1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Orders of Groups

  1. Jan 12, 2008 #1
    Hi i have completed the answer to this question. Just need your verification on whether it's completely correct or not:

    Question:
    If G is a group and xEG we define the order ord(x) by:
    ord(x) = min{[itex]r \geq 1: x^r = 1[/itex]}

    If [itex]\theta[/itex]: G --> H is an injective group homomorphism show that, for each xEG, ord([itex]\theta(x)[/itex]) = ord(x)

    My answer: Please verify
    If [itex]\theta(x)[/itex] = {[itex]x^r: r \epsilon Z[/itex]} then ord([itex]\theta(x)[/itex]) = ord(x).

    For any integer r, we have x^r = e (or 1) if and only if ord(x) divides r.

    In general the order of any subgroup of G divides the order of G. If H is a subgroup of G then "ord (G) / ord(H) = [G:H]" where [G:H] is an index of H in G, an integer.
    So order for any xEG divides order of the group. So ord([itex]\theta(x)[/itex]) = ord(x)


    any suggestions or changes please? thnx :)
     
  2. jcsd
  3. Jan 12, 2008 #2

    HallsofIvy

    User Avatar
    Staff Emeritus
    Science Advisor

    This makes no sense. [itex]\theta(x)[/itex] is a single member of H, not a set of members of G.

    An "index of H in G"? It is not said here that H has to be a subset of G!

    Seems to me you could just use the fact that, for any injective homomorphism, [itex]\theta[/itex], [itex]\theta(x^r)= [\theta(x)]^r[/itex] and [itex]\theta(1_G)= 1_H[/itex].
     
  4. Jan 13, 2008 #3
    i believe we have to show 2 things:

    i) [itex](\theta(x))^a = e'[/itex]
    ii) [itex]0 < b < a \implies (\theta(x))^b \neq e'[/itex].

    ok so basically:

    If ord(x)=a then [itex]\left[ {\phi (x)} \right]^a = \left[ {\phi (x^a )} \right] = \phi (e) = e'[/itex].
    Now suppose that [itex]ord\left[ {\phi (x)} \right] = b < a[/itex].
    Then
    [itex]\left[ {\phi (x)} \right]^b = e' = \left[ {\phi (x)} \right]^a [/itex]
    [itex]\phi (x^b ) = \phi (x^a )[/itex]
    [itex]x^b = x^a[/itex] (injective)
    [itex]x^{a - b} = e[/itex]

    there seems to be a contradiction where if [itex]x^{a} = x^{b}[/itex], then [itex]\left[ {\phi (x)} \right]^b = e'[/itex] which is not what statement (ii) says.
    am i correct in this assumption? any ideas on how to deal with this?
     
  5. Jan 16, 2008 #4
    When you have [itex]x^{a-b} = e[/itex], then a-b is positive since you assumed b<a. But this contradicts the definition of order. Done.
     
    Last edited: Jan 16, 2008
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: Orders of Groups
Loading...