Divergent Notions of Probability in Aristotelian and Epicurean Meteorology

In this paper I want to argue that, although Epicurus was clearly influenced by Aristotle’s meteorology, and especially by the role played by probability in this field of research, in the end he developed a meteorology, and a view of probability, that were radically different from Aristotle’s.
According to Aristotle, meteorology deals with phenomena that are such-and-such for the most part, which is equivalent to saying that they are likely. Likelihood in turn implies the capability of being otherwise. In practice, however, Aristotle only rarely expresses doubt about the certainty of his meteorological theories. A possible explanation for this apparent discrepancy can be found in the way he conceives of for-the-most-part phenomena. According to Aristotle a phenomenon of this kind can be analysed into a regular, natural component, and a random, unnatural component. Science only deals with the natural and hence with the regular, leaving the random component out of account. Accordingly, while specific phenomena are only such-and-such for the most part, general theories about them are necessarily fixed and certain. Underlying this is the assumption that humans are able to detect and define regular patterns even if they are blurred by random noise.
Epicurus, on the other hand, does not recognize this ability. Apart from the atomic substratum of reality there is no absolute regularity underlying meteorological phenomena. Accordingly it not possible to point out a single, uniquely true, explanation. The only thing we can know is whether a theory is possible, in which case it must be accepted. In this way Epicurus accounts for every meteorological phenomenon with a list of possible explanations. There is some disagreement about the status of these alternative explanations. According to some scholars they are all equally possible: we might call this interpretation sceptic. According to others they are all equally true, if not in this world, then in some other world somewhere in the infinite universe: we might call this interpretation pluralistic. A third interpretation we find in the work of the later Epicurean Diogenes of Oinoanda, who concludes his account of the method of multiple explanations with the statement that it is okay to say that one explanation is more probable than another. This version of the theory we might call probabilistic.