Understanding Semi-Direct Products in Group Theory and Lie Algebras

  • Level: Graduate 
  • Thread starter Thread starter haushofer
  • Start date Start date
Join the discussion
Registration is free. Start your own thread to ask a follow-up.
3 replies · 2K views
Science Advisor
Insights Author
Messages
3,086
Reaction score
1,608
Hi,

I try to get a grasp on semi-direct products, by notes written by Patrick J. Morandi ("Semi direct products"). I see that the notion of a semi-direct product is more general than a direct product.

However, the author states that

A group G is a direct product of two groups iff G contains normal subgroups [itex]N_1[/itex] and [itex]N_2[/itex] such that [itex]N_1\cap N_2 = \{e\}[/itex] and [itex]G= N_1 N_2[/itex].

Why is this exactly the case?

And also, how can I translate this for Lie groups on the level of the Lie algebra? (For instance, for the Poincare group). If someone knows good notes or a textbook I'm happy to be informed also :)
 
Physics news on Phys.org
I see that

[itex] N_1\cap N_2 = \{e\}[/itex]
gives that the decomposition is unique, but I don't see why the subgroups have to be normal. What happens if they're not?
 
It is a standard theorem in group theory that if [tex]H[/tex] and [tex]K[/tex] are normal subgroups of [tex]G[/tex] and [tex]H\cap K=\{e\}[/tex], then [tex]HK\cong H\times K[/tex].

see e.g. http://homepage.mac.com/ehgoins/ma553/lecture_21.pdf ("recognition theorem").

You can probably prove the converse for yourself. (just think of {(h,e)|h\in H} and {(e,k)|k\in K})
 
Last edited by a moderator:
Ok, thanks! Yes, the converse is quite clear to me I guess, but I don't see clearly why these subgroups have to be normal. I'll check your link, thanks again! :)