Mechanisms in the Emergence of Data Science: A Comparative Study of Abstract Knowledge

• In Event: 497 - Regular Session. Sociology of Knowledge

Abstract

Universities’ reactions to the recent formation of “data science” as a field of professional expertise have produced a paradoxical situation: A coherent grammar of skills and analytical strategies gets superimposed on the separate scientific disciplines that have developed this grammar’s rules. How can such heterogenous academic underpinning sustain coherent practice? With a role structure similarly embedded in public, private and academic organizations, legal scholarship presents an effective analogy. A comparative research design is proposed that leverages this analogy and integrates evidence from complex citation networks and documentary data to resolve the puzzling tension between heterogenous organization of coherent knowledge. The results show that the mechanisms generally associated with producing the guiding theories of scientific fields neither organize legal nor data science expertise. Simulation tests demonstrate that in each field different processes—individual brokerage and institutionalized journals, respectively—solve these coordination problems. Both processes, however, function as instantiations of the same mechanism. Qualitative analyses of analytical practices in legal scholarship and data science reveal how researchers in both fields use formal abstractions to discover relationships between problems they are concerned with. Implications of such pragmatic formalism for research on state formation, labor markets and political economy are discussed.

Author

• Philipp Soeren Brandt, Columbia University