[proofplan]
We characterize an element $x \in G$ by whether it fixes every left coset $gH$. Fixing the coset $gH$ is equivalent to the conjugacy condition $g^{-1}xg \in H$, which is the same as $x \in gHg^{-1}$. Since this must hold for every $g \in G$, the kernel is exactly the intersection of all conjugates of $H$.
[/proofplan]
custom_env
admin
[step:Translate kernel membership into fixing every left coset]
Let
\begin{align*}
K := \ker \lambda.
\end{align*}
For an element $x \in G$, by the definition of the kernel of the homomorphism $\lambda: G \to \operatorname{Sym}(G/H)$, we have
\begin{align*}
x \in K \iff \lambda(x) \text{ is the identity permutation on } G/H.
\end{align*}
The identity permutation on $G/H$ fixes every element of the set $G/H$. Since every element of $G/H$ has the form $gH$ for some $g \in G$, this is equivalent to
\begin{align*}
(xg)H = gH \quad \text{for every } g \in G.
\end{align*}
Thus
\begin{align*}
x \in K \iff (xg)H = gH \text{ for every } g \in G.
\end{align*}
[/step]
custom_env
admin
[step:Convert fixed cosets into membership in every conjugate of $H$]Fix $x \in G$ and $g \in G$. We first record the coset equality criterion in this special case. The equality
\begin{align*}
(xg)H = gH
\end{align*}
holds if and only if $xg \in gH$, because two left cosets of $H$ are equal exactly when one representative belongs to the other coset. This condition is equivalent to the existence of an element $h \in H$ such that
\begin{align*}
xg = gh.
\end{align*}
Multiplying the equality $xg = gh$ on the left by $g^{-1}$ gives
\begin{align*}
g^{-1}xg = h.
\end{align*}
Therefore
\begin{align*}
(xg)H = gH \iff g^{-1}xg \in H.
\end{align*}
Equivalently,
\begin{align*}
g^{-1}xg \in H \iff x \in gHg^{-1}.
\end{align*}
Combining this with the preceding step gives
\begin{align*}
x \in K \iff x \in gHg^{-1} \text{ for every } g \in G.
\end{align*}[/step]
custom_env
admin
[guided]Fix $x \in G$ and one coset $gH \in G/H$. We want to translate the statement “$x$ fixes the coset $gH$” into a statement about membership in a conjugate of $H$.
The element $x$ fixes $gH$ under the left coset action exactly when
\begin{align*}
\lambda(x)(gH)=gH.
\end{align*}
By the definition of $\lambda$, this means
\begin{align*}
(xg)H = gH.
\end{align*}
Now we analyze this equality of left cosets. The coset $(xg)H$ equals $gH$ if and only if the representative $xg$ lies in the coset $gH$. This means precisely that there exists an element $h \in H$ such that
\begin{align*}
xg = gh.
\end{align*}
Multiplying on the left by $g^{-1}$ gives
\begin{align*}
g^{-1}xg = h.
\end{align*}
Since $h \in H$, this proves
\begin{align*}
(xg)H = gH \implies g^{-1}xg \in H.
\end{align*}
Conversely, suppose that
\begin{align*}
g^{-1}xg \in H.
\end{align*}
Define $h \in H$ by
\begin{align*}
h := g^{-1}xg.
\end{align*}
Then $xg = gh$, so $xg \in gH$. Hence the left coset represented by $xg$ is the same as the left coset represented by $g$:
\begin{align*}
(xg)H = gH.
\end{align*}
We have proved the equivalence
\begin{align*}
(xg)H = gH \iff g^{-1}xg \in H.
\end{align*}
Finally, the condition $g^{-1}xg \in H$ can be rewritten by multiplying the inclusion relation through the conjugation map. It is equivalent to saying that $x$ has the form $ghg^{-1}$ for some $h \in H$, namely
\begin{align*}
x \in gHg^{-1}.
\end{align*}
Thus $x$ fixes the particular coset $gH$ if and only if $x$ lies in the particular conjugate subgroup $gHg^{-1}$.[/guided]
custom_env
admin
[step:Intersect the conjugate conditions to identify the kernel]
From the previous steps, for every $x \in G$,
\begin{align*}
x \in \ker \lambda \iff x \in gHg^{-1} \text{ for every } g \in G.
\end{align*}
By the definition of set-theoretic intersection, the right-hand condition is equivalent to
\begin{align*}
x \in \bigcap_{g \in G} gHg^{-1}.
\end{align*}
Using the definition
\begin{align*}
\operatorname{Core}_G(H) := \bigcap_{g \in G} gHg^{-1},
\end{align*}
we obtain
\begin{align*}
x \in \ker \lambda \iff x \in \operatorname{Core}_G(H).
\end{align*}
Since this equivalence holds for every $x \in G$, the two subsets of $G$ are equal:
\begin{align*}
\ker \lambda = \operatorname{Core}_G(H).
\end{align*}
[/step]