When watching my thoughts, the structure of my knowledge is more a garbled web of associations than a systematic table of contents. This garbled web is self-organizing. While its formation is mostly out of my control, exploring it isn’t. Yet that seems to be the nature of learning, understanding, and creativity. We relate ideas in non-obvious ways. This is something I’ve noticed, but later found is a real theory of cognition called Relational Frame Theory.

Relational Frame Theory describes (dryly) the mechanism to building webs of relationships between pieces of information. If you lived in a vacuum and only learned about apples, you’d only understand and draw connections based on your experiences with apples. But in the real world, apples connect with many things. Apples are foods, fruits, trees, an industry, a company, etc. This means as you learn, the number of connections between concepts grows exponentially. Ultimately everything has a chain of connection to everything else.

Thinking through this theory, our minds seem to have two jobs. Classifying perceptions into objects and generating deeply intricate relationships between those objects. We build a graph of objects and relationships. Increasing intelligence is remembering more things and making more connections between those things.

This is interesting as its a stark contrast between how we typically represent knowledge. Books have chapters, sections, sub-sections, sub-sub-sections, ad infinitum. This is linear and hierarchical. Subsection 1.4.5 may have a relationship with subsection 5.2.1, or with subsection 9.6.3 from a completely different book of a different field, but it’ll never be found. There’s too many potential relationships, limited short term memory, and insufficient compute time to search the full space by just traversing through the book linearly.

If our brains are a jumble of associations, surely linear sequences of monochrome characters isn’t the best medium for the job. How could it be improved?