Humans possess knowledge of causal systems with deep compositional structures. For example, we know that a good soccer team needs players to fill different roles, with each role demanding a configuration of skills from the player. These causal systems operate on multiple object types (player roles) that are defined by features within objects (skills). This study explores how human learners perform on novel causal learning problems in which they need to infer multiple object types in a bottom-up manner, using causal information as a cue for their existence. We model subjects' learning process with Bayesian models, drawing hypotheses from different spaces of logical expressions. We find that, although subjects could sometimes learn object types, they failed at tasks that require learning of more than one object type. Our result identifies the number of object type as a major obstacle for human acquisition of complex causal systems.

Authors:

Feng Cheng: New York University; Bob Rehder: New York University