Abstract

Arguments from inductive risk (AIRs) have widely been accepted as demonstrating that non-epistemic (e.g., social or ethical) values must or ought to play a role at core stages of scientific investigation. However, these classic versions of the AIR focus on only two options for treating hypotheses: accept or reject. In practice, scientists often exercise a third option: formally suspend judgement on the hypothesis. My focus is the “generalization” of the classic AIR to accommodate this third option. I argue that an adequately generalized AIR requires more careful distinctions between the epistemic (truth-oriented) and pragmatic (action-oriented) dimensions of trade-offs involved. More specifically, I distinguish between an epistemic error, failed reticence, as distinct from a pragmatic error, dangerous inaction, both associated with increased rates of suspension of judgement. Throughout, I use the case of scientists refraining from reporting numerical values for sea level rise from disintegration of the West Antarctic Ice Sheet in the Intergovernmental Panel on Climate Change (IPCC) fourth assessment report as an example where trade-offs involving these errors must be determined. I argue that applying the epistemic/pragmatic distinction to this case allows for the beginnings of a framework for evaluating decisions to suspend judgement and clarifies differing roles for scientists and policymakers in cases of uncertainty in policy-relevant science. The overarching aim of the paper is to provide considerations that enable more rational non-epistemic value influences on the selection of evidential standards in climate science.