Abstract

Ancient sign theory was concerned with when and how inferences from the observed to the unobserved were justified. This thesis uses the report of Sextus Empiricus, a skeptical philosopher of the second or third century A.D., to investigate Hellenistic sign theories. In the first chapter I explain how the different schools involved in the debate, chiefly the Skeptics, the Stoics and the medical Empiricists, understood the key terms in the dispute. ;Chapters two and three examine Sextus' treatment of signs in Pyrrhoniae hypotyposes II and Adversus Mathematicos VIII respectively. Both passages distinguish commemorative from indicative signs. The commemorative sign, which Sextus wants to accept, is a matter of projecting observed regularities in the sequence and conjunction of events. The indicative sign, against which he promises to direct his whole argument, depends on a natural power to reveal what it signifies. I argue that the Stoic theory Sextus attacks in the PH II passage is not a theory of indicative signification and that his arguments do not advance his stated purpose of arguing against indicative signs alone because they are equally effective against commemorative signs. Unlike the PH II passage, the M VIII passage does contain arguments directed against the indicative sign. These arguments are suitable for a proponent of the commemorative sign. I trace these arguments to the controversy between the medical Empiricists and Rationalists. The mistaken equation of Stoic and indicative signs which mars the PH I passage, I claim, resulted from a failure this controversy separate from the dispute between the Stoics and their opponents. ;In the fourth chapter I turn to the Stoic theory. I conclude, surprisingly, that the Stoic sign has more affinity with the commemorative than with the indicative sign. In the final chapter I approach the issue of how signs can give a partial warrant to the conclusion of an inference by relating it to ancient concerns about the compatibility of knowledge and failure. I discuss different responses to this problem but lay particular stress on what I argue is the medical Empiricists' development of a probabilistic account of inference