Bowker, Geoffrey C., and Star, Susan Leigh. Sorting Things Out: Classification and Its Consequences. Cambridge, MA: The MIT Press, 1999. xii + 377 pp. $29.95, Paper-back $17.95.
Sorting Things Out is about the relationship between formal systems of knowledge representation-notably classifications-and the practices and situated experiences in which these systems are embedded. According to the authors, the book stands at the crossroads of the sociology of knowledge and technology, history, and information science.
A variety of classification schemes are discussed: WHO's International Classification of Disease (ICD), the Diagnostic Standards and Classification of Tuberculosis, the Classification and Nomenclature of Viruses, the Nursing Intervention Classification (produced by a group of nursing scientists at the University of Iowa), and the racial classifications that were employed in preapartheid and apartheid South Africa. The Diagnostic and Statistical Manual of Mental Disorders is not included.
The authors want to illuminate the moral and political dimensions of classification schemes. The DSM does not fit the bill because its social origins are ostensibly transparent, permitting a "direct read-off from political exigencies to disease categories." The ICD is a better choice because its "politics were quieter. . . [and] show the generalizability of the thesis. . . ." On the other hand, the "politically laden categories generated by the proapartheid government and its scientific apologists"-clearly another case of transparency-is included because they "balance" the ICD case (p. 324).
The book includes historical accounts of the classifications, how they were tailored to meet the interests and agendas of relevant organizations and institutions, and how they were influenced by the interaction between classifiers and the things classified-patients, bacteria, viruses, citizens. However, the authors' true subject is at a higher level of conceptualization, namely the point at which multiple classifications coalesce into "information infrastructures." These infrastructures are described as being "big, layered, and complex." They are "sunk into" social formations and technologies. They are continually changing, via processes of negotiation and resistance, and through adjustments to other elements within the respective mega-systems.
Infrastructures do not coincide with clearly defined systems of control. There is no master command, no monopoly of power. They are occupied by multiple groups of classifiers-diagnosticians, therapists, researchers, bureaucrats, health policy planners, etc.-and the groups are themselves internally heterogeneous. Consequently, classifications mean different things and produce different realities from one locale to another.
Infrastructures are likewise invisible and lose their transparency only when they begin to break down. According to the authors, this makes it difficult for us to visualize or describe these information systems and the structures that influence their design and utilization. Their solution to this problem is "long-term and detailed ethnographic and historical studies of information systems in use." The goal is to build an analytic vocabulary "that recognizes that the architecture of classification schemes is simultaneously a moral and an informatic one." The analytic vocabulary developed in Sorting Things Out consists largely of metaphors. These are said to "speak to the densely patterned interaction of infrastructures and the experience of living in the "classification society" (p.232). Here are two of them, "texture and torque": "[We] move on from exploring the seamless web of science and society, of nature and knowledge to an analysis of the information infrastructure that acts as matrix for the web. The web itself is textured in interesting ways by the available modes of information storage and transfer" (p. 193).
It is one thing to be ill and in the hospital with an indefinite release date. It is quite another when the date of release depends on one's ability to negotiate well with the physicians, their interpretation of the latest research, and the exigencies of public health forms and red tape. We call this agglomeration torque, "a twisting of time lines that pull at each other, and bend or twist both patient biography and the process of metrication. When all are aligned, there is no sense of torque or stress; when they pull against each other over a long period, a nightmare of texture emerges" (p. 27).
Sorting Things Out concludes with advice for "designers and users... of complex information spaces." It is "politically and ethically crucial to recognize the vital role of infrastructure in the "built moral environment." Classification schemes represent multiple constituencies. To do this, they must incorporate ambiguity. Give "presence to the voice of classifiers and their constituents.... (be) sensitive to the exclusions that are intrinsic to all classifications. More specifically, pay attention to who gets to determine what is 'other'" (pp. 324-326).
Allan Young, Ph.D.
McGill University; Montreal, Canada