Hume scholars have long disputed how we should understand his famous "two definitions" of causation. A serious problem with existing accounts is that they fit uneasily with Hume's claim in the Treatise that the two definitions correspond to causation considered separately as a "natural" and as a "philosophical" relation. This paper advances a new interpretation of the two definitions, according to which they represent an account of two different psychological mechanisms that generate causal judgments. This interpretation is fully consistent with Hume's claim that the two definitions map onto his distinction between natural and philosophical relations, once that distinction is itself properly understood