Modal Fragility and Counterfactual Causation

Abstract One of today’s foremost counterfactual analyses draws many causal links where it shouldn’t—not only towards the future, but also towards the past. The root of these problems is that unlikely events are too modally fragile to ground reasonable causal claims. However, I will argue that these issues arise just as readily for likely events, spelling double trouble for this analysis. By recentring this approach around probabilities instead, causation tracks structure whose modal robustness justifies its name, sidestepping the old problems and the new.