Bowker's "The History of Information Infrastructure"

Bowker, Geof­frey C. “The His­tory of Infor­ma­tion Infra­struc­tures: The Case of the Inter­na­tional Clas­si­fi­ca­tion of Dis­eases.” Infor­ma­tion Pro­cess­ing & Man­age­ment 32, no. 1 (Jan­u­ary 1996): 49–61.

This arti­cle is a lead up to a few chap­ters that ended up in Bowker and Star’s Sort­ing Things Out. I found my read­ing copy in a book called His­tor­i­cal Stud­ies in Infor­ma­tion Sci­ence.

There are sev­eral pay­offs for read­ing this arti­cle. One is that this is an early piece that shows how Bowker was think­ing about writ­ing his­to­ries of infor­ma­tion infra­struc­ture. Because the arti­cle is in a bit of a pri­mor­dial state, read­ing it helps to see how Bowker was think­ing dur­ing some of his later work.

This arti­cle also describes sev­eral key terms and con­cepts that are enor­mously help­ful for ana­lyz­ing infor­ma­tion infra­struc­ture: imbri­ca­tion, boot­strap­ping, figure/ground, and a short dis­cus­sion of infra­struc­tural inver­sion. “Imbri­ca­tion” is an ana­lytic con­cept that helps to ask­ing ques­tions about his­tor­i­cal data. A good exam­ple from my own work: I’m cur­rently work­ing back­wards through meth­ods of com­puter pro­gram­ming to see which pro­gram­ming con­cepts, clas­si­fi­ca­tions, and tech­niques are passed for­ward to new lan­guages and stan­dards. Com­puter lan­guages (unsur­pris­ingly) bor­row lots from the work that peo­ple have already done. Con­cerns, val­ues, and pol­i­tics of the past are passed for­ward with those concepts.

“Boot­strap­ping” is the idea that infra­struc­ture must already exist in order to exist. Another way to say this is that infra­struc­ture forms an impor­tant part of social con­text. To cre­ate an infra­struc­ture as part of social con­text, peo­ple must already be work­ing with the imag­i­nary con­cepts of that con­text. This becomes par­tic­u­larly messy as an infra­struc­ture is used across large spans of time and space, yet its orig­i­nal design is context-dependent and built with the knowl­edge of the cul­tures that insti­gated the project. So to boot­strap is to build the exist­ing assump­tions and pol­i­tics about what infra­struc­ture is use­ful for into infra­struc­ture while real­iz­ing that the design of infra­struc­ture will simul­ta­ne­ously build those assump­tions and pol­i­tics into other times and spaces as the infra­struc­ture is adopted else­where. An exam­ple might be use­ful here. Bowker’s arti­cle deals mostly with the Inter­na­tional Clas­si­fi­ca­tion of Dis­eases (ICD), which has a rich clas­si­fi­ca­tion sys­tem for cat­e­go­riz­ing death. These clas­si­fi­ca­tions reflect the time that they were cre­ated (more peo­ple dying of tuber­cu­lo­sis at the begin­ning of the 20th cen­tury; fewer deaths can be attrib­uted to mul­ti­ple causes) and the pol­i­tics (why exactly does a state want to know that peo­ple are dying in “acci­dents from trans­port, acci­dents in mines and quar­ries, agri­cul­tural and forestry acci­dents, or acci­dents due to machin­ery” and why is it ok to group those together?).

“Figure/ground” builds on the idea of boot­strap­ping. It’s a way of think­ing about infra­struc­ture that doesn’t see it solely as the prod­uct of human con­struc­tion or as a con­text shap­ing human activ­ity. Infra­struc­ture is both at the same time. When a clas­si­fi­ca­tion becomes a part of an infra­struc­ture, for exam­ple, it is embed­ded by some­one apply­ing agency.That agent is influ­enced by the larger assump­tions of the larger infra­struc­ture. That is, new clas­si­fi­ca­tions and stan­dards only make sense within the larger ecol­ogy of exist­ing infra­struc­ture. That clas­si­fi­ca­tion then simul­ta­ne­ously becomes part of the ecol­ogy for think­ing about the infra­struc­ture. Fur­ther devel­op­ment of infra­struc­tures will there­after have to fight with the orig­i­nal stan­dards and classifications.

Figure/ground is impor­tant for think­ing broadly about infra­struc­ture. It’s not that some­one can’t remove a clas­si­fi­ca­tion such as “died from tuber­cu­lo­sis,” although this could become dif­fi­cult as well if the stan­dards have become widely mate­ri­ally and socially enforced (think elec­tri­cal power lines, pipe fit­tings, or vot­ing procedures) it’s more that the entire infra­struc­ture was con­cep­tu­ally estab­lished as a whole with that clas­si­fi­ca­tion as a part of it. If you’re famil­iar with the parole/langue idea from Saus­sure, figure/ground is an elab­o­ra­tion that has been fil­tered through infra­struc­tural theory.

Bowker calls the sum of these ana­lytic meth­ods infra­struc­tural inver­sion, a con­cept with con­tributes to a good part of his work.

So noth­ing new for infra­struc­tural the­o­rists, but a good break­down and dis­cus­sion of key con­cepts. The arti­cle serves as primer/recap  for those inter­ested in think­ing the­o­ret­i­cally about infrastructure.