I understand that if $R$ is an equivalence relation on $G$, the resulting partition cells are either equal or disjoint. I think I understand that the graph of the quotient set $G^R$ is constructed with a representative vertex from each cell.
Then, why does $G^R$ have edges if the cells are disjoint?
Suppose that $C_0$ and $C_1$ are distinct cells of the partition, with representative vertices $v_0$ and $v_1$, respectively. Then in $G^R$ there is an edge $v_0v_1$ if and only if there is at least one edge of $G$ joining a vertex in $C_0$ and a vertex in $C_1$. For a simple example, let $G$ be the graph shown below:
Partition the vertices of $G$ into cells $\{a,e\},\{b,f\},\{c\}$, and $\{d,g\}$, and let $a,b,c$, and $d$ be the representatives of these cells.
That accounts for all of the edges of $G$ that link different cells, so it accounts for all of the edges of $G^R$. $G^R$ therefore looks like this:
Or if you prefer, like this: