The ancient notion of an art (τέχνη) embraced a wide range of pursuits from handicrafts like shoemaking and weaving to more exalted disciplines not excluding philosophy (cf. Plato Gorgias 486b; Hippolytus Refutatio. 570,8 DDG; Sext. Emp. Μ II13). Nevertheless, there was a sufficient amount of agreement about what was expected of an art to permit debates about whether different practices qualified as arts. According to the conception which made these debates possible, an art is a body of knowledge concerning a distinct subject matter which enables the artist to achieve a definite type of beneficial result. Obviously, the failure of a practitioner to achieve the aim of an art can form the basis of a fairly simple challenge to his artistic competence. But when all of a profession's practitioners are prone to failure, even those who fully satisfy its internal standards, the artistic status of the practice can be called into question. Doubts then shift from the competence of an individual practitioner to the assumption that his practice is an art. The aim of this paper is to show in outline how the persistent occurrence of failures of this sort in some arts led to important developments in the ancient conception of an art. The developments we will be concerned with are not the specific and substantive changes in the theory and practice of particular arts which failure may have prompted, e g., the kind of improvement in surgical practice we would expect surgical failures to suggest. Rather, our concern will be with the way in which regular and apparently ineradicable failures led to the revision of the conception of an art itself. One such development, I will suggest, was the emergence of a crude but recognizable notion of probabilistic knowledge.
Allen, James, "Failure and Expertise in the Ancient Conception of an Art" (1989). The Society for Ancient Greek Philosophy Newsletter. 324.