The Devil’s Software

Robert Lake levels a scathing postpositivist criticism of GIS, as he sees it in the early 1990s, as being fundamentally ethically flawed.  His major ethical sticking point is that the underlying positivist data and analysis models of GIS by necessity objectify the subjects of research and are unable “to comprehend and respect the subjective differences among the individuals who constitute the irreducible data points at the base of the GIS edifice.”  Lake makes a note of prioritizing these deontological concerns over the consequentialist ethical considerations of the field of GIS and planning/applied geography more broadly.  Thus, the “internal correctives” to adopt ethical codes among GIS practitioners is meaningless to Lake, who calls it “insufficient on ethical grounds if it focuses exclusively on the ends to the exclusion of the means”.

Lake’s argument is grounded in the more general postmodern imperative, which holds that positivist (or even quantitative) tools such as GIS must be eliminated, or at the very least, reconstituted from the ground up.  Not even the most benevolent use of GIS can be tolerated, because of the deontological abhorrence—the mortal sin—of representing people as undifferentiated digital objects without any consideration of positionality or subjectivity.  Lake hammers at this singular point repeatedly throughout the article.

I concede that it is vitally important to identify and account for the assumptions and values implicit in any analytical tool, including GIS.  That being said, I find Lake’s argument overly alarmist, generally unconvincing, and even potentially harmful for several reasons.  First, the argument’s unyielding deontological edicts ignore potential applications of GIS in purely physical domains of geography; and even in human geography, there is nothing that precludes researchers from tying back the results of GIS analysis to more qualitative and critical considerations.  Lake dismisses the potential of participatory GIS—an incredibly pertinent and empowering field today—as capable of nothing greater than token consultation.  Second, taken as a wider postmodern assault on abstracted, quantitative record-keeping, Lake’s argument quickly becomes dubious and unwieldy: should we also tear down the subject-object dualistic institutions of paper filing systems and library catalogues?  Finally, I believe that Lake’s appeal for geographers to refrain from adopting GIS in their research is a serious mistake.  Though harping on the deontological ethics may convince academics to avoid GIS applications, it certainly won’t dissuade the militaries, governments and corporations who are waiting in the wings to use GIS for unequivocally evil ends, to say nothing of GIS’s problematic means.  Without a critical and constructive GIScience approach that actually engages with GIS instead of talking past it, there will be nothing standing in the way of these other potentially oppressive actors—now there’s a consequentialist argument worth heeding!


Comments are closed.