Ben Hammersley is having a discussion about ways of creating “emergent taxonomies”. This is very interesting stuff, not least because it represents a (reasonably) practical approach to the problem that most people simply on the web will not be invested or skilled enough to follow precisely the demands of a predetermined taxonomy. An emergent taxonomy (or categorisation, if you like) looks like a decent way of semantically linking fundamentally unstructured information.
Great. But I work for a publisher, and we produce fundamentally structured information. Not only that, but in a field which looks for fundamentally structured hierarchies of information. If we trusted to an emergent taxonomy, it might relate bags of information, but with no meaningful hierarchy, or a structure that looks rather too much like Borges’ famous joke taxonomy.
At the moment, technologically speaking, we could do either thing. But, as a practical way forward for the big purveyors of highly-ordered information (i.e. those who make money out of it), bet on taxonomies. They’re mostly already out there. The big challenges are:
1) Automating their application to pre-existing data
2) Linking the top-level taxonomies (e.g. cross-linking the art history taxonomy with the religious one, the UK tax taxonomy with the German one, and so on)
I think the first is more difficult, but it will happen, no question. I just hope that the rest of us (i.e. the Ben Hammersleys, Jim Leys, Shelley Powers, me in my private life) don’t just get left with the Borgesian category of “Other”, or even “Stray Dogs”.