9
$\begingroup$

In the spirit of this question, it would be interesting to give a characterization of HNN extensions as a 2-colimit. If $G$ is a group and $\alpha:H \xrightarrow{\cong} K$ is an isomorphism between two subgroups of $G$, then I think that the HNN extension $G_{\alpha}$ has the following universal property: if we let $i_1:H \hookrightarrow G,i_2:K \hookrightarrow G$ be the canonical inclusions, then the set of group homomorphisms $G_{\alpha} \to T$ are in natural bijection to pairs $(f,t)$ where $f:G \to T$ is a group homomorphism and $t:f \circ i_2 \circ \alpha \Rightarrow f \circ i_1$ is a 2-morphism. Here we're considering $\mathbf{Grp}$ as a full subcategory of $\mathbf{Cat}$ which carries a standard 2-category structure.

I'm just wondering if this universal property can be phrased as some kind of 2-colimit.

$\endgroup$
  • $\begingroup$ I assume you are meaning to think of groups as one-object categories? It's not obvious to people who haven't thought about it this way before. $\endgroup$ – David Roberts Feb 17 at 6:26
  • 1
    $\begingroup$ David, I think this is clear from the last sentence "we're considering Grp as a full subcategory of Cat". $\endgroup$ – Martin Brandenburg Feb 17 at 9:18
10
$\begingroup$

Assuming your universal property is true, it exactly says that the HNN extension is the coinserter of $(i_2 \circ \alpha,i_1) : H \rightrightarrows G$.

$\endgroup$

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.