A long essay about structures in tags networks revealing cognitive processing mechanisms. Not quite - software can do this kind of semantic association, and it doesn't necessarily reveal much about the brain (!).
But the post does obliquely raise one very good point - its incredibly cheap to distribute the cost of categorizing semantic information massively among a bunch of humans, because coordination and transaction costs to doing so have been vaporized.
Which begs the question - how does this impact the dynamics of competition on the Net?
Answer: interconnected (ie, compatible between apps) tagging is a key leverage point which more and more firms will start to exploit - because, like open-source, it creates absolutely huge economies of scale. In fact, I think we'll see many more kinds of massively distributed coordination mechanisms, not just tags (viz all the talking I was doing about coordination machines a few months back).