I’ve recently begun a new project with Camilla Hrdy about government procurement as a local innovation incentive. Serendipitously, this has exposed me to literature around innovation clusters and spillover effects, as in this Fallah and Ibrahim literature review. This has been an “aha!” moment.
Innovation clusters are, in the literature, geographic places like Silicon Valley, Cambridge, Massachussetts, and other urban areas where there is a lot of R&D investment. Received wisdom is that these areas wind up driving the local economies in those areas by spillover effects, a market externality in which those beyond the intended beneficieries of an innovation benefit from it, normally through an informal knowledge transfer.
To economists in the 90’s, this was a significant and exceptional property of certain geographic places. To the digital native acclimated to Free Culture, this is a way of life. Spillovers are defined in terms of the intended boundaries of the recipients of innovative information. However, when there is no intended boundary you still get the spillover effect on innovation itself (however incomprehensible the incentives are to the 90’s economist). The Internet provides a virtual proximity that turns it into an innovation cluster. Advances in human computer interaction further enable this virtual proximity. We might say that Github, for example, is an innovation cluster with a higher degree of virtual proximity between its innovators within the larger virtual innovation cluster that includes SourceForge, Bitbucket, and everything else. (Considering software engineering as the particular case here.)
By binding together other innovation clusters, this virtual proximity leads to the innovation explosion we’ve seen in the past 10 or so years. “Everything changes so fast.” It’s true!
Outside of the software environment, we can point to other virtual innovation clusters such as Weird Twitter, where virtual proximity and spillover effects are used to innovate rapidly in humor.
The drive to open access academic research is due in part to an understanding of these spillover effects. You increase impact by encouraging spillover. I.e., you try to make waves. Academic research becomes more like speciality journalism in the sense that you try to break a story globally, not just to a particular academic community. The speed of innovation in such a dynamic environment is bewildering and perhaps the university tenure-based incentive system is not well designed to handle it, but nevertheless these are the times.
Jack Burris at the Berkeley D-Lab likes to say that the D-Lab is designed to support ‘collisions’ between researchers in differnet fields. “Spillovers” might be a term with more literature behind it. Indeed, interdisciplinarity needs to start with collisions or spillovers because that is what creates mixing between siloed innovation. I’ve heard that Soo and Carson’s paper about Clark Kerr as an industrial organizer explain some of the idiosyncracies of Berkeley in particular as an accumulation of silos.
Which explains the D-Lab’s current agenda as a mix of open source evangelism, reproducible research technology adoption, technical skills training for social scientists, and eshewer of disciplinary distinctions. If Berkeley’s success as a research institution depends on its being an effective innovation cluster, even within the larger innovation cluster that is the coast of Northern California, then it will need to increase the virtual proximity of its constituent innovators. Furthermore, this will expose non-local actors to spillovers from Berkeley, and perhaps Berkeley from spillovers from other institutions. This is of course a shift in degree, not kind, from the way the academic system already works in the economy. But what’s new is the use of disruptive infrastructure to accelerate the process.
This would all be wonderful if it were not also tilting towards a crisis, since its unclear how the human beings in the system are meant to adapt to these rapid changes. What is scholarship when the body of literature available on a particular topic is no longer strictly filtered by a hierarchical community of practice but rather is available to anybody with the (admittedly often specialized, but increasingly available) literacy? Is expertise just a matter of having the leisure and discretion to retweet the latest and greatest? Or do you make it by positioning yourself skillfully at the right point of the long tail? Or is this once-glorified social role of intellectual labor now just a perfunctory routine we can replace with ranks of amateurs?
To be a good expert is to be a good node. Not a central node, not a loud node, just a good node. This is humbling for experts, but these are the times.