Have you just stumbled into some archiac throwback maths lesson of some peculiarly rebarbative kind?
GRAPHS?, I hear you say with disgust / revulsion / incomprehension.
Bear with me.
First things first: we're not talking about graphs with x and y axes, bell-shaped curves or any of that stuff.
We're using the term graph in its wholly mathematical sense - does that make things worse? - of a structure of nodes and links.
The interrelationship digraph
Imagine that you've done your affinity diagramming. You will now probably want to try and formalise the messy, intuitive, perceptual affinity groupings that you've created.
How do you do this? The common way, as already discussed, is to create a form of visual representation - the affinity diagram - that expresses the relations between the affinity sets that your affinity diagramming has revealed.
Or in other words, the common way to do this is to create an interrelationship digraph - because affinity diagrams are a nothing more than a particualr kind of graph.
And a convenient shorthand way of refering to an interrelationship digraph (which is a bit of a clunky term, let's face it) is as a network.
Bear with me.
The nature of relationship
Think about this. What kind of relations are expressed by an affinity diagram?
Or, to turn the question around, what specifies why one item or affinity set should be grouped with another? What is the relation between them? Is it hierarchical or associative? Is it one of logical order, or natural similarity?
Thse questions are at the centre of what defines a network. Because a network, basically, is just nodes (or blobs) connected by vertices (or lines). And there is nothing intrinsic to a network that renders it explanatory. The explanatory power of a network derives from the interpretation that is imposed upon it.
What kinds of relation are expressed by an affinity diagram? Anything that you like! The convention is that linkage indicates some measure (along some value scale) of similarity, and that increasing distance between nodes indicates increasing dissimilarity.
Blobs and lines?
There we go, nice and easy. But the devil is always in the detail, and so it is with the network.
The nodes (blobs) of a network can represent anything you like: concepts, ideas, words, values, locations, people, etc. Similarly, the vertices (lines) can represent any form of relationship between these items that you like: hierarchy, similarity, distance, resemblance.