p. 128, 129. Theorem 13.12. Let $h$ be a homomorphism of groups $G \to G'$.
III. If $S \le G$, then $h[S] \le \color{red}{G'}$.
IV. If $S' \le G'$, then $h^{-1}[S'] \le G$.

p. 149. Theorem 15.16. Let $h$ be a homomorphism of groups $G \to G'$.
III. If $N \unlhd G$, then $h[N] \unlhd \color{red}{h[G]}$.
IV. If $N' \unlhd \color{red}{h[G]}$, then $h^{-1}[N'] \unlhd G$.

Note that $h[N]$ may not be normal in G', even though $N \unlhd G$. For example, define $h: \mathbb{Z}_2 \to S_3$ by means of $h(0) = id$ and $h(1) = (1)(2,3)$. h is a homomorphism. Any group is a normal subgroup of itself, but $\{id , h(1)\}$ is not a normal subgroup of $S_3$.

(1.) What's the intuition behind the first difference in red? Why $h[G]$ instead of just $G'$?

(2.) What's the intuition behind the second difference in red?

(3.) How do you envisage and envision Fraleigh's counterexample for $h[N] \not \unlhd G'$?

(4.) To prove $\{id , h(1)\} \not \unlhd S_3,$ I can use the Normal Subgroup Test.
Hence I need to find exactly one $g \in S_3$ so that $g\{id , h(1)\}g^{-1} \not \in S_3$.
How do you find $g$ craftily, without trial and error? My Cayley diagram by dint of Group Explorer:

enter image description here


If $S\leq G$, then $h[S] \leq G'$: This is because the map is a homomorphims.ie It preseves the fact that a subgroup is a group.

If $N⊴G$, then $h[N]⊴ h[G]$: Normality of a subgroup depends on the group containing it. The group structure of $G$ is preserved by $h$ and so it preserves its subgroup structures as well. The third case is similar.

Lastly, to show the group is not normal, recall that conjugacy classes in $S_3$ can be seen from cycle type. In particular, if you conjugate by an element with different cycle type e.g $(123)$, you'll leave the subgroup.

  • $\begingroup$ The first part doesn't make sense. If $h\colon G\to G'$ then $h[S]$ is by definition contained in $G'$, not $G$. $\endgroup$ – Hagen von Eitzen Feb 21 '14 at 19:01
  • $\begingroup$ that must be a typo $\endgroup$ – Hesky Cee Feb 21 '14 at 19:04
  • $\begingroup$ @IhechukwuChinyere: your comment seems to answer the question. You answer seems to miss the point. $\endgroup$ – robjohn Feb 24 '14 at 17:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.