Fred and I were talking the other day about group cognition in the blogosphere – a topic which yes, is starting to interest me. It interests me because I believe that multiple heads are better than one, and that a community can make a more thorough contribution to a topic than any one individual. Fred and I discussed this in our paper Lyceum: A Blogsphere for Library Reference in the context of library reference, that multiple librarians can provide a broader and deeper answer to a question than could any one individual librarian. But it has since struck me that this is also the model of employed by the scientific community: one person working in an area is an eccentric who never gets published, a whole community working in an area has journals and conferences and whatnot. Science is fundamentally a social enterprise – just ask Bruno Latour.

Multiple heads are better than one, but I’ve been pretty vocal here about the downside of collective decision-making. So whose heads do you harness? I would suggest those of experts. You want broad and deep answers to reference questions? Get a bunch of reference librarians collaborating. You want to collect quality news coverage, Fred? Get a bunch of journalists collaborating, or a bunch of historians: the former can tell you what makes for good coverage, the latter can tell you what will probably be useful to know 100 years from now.

I would be remiss here if I didn’t mention my former advisee Marianne Gouge‘s Masters paper, entitled Blogs as a Means of Preservation Selection for the World Wide Web. Marianne asked the research question, can blog aggregators be used to select content of archival quality? What she found was that different aggregators selected content that was of archival quality according to one or a few criteria, but not according to all of the criteria she used.

So – and now I’m going beyond Marianne’s conclusions – if you wanted to use aggregators as archival filters, you would have to (A) use several aggregators that do well by different criteria and hope that they have collected some overlapping content, or (B) use aggregators as the “first line of defense” to filter content and then rely on human expertise to vette the content collected automatically. Choose a different set of criteria and you can presumably select quality content in an different arena. But since aggregators aren’t doing anything except skimming content from a subset of the blogosphere, the blogosphere can’t be used as a source of quality content, except (A) by using overlapping sampling strategies, or (B) by requiring human filtering. And so we’re back to requiring expertise, either on the front- or the back-end: you need expertise either to determine your selection criteria or to vette what’s retrieved from the pool of collectively-created content.