The meaning of ‘melancholia’ in classical antiquity is opaque and has little in common with 20th-century psychiatric usage (Drabkin, 1955; Heiberg, 1927). At that time, melancholia and mania were not polar opposites (i.e. one was not defined as having opposite features to the other). Melancholia was defined in terms of overt behavioural features such as decreased motility, and morosity (Roccatagliata, 1973; Simon, 1978). Hence, in medical usage, ‘melancholia’ referred to a subtype of mania and named, in general, states of reduced behavioural output. These included disorders that might “exhibit depressed, agitated, hallucinatory, paranoid and even demented states … the ancient diagnosis of melancholy has no correct analogue in modern psychiatric practice …” (Siegel, 1973, p. 274).