Philodemus’ De signis is one of the classical texts of greatest semiotic interest. It reports the debate which arose between the Epicureans and an opposing school, usually identified as the Stoics, concerning semiotic inference. The Epicureans proposed to construct semiotic inferences based on generalizations resting on similarity, ultimately configuring their method as a form of induction. Their opponents attacked the Epicurean proposal in a twofold way: on the one hand, they argued that the Epicureans’ method intrinsically lacked cogency, invalidating their (...) inferences from a logical point of view. On the other, they criticized the notion of similarity, arguing that it is generally a vague notion, and in some cases impossible to implement, as when one is faced with unique cases. The debate is very complex and is divided into replies and rejoinders. The ultimate impression one gets is that the Epicureans were able, for the first time in antiquity, to propose a real method to construct semiotic inferences, even though the latter were subject to fallibility, while their opponents did not propose a method, but a test, “elimination,” able only to check the logical soundness of semiotic inferences. In doing so, they placed themselves in a tradition extending back to the theory of signs formulated – albeit in a significantly different way – by Aristotle. (shrink)