(1.) Why did Fraleigh shirk the proof for $(2) \implies (1)$?
By dint of Arthur's comment, $(2) \iff \color{crimson}{gHg^{-1} \subseteq H} \quad \wedge \quad gHg^{-1} \supseteq H \implies \color{crimson}{gHg^{-1} \subseteq H} \iff (1)$
(2.) In $(1) \implies (2)$, how does $\{ghg^{-1} : h \in H \} \subseteq H$?
(3.) I know left cosets $\neq$ right cosets. The same $H$ appears on both sides $gH = Hg$ in:
$gH = Hg \iff gh_1 = h_2g$, hence why isn't $h_1 = h_2$ always?
I read $gH = Hg$ is a set equality and not an equality elementwise. But I'm confounded.
To boot, I know $gh_1 = h_2g \iff G$ Abelian $\iff \color{magenta}{g^{-1}}gh_1 = \color{magenta}{g^{-1}}gh_2 \iff h_1 = h_2$.
$G$ can be nonAbelian hence if it is nonabelian, the previous line muffs.

Let $f_g:G\mapsto G$ defined by $f_g(x)=gxg{-1}$. Note that $f_g$ is an bijection from $G$ to $G$(actually it is an isomorphism from $G$ to $G$) Thus,restriction of $f_g$ on any subset $S$ of $G$ is also a bijection from $S$ to its image.
Now,if $G$ is finite your question is very trivial since if $gNg^{-1}\subseteq N$ then we know that $|gNg^{-1}|=|N|\implies gNg^{-1}=N $.(their order equal since it is an bijection)
For general case,Since $f_g$ is bijection so it has inverse namely $f_{g^{-1}} $ or vice versa.Thus,restriction of $f_g$ on a subset $S$,$f_g:S\mapsto Im(S)$ is invertable.
Now, it is given that $f_g(N)\subseteq N$ for all $g\in G$.So,
$$f_{g^{-1}}(N)\subseteq N$$ $$f_g(f_{g^{-1}}(N))\subseteq f_g(N)$$ $$N\subseteq f_g(N)\implies N=f_g(N)$$ Which is showing that $gNg^{-1}\subseteq N\implies gNg^{-1}=N$ we are done.