Rabbit Holes and Butterfly Effects: Narrative Probabilities and Climate Science

Abstract

Probability is a central concept in scientific models of causation. When I say, for instance, that ‘smoking causes lung cancer,’ I am not implying that smoking will necessarily result in lung cancer; I am only suggesting a strong (causal) relationship between smoking habits and the incidence of lung cancer. That relationship is probabilistic and based on statistical models which help scientists distinguish merely accidental outcomes from causal linkage. In this article, I focus on how this kind of probabilistic relation brought into view by scientific models puts pressure on the ‘folk’ understanding of causation that underlies storytelling. Probability also plays an important role in narrative: when readers or viewers parse a se-quence of narrated actions, their interpretation will build on assumptions about both causation (action 1 led to action 2) and probability (how likely action 2 is as an outcome of action 1). Crucially, however, narrative probability reflects cultural expectations surrounding human behavior, not statistical regularities. By contrast, narrative engagements with complex phenomena (particularly, in this article, climate change) call for new ways of thinking about narrative causation—ways that approximate the probabilistic understanding of statistical models. We