I've got a problem and i'm not sure how to start this. Could somebody give me a hand?

Let G be a group of odd order.

Show that the function f: G --> G defined by f(x) = x^2 is one-to-one.

Printable View

- April 9th 2011, 04:42 PMRelmiwShowing a Function on Groups is One-to-One
I've got a problem and i'm not sure how to start this. Could somebody give me a hand?

Let G be a group of odd order.

Show that the function f: G --> G defined by f(x) = x^2 is one-to-one. - April 9th 2011, 09:27 PMroninpro
To show that a function is one-to-one, you need to show that the only element that maps to the identity is the identity. In other words, . You can try to combine this with Lagrange's theorem: the order of a subgroup divides the order of the group.

Can you see what to do? - April 10th 2011, 12:28 AMDeveno
if x^2 = e (trivially, e^2 = e in any group) and x is not e, what can we say about the order of G?

- April 10th 2011, 01:42 AMtonio
- April 10th 2011, 01:43 AMtonio
- April 10th 2011, 11:29 AMDeveno
it's a hint, not the answer. my intent was to get the asker to think about the implications of G having odd order.

if x is any element in such a group, x has odd order, because....

continuing, this means that if x^2 = y^2, and if the order of x is k = 2m+1, x = xe = x(x^k) = x(x^(2m+1)) = x^(2m+2) = (x^2)^(m+1)

= (y^2)^(m+1) = y^(2m+2) so x is in <y>, so <x> < <y>. but by the same token, if the order of y is r = 2n+1, y = ye = y(y^(2n+1))

= y^(2n+2) = (y^2)^(n+1) = (x^2)^(n+1).

so y is in <x>, hence <x> = <y>, so k = r, and m = n. hence x = y^(2m+2) = y^(2m+1)y = ey = y, so g-->g^2 is injective, homomorphism or not. - April 10th 2011, 11:48 AMRelmiw
- April 10th 2011, 02:01 PMSwlabr
What you said isn't true - the function you give maps two elements to the identity, f(-1)=1=f(1), so what you've written doesn't form a counter-example. Specifically, it doesn't form a counter-example because your map is a homomorphism (as your group is abelian)...

I am not sure what your point is - are you perhaps getting confused with 1-1, which means injection, and a "there exists a 1-1 correspondence between the sets", which means there exists a bijection?

Alternatively, counter-example to the comment: `A function is an injection if the only element which maps to the identity is the identity' would be . This isn't a homomorphism, and so your statement works. In general, the kernel of a homomorphism is trivial if and only if the function is an injection. This is easy to prove, and was what Roninpro was trying to say, and indeed did say when you realise he said ker(f)=e...normal functions don't have kernels... - April 10th 2011, 02:46 PMDeveno
i think the point tonio was trying to make was, there is a distinction between:

a) a homomorphism f is injective iff ker(f) = {e}, and

b) a function f is injective only if f^-1(e) = {*}, a singleton set (note the "only if").

the thing is, if f is only a function, and not a homomorphism, it could well be that the pre-image of the identity is a singleton set, without f being 1-1. tonio's example

is not the best one, since for his map f^-1(1) = {-1,1}. swlabr's example is a good counter-example for a function, but not for a function of the form x-->x^2.

in my orginal reply, i hinted at the fact that any group of even order must have an element of order 2 (the map x-->x^-1 on G-{e} must have a fixed point),

so there aren't any such counter-examples for groups of even order, x-->x^2 is never injective. as he (rightfully) pointed out, this is not the same as saying for

groups of odd order, x-->x^2 is always injective (~A --> ~B is not the same as A-->B, we have to rule out A --> ~B. here "A" is: G is of odd order,

and "B" is: the squaring map is injective).

judge's decision: tonio gets 5 points for pointing out the logical subtleties of the problem, and loses 2 for improper rebuttal. - April 10th 2011, 07:19 PMtonio

I think the rebuttal was not only proper but necessary and important since what was being

rebutted was incorrect, so I should lose only 1 point for giving an incorrect counterexample.

Tonio

Pd. You know what? Take away an other point from me for being a pest and rebutting again... - April 10th 2011, 07:25 PMtonio
- April 11th 2011, 12:21 AMSwlabr
A counter-example where the function maps only the identity to the identity but isn't 1-1 is group with presentation,

.

It is torsion free (a group with one-relator is has torsion if and only if the relator is a true power) and so the only element which maps to the identity is the identity. If a and b are non-equal elements then the function is not 1-1 since a and b are mapped to the same element. However, a and b*are*non-equal elements as the group is not infinite cyclic. This is because it isomorphic to the group with relator (replace with via a Tietze transformation) and so the group has abelianisation , and so cannot be infinite cyclic.

Also - my statement `what you said isn't true' was a remenant of a previous edit, when I was still trying to work out what you were saying. It was late. I was tired... - April 11th 2011, 08:46 AMDeveno