And I don't mean graphs in the sense of x and y axes, straight lines through points, Bell shaped curves, and such like.

Instead, I mean graphs in a (as odd as it may sound, because I'm not) mathematical sense, as a thing of nodes and vertices.

Or blobs and lines, more simply.

Why blobs and lines are so important
More here...
Before the web came hypertext, and before hypertext came mind maps.

Topic mapping is essentially the information architect's version of mind mapping, i.e. it's simply mind-mapping by another name. It's a very easy and useful brainstorming tool that can be used to find structure in what might otherwise appear to be an unstructured problem space. Use it for expressing ideas, seeking regularities, identifying organising principles and so on.

A mind map (qua topic map) is a form of semantic network, which even more simply is just a form of graph or network, i.e. a structure of nodes and links. You can put anything you like at the nodes of a network or topic map: attributes, concepts, actions, descriptors, emotions, judgements, qualities, and so forth. Equally, the links between nodes can asssume any kind of relation that you like: similarity, dissimilarity, distance, likeness, related-to-ness, and so forth. Generally speaking, the further the further a node is from another, the weaker the relation (whatever it may be) between them is.

Read more about graphs