Social construction of algorithmic success: between good science and political feasibility in marine conservation planning.

Abstract

In the decade between the mid-1980s and mid-1990s, critical voices within the conservation biology community argued that site selection for protected areas was most often done in a way that was unscientific. Conservation practitioners, many of whom became acutely aware of the constraints of the policy world through direct participation, believed that they needed to think pragmatically about establishing a scientific basis for the design of protected areas. Some of the conservation practitioners came to see rationalistic tools such as optimization algorithms embedded within decision-support systems as means of reconciling social, economic, and environmental interests. This paper recapitulates the history of the first significant policy initiative that purported to use algorithmic decision support software, MARXAN, by interweaving environmental history, history of computing, and history of science. Specifically, it is a historical reconstruction of the use of MARXAN in its first large-scale conservation policy project: a rezoning of Australia's Great Barrier Reef that took place between 1998 and 2004. This paper asks: how exactly was MARXAN used in the conservation policy planning initiative? And, what role did MARXAN play in narratives about the success of the policy initiative? I argue that in Australian case, it was the commitment to political value of democratic deliberation and not the allure of algorithmic objectivity that stood behind what was by many considered an agenda-setting marine conservation policy. These findings add support to the growing consensus in critical algorithmic studies against algorithmic determinism by situating the agency of the users of MARXAN within a larger context of a "drama" as reported (Hilgartner in Science on Stage: Expert Advice as Public Drama. Stanford University Press, Stanford, 2000) of science advice.