Let $\pi: G\to G_1$ be a surjective group homomorphism to a finite group $G_1$ and let $S_1$ be a (finite) generating set of $G_1$. Assume $\mathrm{Cay}(G_1,S_1)$ is the Cayley graph of $G_1$ with respect to the generating set $S_1$. Define $S:=\pi^{-1}(S_1)$ and consider the Cayley graph $\mathrm{Cay}(G,S)$.

I would like to know if there is any criteria that allows one to construct an embedding of $\mathrm{Cay}(G_1,S_1)$ into $\mathrm{Cay}(G,S)$. Surly this is not always possible but I am looking for an example that existence of such embedding is possible. For example when $\pi: \mathrm{GL}_2(\mathbb{Z}_p)\to \mathrm{GL}_2(\mathbb{F}_p)$ is the reduction map and one can use for instance Hensel lemma to construct fibers.


1 Answer 1


This is possible (and easy), because you have defined $S$ to be the entire inverse image of $S_1$. (An embedding is unlikely to exist if $S$ is some small subset of $G$ that is mapped onto $S_1$ by $\pi$.)

For each $g_1 \in G_1$, let $\varphi(g_1)$ be any element of $\pi^{-1}(g_1)$. Then $\varphi$ is an embedding of $\mathrm{Cay}(G_1,S_1)$ onto an induced subgraph of $\mathrm{Cay}(G,S)$: for $g_1, h_1 \in G_1$, we have $$ \varphi(g_1) - \varphi(h_1) \Leftrightarrow \varphi(g_1) \varphi(h_1)^{-1} \in S = \pi^{-1}(S_1) \Leftrightarrow \pi \bigl( \varphi(g_1) \varphi(h_1)^{-1} \bigr) \in S_1 \Leftrightarrow g_1 h_1^{-1} \in S_1 \Leftrightarrow g_1 - h_1 .$$

  • $\begingroup$ Another way to say what Dave said is that, since $S$ is the full preimage of $S_1$, as a graph, $\mathrm{Cay}(G,S)$ is the lexicographic product of $\mathrm{Cay}(G_1,S_1)$ with the edgeless graph on $G/G_1$ (and thus any choice of one vertex per preimage will induce an embedding). $\endgroup$
    – verret
    Mar 11, 2015 at 8:21

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.