$ are group elements, but did not find the relevant mappings. So, need to name them. Similarly, for $x',y',z',m',n',p'$. – jiten Apr 08 '20 at 22:32
  • 1
    The diagonal being equal just means every element has order $2$, not that it's abelian. This is what you're being asked to prove, if the diagonal is of this form then it's abelian. – CyclotomicField Apr 08 '20 at 22:33
  • 1
    Why did you use a table for 5 elements? There is nothing that specifies anything about the order of the group. Although it is useful to try small examples, in this case it turns out to be impossible. If every non-identity element in a group is an involution, then it's order must be equal to $2^n$, for some integer $n$. – cxx Apr 08 '20 at 22:33
  • @CyclotomicField It is my self created question - as could not understand why it is intuitively so. Say, this is here at : https://math.stackexchange.com/q/238171/424260 – jiten Apr 08 '20 at 22:36
  • 1
    There is in fact no group with $5$ elements in which every non-identity element has order $2$. So it is not a surprise that you cannot construct such a table! A group in which every nonidentity element has order $2$, if finite, must have order a power of $2$, but you are probably too early in your study of groups to know this. – Arturo Magidin Apr 08 '20 at 22:40
  • @cxx So, it all boils down to order of the group being a power of $2$. I will search for reason behind it, or better you post an answer to it. – jiten Apr 08 '20 at 22:40
  • 1
    Well, no; it does not boil down to that, because there are groups of order a power of $2$ that are not abelian... – Arturo Magidin Apr 08 '20 at 22:41
  • @ArturoMagidin Request an answer that details your points. – jiten Apr 08 '20 at 22:45

4 Answers4

3

The reason you can’t make it work is probably unknown to you at this stage. You are trying to construct a group with $5$ elements in which every element has order $1$ or $2$. However, no such group exists! As a consequence of Cauchy’s Theorem, in a finite group $G$, if a prime $p$ divides the order of the group $G$, then $G$ contains an element of order $p$. As a consequence of Lagrange’s Theorem, the order of any element of $G$ divides the order of $G$. And so if a finite group satisfies the condition that every element is of order $1$ or $2$, then the group must be of order $2^n$ for some $n\geq 0$. In particular, it cannot have order $5$.

But of course those two theorems are probably material you have not yet reached in your study.

It is also not the case that the answer boils down to the group having order a power of two, because there are finite groups of order a power of $2$ that are not abelian. For example, the multiplicative group of simple quaternions, $$ Q_8 = \{1, -1, i, -i, j, -j, k, -k\}$$ is a group of order $2^3=8$ under multiplication (with rules $i^2=j^2=k^2=ijk=-1$) is not abelian. Of course, it is also not the case that every element is of order $1$ or $2$, but the point is that you really need “every element is of order $1$ or $2$”, and you can’t just get away with “the order of the group is a power of $2$”.

There are many ways of proving that this condition suffices. Here is a slightly different one, using the weaker property that $(ab)^2 = a^2b^2$ for all $a,b\in G$ (if every element is of order $1$ or $2$ then this holds, but this may hold even if not every element is of order $1$ or $2$): given any $x,y\in G$, we have $$xyxy = (xy)^2 = x^2y^2 = xxyy.$$ Then multiplying by $x^{-1}$ on the left and $y^{-1}$ on the right, we obtain $yx=xy$. Thus, $G$ is abelian.

Arturo Magidin
  • 398,050
  • Please tell the reason why the property $(ab)^2 = a^2b^2$ is a slightly weaker property. – jiten Apr 08 '20 at 23:17
  • 1
    @jiten: Because if every element has order $1$ or $2$ then the condition holds, but the condition can hold in groups in which not every element is of order $1$ or $2$; namely, it holds in every abelian group. – Arturo Magidin Apr 08 '20 at 23:23
2

Just use the property $a^2=e$ for all $a\in G$. Let $a,b\in G$ then $$ ab=a(ab)(ab)b=(aa)ba(bb)=ba. $$

maciek97x
  • 295
1

I must admit I found our OP jiten's proof a little difficult to parse.

As an alternative, one might say:

$\forall x, y \in G, \; x^2 = y^2 = (xy)^2 = e, \tag 1$

from which

$x^2 = e \Longrightarrow x = x^{-1}x^2 = x^{-1}e = x^{-1}, \tag 2$

and similarly

$y = y^{-1}; \tag 3$

thus, since

$(xy)^2 = e, \tag 4$

we have

$yx = (yx)^{-1} = x^{-1}y^{-1} = xy, \; \forall x, y \in G, \tag 5$

and $G$ is abelian.

Now suppose the order of $G$ is not a power of $2$; then there is an odd prime $p$ such that

$p \mid \vert G \vert; \tag 6$

but then by Cauchy's theorem, $G$ contains an element of order $p$; however this contradicts (1), i.e., the hypothesis that every element of $G$ is of order $2$. Thus

$\vert G \vert = 2^n, \; n \in \Bbb N. \tag 7$

$OE\Delta$.

Robert Lewis
  • 71,180
1

Here's one way to do it: (I have left out a lot of details, though.)

Suppose $(G, +)$ is a finite group such that for every $g \in G$, $g^2 = e$. We can make $G$ into a vector space over $\mathbb{Z}_2$ by defining addition to be the group operation and multiplication to be $0 \cdot g = e$ and $1\cdot g = g$ for any $g \in G$. Indeed, one can check that this properly defines a vector space structure on $G$. It immediately follows that $|G| = 2^n$, where $n$ is the dimension of $G$.

cxx
  • 524
  • Please provide details, that show it clearly how the group order must be a power of $2$. – jiten Apr 08 '20 at 23:01
  • @jiten Actually, I've intentionally left out the details. You should try to fill them in yourself. It's a good exercise! – cxx Apr 08 '20 at 23:05
  • Please add more details so as to complete the exercise, as not clear about how $G$ has been made into a vector space over $\mathbb{Z_2}$. Your answer is intuitive and request more details so as to be able to work on that. – jiten Apr 08 '20 at 23:18
  • I am waiting for your response by a detailed answer, as otherwise confused how to pursue. Searched literature fit for me. In Mirman's book mention occurs on pg. 59; but only a half page on that. Just underlies that for groups, the concept of a vector space is needed. In Lovett's book; it occurs in section 8.6 and seemingly concerns with Representations of Groups. Also, a complete section is devoted with many examples. Please comment if am going in correct direction? – jiten Apr 08 '20 at 23:54
  • Seems like the book by James B. Carell titled Groups, Vectors and Metric Spaces is the best possible choice. Please respond somewhat as your input will help a lot. – jiten Apr 09 '20 at 00:26
  • Well it seems that you haven't really covered material up to vector spaces, so I'm hesitant to tell you what resources you should use. It would be better to just continue learning from whatever course/textbook you're following. Right now it wouldn't make sense for you to dive into stuff about vector spaces, if you haven't covered the preliminary topics. (Or maybe you could if you really want, but I don't really know, either.) – cxx Apr 09 '20 at 00:31
  • I have covered, but am out of touch now. – jiten Apr 09 '20 at 00:40