Comments on Discussions of Entropy and Art in "Leonardo" in 1973

Abstract

Arnheim's book 'Entropy and Art: An Essay on Disorder and Order' [1] has stirred discussions in Leonardo [2-9] that appear to require considerable understanding of physics before one may grasp the criticisms and arguments that have been raised. Many of the confusions in the book and the justified criticisms arise from an incorrect use of the word entropy as related to order and meaning. At best, the concept of entropy is unfortunately awkward and challenging to the student of physics or engineering [10]. It is misused in many books and I am sorry to say that the trend is compounded by Arnheim. I assert that the concept of entropy has very limited significance or metaphorical power outside of the energy considerations that brought it into existence. Indeed, as a figment of our imaginations and as a quantity that can only be inferred from several direct measurements (certainly not observed by inspection), entropy is useful only when carefully applied. In science, regularity is observed to many levels of penetration. The quantification and specific identification at each level and between levels must be unambiguous to be meaningful. My experience in the arts, in contrast, has led me to accept considerable ambiguity in addition to often superficial regularity. Many of the specific topics chosen for discussion in the book, such as entropy, equilibrium, diffusion, order and simplicity, require careful definitions and consistent use. The apparent contradictions arise from inconsistencies of meaning between the scientific use and a more common or free use of these words; they are strongly context dependent and the context must be consistent and clear. Further, an inaccurate concept of cosmic tendencies leads to a view in contradiction with scientific observations; most natural events tend toward diffusion, decay and homogeneity, and it is only with the expenditure of extraordinary amounts of energy that organic systems, man in particular, rescue some events so that they result in highly differentiated and so-called 'ordered' products. At the root of the problem is the very concept of