Perhaps following up on a distributed conversation on citation indexing among academic blogs (me, David Brake, Seb Paquet), Thomson ISI has announced a new scholarly index of web publications, largely based on the CiteSeer project. I see this as a major positive move, in part because it means that publishing your work to the web no longer carries quite as many disincentives, from the perspective of the tenure process, as it has in the past.
Both Seb and David raise the issue that citation is a good measure of popularity, and that is helpful for some things and not for others. In general, this is a view that I am sympathetic to. Seb suggests that someone would do well to collect both from the core of the elite blogs and from the periphery of lesser-known ones.
I’ll fall back on a homology that I use fairly consistently: the city as an information-processing machine. We like to think that innovation happens at the city core, while the hinterlands remain a kind of social memory, a more conservative force. This is true to an extent, but it discounts the nature of the flow of innovation. Invention is more observable at the core of a city, because it is at this core where the greatest amount of fusion, translation, and reconfiguration takes place. The city center acts as a solvent, reducing the temporal and spatial barriers to interaction that permeate the rest of the society.
However, the hinterlands (and this is not a binary opposition: the “hinterlands” here means anything from Harlem to Peoria to Vladivostok) provide the framework for innovation, for the development of inventions into something more permanent, sustained, and ultimately more influential.
So, some kids come to town with their pants hanging a bit lower than usual. In the maelstrom of interactions in the city, this becomes a trope. It is reified and replicated on a small scale. Others — either for commercial gain or search of social currency — take these trends, concretize them into mass media depictions or commercial products, and redistribute them in a more or less stable way to the broader culture. And this broader culture provides the friction and isolation needed to allow ideas to mutate and incubate. The city, left to itself, eats itself. It needs to draw on young ideas from the surrounding hinterlands.
It’s both true that you need to pay attention to the concentration, and that you need to pay attention to the periphery: both are of value.
Now the question is whether concentrations of citations are really analogous to concentrations of communication and transportation that happen in cities. New York is the center of FedEx packages, telephone calls, and web links in the world. Does it follow that in-links indicate centers of invention? Clearly not.
The A-list bloggers represent the broadcasters of today, and as such bore me. That is, while every journalist is most interested in the handful of blogs that seem to be supplanting journalists, I find this to be the least interesting effect of blogging. Most interesting to me is the possibility that blogs are taking on more local seats of opinion and innovation leadership.
The process of drawing ideas and practices from the hinterlands and spewing back translations is repeated at every bracketed level–that is, it is fractal. I am talking not just about the world city (the concentration of between 8 and 20 global centers), but about neighborhoods and groups that draw on the isolated and process this into something that can be understood widely.
Katz & Lazardsfeld’s work on 2-step and multi-step flow got at some of this, perhaps. The interesting step is no so much from the mass to opinion leaders, but to and from opinion leaders and their constituency. This is still largely an invisible, interpersonal process, but I think that is changing with blogs, and I think that change will be much more interesting.
This is all, in some way, a defense for the Scholarati stuff I talked about below. What is interesting in the power law family of distributions of web links is that they seem to be self-similar at scale. I think looking at ways of bracketing off subnets is a valuable process. For me, this centers on how scholars use their blogs, but it could just as easily be Buffy fans or chicken farmers. The question is whether there is some relationship between a subnetwork and the composition of that network? Is there some network structure of hyperlinks (admittedly an estimate, but an available measure) that correlates to topical or other cohesiveness?