Ancient Automata and Mechanical Explanation

Abstract

Many twentieth century scholars describe ancient natural philosophy in terms of a dichotomy between so-called 'mechanistic' and teleological approaches to explanation. The atomists in particular are taken as representative of the former. However, scholars have pointed to some difficulties with this practice;' and different things are meant by the claim that there are 'mechanistic conceptions' or 'mechanical explanations' of the natural world in early natural philosophy. Often the term is taken to refer to the presence of some specific feature in an ancient account, such as explanation by material properties alone, rejection of teleological explanation, or the view that causation should be restricted to contact action. The characterizations of mechanistic thought typically offered can be quite diverse perhaps more so than is commonly recognized and it is not clear that any one definition or subset of necessary and sufficient characteristics would encompass the different ways the term is used. While there surely are similiarities between ancient atomism and later corpuscularian philosophies, care must be taken with the use of an omnibus term like 'mechanistic' to describe these similarities, lest it obscure as much as it reveals. Here I do not attempt an analysis of the manifold ways the terms 'mechanical' or 'mechanistic' are used.2 Rather, I intend to focus attention on only one possible meaning, one that identifies a method of investigation rather than some substantive feature of a natural philosophy. The term 'mechanistic' can be taken to refer to a method of investigating the natural world through terms and principles drawn from the discipline called 'mechanics'. So understood, it is only to be expected that the characteristics associated with the term should be diverse and open-ended and vary through time, as the understanding of mechanics has evolved. This is not the way the term is most commonly used; nonetheless, I suggest that this sense of