Factor (Quotient) Space definitions.

Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
1 reply · 2K views
wotanub
Messages
230
Reaction score
8
I'm learning algebra by myself and this concept is confusing me. Please excuse me if I define anything wrong... I've never expressed myself in this language before.

Lets say we have a group [itex]G[/itex] and a group [itex]G'[/itex] and there exists a homomorphism [itex]R: G → G'[/itex] and for any element [itex]g \in G[/itex], the equivalence class of g is denoted as [itex][g]_{R} = \{h \in G \:|\: f(h) = f(g)\}[/itex]

I understand the factor space [itex]G/R[/itex] as the set of all equivalence classes of [itex]G[/itex]:
[itex]G/R = \{[g]_{R} \:|\: g \in G\}[/itex]

but another way I always see this explained (that I'm not clear on) is if we have a subgroup [itex]H \subset G[/itex] then we can define a factor space with left cosets.

[itex]G/H = \{gH \:|\: g \in G\}[/itex]

How are these definitions stating the same thing? Does it have something to do with [itex]H[/itex] being the kernel of a homomorphism? I don't really understand what cosets have to do with equivalence relations.
 
Physics news on Phys.org
The two are closely linked.

Take the factor group approach. The quotient set is defined as ##G/H = \{gH~\vert~g\in H\}## with ##H## a subgroup. This ##H## has to be normal if you want ##G/H## to have a natural group structure, so we will do this. This actually corresponds to the following equivalence relation: we define ##g\sim h## iff ##g^{-1}h\in H##. The equivalence classes correspond exactly to the cosets. That is: ##[g]_\sim = gH##. So the coset definition actually does correspond to an equivalence relation.

Now, the link with homomorphisms is the following:
Given a homomorphism ##f:G\rightarrow G^\prime##, then we set ##gRh## iff ##f(g) = f(h)##. But we can take ##H = \textrm{Ker}(f) = \{h\in G~\vert~ f(h) = e\}##. This is a normal subgroup. Then we see that
[tex]f(g) = f(h)~\Leftrightarrow f(g^{-1}h) = e~\Leftrightarrow g^{-1}h\in H~\Leftrightarrow g\sim h[/tex]
So this equivalence relation is nothing more than the one defined above.

Conversely, if we are given a normal subgroup ##H## of ##G##, then we can always find a group ##G^\prime## and a homomorphism ##f:G\rightarrow G^\prime## such that ##H = \textrm{Ker}(f)##. Indeed, just take ##G^\prime = G/H## and take ##f(g) = gH##.

So the two methods outlined by you are equivalent.
 
  • Like
Likes   Reactions: 1 person