Computational humanities analysis depends on biased choices about what counts as data

Evidence

Computers “cannot tolerate the ambiguity typical of humanities texts and interpretative methods,” argues Johanna Drucker, and thus, “what is considered data—that is, what is available for analysis—is as substantive a consideration as what is revealed by its analysis.” Consequently, humanists who work with digital tools are challenged to “make explicit many of the premises on which those understandings are based in order to make them operative in computational environments.” Without careful consideration, digital humanists can perpetuate a bias of the empirical sciences, in which data are held to be mere representations of preexisting facts. On the contrary, “humanistic inquiry acknowledges the situated, partial, and constitutive character of knowledge production.” — Andrew deWaard, Derivative Media, 325

Categorization

• QUESTION: The ‘raw data is an oxymoron’ point in 324 and Drucker here could ground a critique of the author’s own database methods in chapters 5-7 — worth checking how he handles his own data selection bias.