Imitation research has been hindered by (1) overly molecular analyses of be
haviour that ignore hierarchical structure, and (2) attempts to disqualify
observational evidence. Program-level imitation is one of a range of cognit
ive skills for scheduling efficient novel behaviour, in particular enabling
an individual to purloin the organization of another's behaviour for its o
wn. To do so, the individual must perceive the underlying hierarchical sche
dule of the fluid action it observes and must understand the local function
s of subroutines within the overall goal-directed process. Action-level imi
tation, copying strings of actions linearly without any such understanding,
is less valuable for acquiring complex behaviour and may often have other,
social functions. At present, we lack a mechanistic understanding of the a
bilities underlying program-level imitation that make it possible for the u
nderlying structure of complex actions to be dissected visually and recreat
ed in behaviour.