The word “imitation” has seen its meaning change many times over the years, a process that has narrowed down its scope onto an ever-decreasing range of referents. Most of these revisions have been driven by debate in the literature concerning nonhuman animals (hereafter, “animals”), but may have implications for those studying imitation in humans, especially at preverbal stages where many of the same issues occur. Before Thorndike's (1898) definition “learning to do an act from seeing it done,” imitation seemed not even to require a causal role for observation – and this sense is preserved in everyday talk, when people say, for instance, “this is a case of nature imitating art.” More recently, the notion of stimulus enhancement was introduced (Spence, 1937): seeing some act done in a particular place, or to some particular object, has the effect of increasing the observer's probability of going to that place or interacting with that object. This, as Spence noted, would make subsequent individual learning very much more efficient, and many behaviors that were formerly (and are still popularly) considered imitation were thereby explained away – as trial and error learning aided by social circumstances (Galef, 1988). Tomasello further drew attention to the possibility of learning about the physical situation as an indirect consequence of another's behavior, introducing the concept of emulation (Tomasello, 1990).