This paper reassesses the syntactic and semantic analysis of unergative verbs with inanimate subjects. Two main problems are present: First, what component of meaning predicts their unergativity? Contra Reinhart 2002, this paper argues that inanimate unergatives involve a causal implication, hence their subject merges externally without further ado. Second, a subset of inanimate unergatives exhibits transitive alternates (e.g. The doorbell buzzed / The postman buzzed the doorbell). How is the alternation formed? This paper shows that the process involved here is a productive, non-derivational pattern, which has been independently attested.