A RECENT book by Hermann Koller' puts forward a radically new and challenging account of the origin and development of the concept of imitation (mimesis) among the Greeks: challenging because it turns our usual understanding of the word and its history almost exactly upside down. According to the prevailing view, mimesis-or mimeisthai-began by meaning imitation, Nachahmung, and then took on other senses by extension or adaptation, for example, in the Poetics, where it clearly denotes something more than a mere copying of nature. Koller maintains, on the contrary, (1) that the original sense was not imitation, but Darstellung, Ausdruck(sform}, Formwerdung des Seelischen ;2 (2) that the ambit of the word was originally limited to music and dancing, denoting the expressive power of mousike in its primeval unity;;3 and (3) that the meaning imitation is a later development, a watered-down application of the idea to fields (e.g., painting) where it did not properly belong.4 The primeval idea was expanded by Damon and the fifth-century Pythagoreans (identity unspecified, as is their relation to Damon) into a grandiose Ausdrucks- or Ethoslehre, embracing the whole range of emotional, therapeutic (cathartic), and educational uses of music. This PythagoreanDamonian doctrine was partly adopted, partly adapted and distorted, by Plato and Aristotle, and provides the necessary background not only for their discussions of mime'sis but for the conception of music which was dominant throughout antiquity. I am far from wishing to challenge everything in Koller's book. It is full of interest and sets off in bold relief the crucial importance of Pythagoreanism for the Greek view of music, not that that importance was unrecognized heretofore. But much of Koller's material has relatively little to do with the concept of imitation per se; and at a number of points it seems to me that he has seriously misinterpreted the evidence, both for Plato and for the period before Plato, while other evidence of at least equal importance is neglected. Thus, following and developing a lead suggested by Deiters and Schafke,5 Koller relies heavily upon the second book of Aristides Quintilianus On -Music as a prime source for Damon and the fifth-century Pythagoreans, while on the other hand he considers only a small fraction of the instances of mimeisthai and mime'sis in actual fifthcentury authors. The result is an inter-