Folksonomy, Peirce, AI, and Collective Intelligence plus Interfaces

Yesterday in the Interface Seminar at the Franklin Humanities Institute at Duke Harry Halpin and Paolo Mangiafico did a great presentation on collective intelligence and collaborative knowledge-building websites. It was not just a crash course in Web 2.0 but a theoretical and philosophical interpretation of a history of post-Frege analytic philosophy, post-cybernetic debates around AI, structural linguistics, taxonomy, folksonomy, and then a tour of some incredible collaborative sites, like the mashup of Craig's list real estate offerings and Google map or new library sites. (I'll try to write more about those another time.)

For now, I want to follow up on some extremely interesting ideas that Harry Halpin presented. He argued that folksonomy follows the "long tail" model of distribution, where the significant majority of users tag websites similarly or even identically and then there is a very long tail of users who each come up with distinctive tags. This made me think about collective intelligence, representation, and consensus. If in the social logic of the internet, users feel represented even when they are not in the majority, if they feel comfortable with representation somewhere way out there on that long tail (i.e. where customization and nonconformity become cherished norms of the internet as opposed to, say, representative democracy where customation and nonconformity become homogenized into a representational "vote" where power to make decisions is given to an elected official or body), then it is possible that it is precisely the long tail that promotes enfranchisement, not consensus. That is fascinating to think about for social theory and political theory. Let's go further. If AI is based on a rigorous dis-ambiguation (ie rendering in a clear and agreed upon symbol-system ambiguities of language and other forms of "fuzziness"), then Harry's suggestion that folksonomy might even replace more rigorous taxonomic systems may well rest precisely on the long tail. I'm making an assumption here and it may be too big of one (I need to think it through more carefully): I'm assuming that the long tail of "misfit choice" correlates with the fuzziness of what has to be discarded when real-world language is rendered in a rigrous, non-ambiguous (but ultimately disappointing) taxonomic system. Hmmmmm.

Okay, since I'm freefloating, I'm going to go elsewhere, back to one of the heroes of my youth, Charles S. Peirce. Peirce, almost heartbreakingly, believed that someday we would arrive (via pragmatacism, not pragmatism) at a community of thinkers, at collective intelligence. In his philosophical works, to my mind, he's never particularly convincing about how that communal apotheosis will occur. But there is a section in his geodesic theories (that preceeded just about everyone else's and weren't found in the basement of Harvard till around 1970, at which time they revolutionized thinking on geodesics) where he describes the nodes that connect the parts of a geodesic dome and uses an analogy of community that is unforgettable: that these nodes take the collective weight of the dome and distribute it across the entire network of other nodes through an intricate system of checks and balances that account not only for the weight of the materials used, gravity, and so forth, but also for contingency: rain, snow, shifting ground, tremors, water table changes, temperature (and material changes as a result), and even accidents such as a large object [horse and buggy] crashing into one part of the dome, etc. All the subtle weight shiftings and dependencies and realignments and reactions and counter-reactions counterbalanced so, in the end, the structure, however shaken, still stands. He says that is collective intelligence. The community of ideas.