Knowledge Infrastructures and the Two Cultures

C.P. Snow talked in the 1950s about the two cultures – the natural sciences and the humanities – which formed their own separate universes of discourse. There’s a certain institutional truth to this: few higher learning institutions (

Santa Clara

is an exception) have a commitment to breaking down campus divides for their undergraduates. I was reminded of this when giving a talk at Rice last week (webcast here: http://webcast.rice.edu/webcast.php?action=details&event=1375; I’ll write about forgetting, my topic there, when next I remember): folks there talked about the lecture series I was part of being aimed at bringing together the North and South parts of campus. At UIUC this is translated as North of Green Street and South of Green Street – Champaign Urbana being quite the model for seggregation, with very little crossover between the black (well south of Green Street) and the white communities. 

Now the ‘two cultures’ argument does not hold in some fundamental ways. To get a tad technical social scientesque, the ways it doesn’t hold follow from the arguments of Latour’s We Have Never Been Modern. They may consider they are in separate camps, but in fact there is much more traffic than most imagine – indeed the opposing camps could not exist without this traffic. A starter book for exploring this is Holton’s Thematic Origins of Scientific Thought – he points out that the theme of discontinuity was being explored on both sides in a strangely united fashion in the early years of the twentieth century: from the growth of non-linear mathematics to the rise of quantum mechanics to historiographical writings about revolution (discontinuous historical change). More recently, representing knowledge in a tree form has been countered by molecular biologists (lateral gene transfer upsets the ‘tree of life’) and critical theorists (rhizomatic descriptions of knowledge upset the ‘tree of knowledge’). 

It’s a whole other argument why this is true (curious coincidences tend to have very material roots), but for now let’s segue into knowledge infrastructures. First a definition. A knowledge infrastructure (KI) is the set of practices, institutions, and technologies that undergird knowledge production (it’s a more useful term than information infrastructure since it is deliberately more extensive). We’re doing a lot of KI building right now – two significant movements are the very high end work being done through the Office for Cyberinfrastructure to enable new forms of scientific knowledge production, and the potentially transformative work of the Hewlett Foundation in building ‘open participatory learning infrastructures’ for education. And the world of folks who do this building has historically been separated into two separate cultures (supercomputer folks for cyberinfrastructure or mesh network developers for the rural education opposed to working scientists and local teachers on the other). Just as in Snow’s case, the cultures seem to be divided. Typically, when a cyberinfrastructure fails (a relatively common occurrence), it’s seen in my world as because of a lack of attention to organizational, social and cultural factors. Equally, those who work on those very factors tend to be technology-shy – as I found out a few years back when being part of a failed committee at UCSD to address new publication forms in assessment of faculty promotion cases. 

And again as in Snow’s case, there is a lot of common cause between the two camps. The point is that you cannot neatly separate off technical developments from social, political and organizational change. We can’t change the nature of scientific knowledge production without working together across the boundaries. In Hughes’ terms, the ‘reverse salients’ might be organizational (you don’t get rewarded for interdisciplinary work) or technical (it’s very hard to merge databases from different disciplines with different information standards). Often now it’s the computer systems designers who do the heavy lifting here – they are forced to negotiate with their ‘domain’ scienstists, and as a result they are becoming much more sensitive to reward structures, institutional constraints and so forth. But the most important point is that a ‘technical problem’ can have a ‘social solution’ and vice versa. Thus you can make databases seamless either by doing some technical magic on computer ontologies, or by getting scientific groups to come together and agrree to a common set of standards. Or you can solve the disciplinary camps problem either by visionary university leadership or by the technical fix of a ‘boundary infrastructure’ which allows each community to continue to operate as they did but provides full access to the relevant knowledge and practices of other disciplines.   --Geof Bowker

Comments

[Sally Lehrman]

Your post is an "aha!" moment for me as I think about the divide between social scientists and molecular biologists on the subjects of population genetics, biogeographical ancestry and race. Both sides are committed to a certain way of seeing the issues -- and building a knowledge infrastructure accordingly. A number of joint conferences have failed to bridge the divide. In this case of "two cultures," are there solutions suggested by a reading of Holton or Latour?