Domains and Glue

First, think of the body of knowledge as a large amorphous blob. Then, as you identify parts of the blob that have structure, think of those parts as crystallizing, forming regular crystal domains. Almost always, there will be an irregular region that remains uncrystallized; it would be technically correct to call this region the matrix, but I prefer to think of it as glue. (A usage I'm sure I got from The TeXbook.)

There is a dynamical aspect as well. As the body of knowledge grows or changes, or as your understanding of it changes, new domains may appear, or old ones may readjust themselves.

This is all very vague, but I don't have any really good examples on hand. The best I can do is point to these essays themselves. As I add more of them, it becomes clearer and clearer what the correct categories are, but there are always some weird ones left over that don't really fit anywhere.

There are examples in the sciences, as well. As I understand it, over the past few decades what was once a set of isolated facts about particular systems has grown and crystallized into the field of chaotic dynamics. I'm sure there is some connection here to Kuhn's The Structure of Scientific Revolutions, but I haven't read it recently enough to say anything more.

As a meta-example, The Structure of Scientific Revolutions itself is part of the glue that holds the various scientific domains together. Although I know people try to make a whole domain (history of science) out of it, I've never seen anything to convince me that Kuhn didn't say everything there was to be said, and I don't think his book is a domain unto itself.

When thinking in terms of computers and data, I once decided that anything that has enough structure to be put into database tables is a domain, but I'm not sure that's quite the same thing.

As a final thought, it might be possible to classify glue into two types: permanent overarching glue, which contains information about the domains, and temporary glue, which is specific enough to belong in a domain, but for which no domain has yet been identified.