Rabbit Holes and Butterfly Effects: Narrative Probabilities and Climate Science

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 sequence 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 can start from the observation that climate change is an elusive phenomenon that unfolds on a global scale and whose local consequences may be hard to predict or ascribe to climate change with absolute certainty.This spatiotemporal distribution of climate change can give rise to manifestations that are far more surprising than the probabilistic link between Narrative Probabilities and Climate Science 2 smoking and lung cancer, both of which are locally observable events.In the science of complexity, the phrase "butterfly effect" (coined by meteorologist Edward Norton Lorenz) refers to how a seemingly minor disturbance within a complex system can have large-scale ramifications at a spatiotemporal remove.By extension, I use the term to describe the unanticipated consequences of a complex probabilistic phenomenon. 2 The butterfly metaphor comes from Lorenz's talk at the 1972 convention of the American Association for the Advancement of Science, where he asked: "Does the flap of a butterfly's wings in Brazil set off a tornado in Texas?" (qtd. in Hilborn 425).For an example that is directly relevant to the climate crisis, consider the devastating wildfires that swept through Australia in early 2020.A Nature article states that "human-induced climate change increased the risk of the weather conditions that drove the fires [in Australia] by at least 30%" (Phillips).This is a butterfly effect in that the causal history of these wildfires is complex and can ultimately be located far away from Australia, in the processes of industrialization and capitalism that, emerging in Western Europe at the turn of the eighteenth century, have become the main driver of greenhouse gas emissions and therefore climate change around the globe.It is only in probabilistic terms that we can say that climate change caused these particular wildfires.This probabilistic causation differs fundamentally from the notions of probability at the heart of narrative.This problem has been articulated with particular lucidity by Indian writer Amitav Ghosh in The Great Derangement (2016), which discusses the contemporary novel's confrontation with climate change.For Ghosh, the realist novel struggles to account for the catastrophic consequences of rising sea levels or global temperatures: "To introduce such happenings into a novel is in fact to court eviction from the mansion in which serious fiction has long been in residence," he contends (24).Ghosh's claims have encountered a great deal of opposition in