Actually, three more things:
1. Manifolds also seem like a weird inclusion; they're just one topological concept, and they aren't essentially... irregular? tangled? pick your term.
2. George Landow's writing on hypertext (applying esp. Barthes, Derrida) is relevant philosophy, notable because it's about meaning-making in a graph of texts rather than traversal in a graph of states.
3. "Graphs are generically useful" isn't surprising per se, and bucketing a bunch of applications of graphs without stressing the (dis)similarities between them risks just making that claim. These various authors have models built on the same substratum, but they're using graphs in wildly different ways––is the collection meaningful (i.e. are there theories referencing graphs or tangles that you'd exclude)? Does your idea of a narrative-forming traversal mean compatible things among very different graphs (a graph of states, of texts, an abstract topological category...)?