Nothing to do with category theory. A homomorphism of linear graphs is a fairly concrete object, and Conway only uses graph theoretic terminology to clarify his semi-formal exposition. Dunno if I’d say there’s much math not being respected.