Robert Frühstückl

This paper introduces a novel framework for causal selection based on an analysis of different ways in which causes interact. Some causal interactions function to enable the operation of a mechanism, while others modify the behavior or outcome of that mechanism, allowing fine-grained descriptions of causal relationships. Distinguishing between enabling and modifying makes it possible to separate distinct causal functions that are wrongly grouped into an undifferentiated category of background conditions. Drawing on case studies from ecology, the framework offers new insights into why some factors should be cited in explanations while others remain implicit, despite being causally indispensable.

This paper introduces a novel framework for causal selection based on an analysis of different ways in which causes interact. Some causal interactions function to enable the operation of a mechanism, while others modify the behavior or outcome of that mechanism, allowing fine-grained descriptions of causal relationships. Distinguishing between enabling and modifying makes it possible to separate distinct causal functions that are wrongly grouped into an undifferentiated category of background conditions. Drawing on case studies from ecology, the framework offers new insights into why some factors should be cited in explanations while others remain implicit, despite being causally indispensable.

In current philosophy of science, extrapolation is seen as an inference from a study to a distinct target system of interest. The reliability of such an inference is generally thought to depend on the extent to which study and target are similar in relevant respects, which is especially problematic when they are heterogeneous. This paper argues that this understanding is underdeveloped when applied to extrapolation in ecology. Extrapolation in ecology is not always well characterized as an inference from a model to a distinct target but often includes inferences from small-scale experimental systems to large-scale processes in nature, i.e., inferences across spatiotemporal scales. For this reason, I introduce a distinction between compositional and spatiotemporal variability. Whereas the former describes differences in entities and causal factors between model and target, the latter refers to the variability of a system over space and time. The central claim of this paper is that our understanding of heterogeneity needs to be expanded to explicitly include spatiotemporal variability and its effects on extrapolation across spatiotemporal scales.