A toolbox of methods for probabilistic inference

We propose that probabilistic inference is supported by a mental toolbox that includes sampling and symmetry-based reasoning in addition to several other methods. To flesh out this claim we consider a spatial reasoning task and describe a number of different methods for solving the task. Several recent process-level accounts of probabilistic inference have focused on sampling, but we present an experiment that suggests that sampling alone does not adequately capture people’s inferences about our task.