Epíkouros' On Nature, Book 28 vs. Plato's Cratylus

It is clear that Epíkouros, in Book 28, takes up Plato's discussion of "conventionalist" vs. "naturalist" views of language.

A linguistic conventionalist thinks that languages come about only by convention. The conventionalist denies a word's intrinsic relationship to a class of objects and therefore asserts that any word can refer to a particular class of objects—and that no particular word is better per se for this purpose.

In contrast, Epíkouros recognizes that linguistic conventions originate from an intrinsic relationship between a class of objects and the primary expression naturally used (specific to race and location) to identify that class. Languages originated from natural impulses. Because of this, most words inherently apply to an originally specific concept. Therefore, Epíkouros recommends that we base our vocabulary upon these fundamental meanings.

So even though Epíkouros and Cratylus both disagree with Hermogenes regarding convention, Epíkouros and Cratylus disagree on the basis of "correct" usage (with Epíkouros basing this in automatic association of fundamental etymologies, and Cratylus in the divine origins of language).

Quote from Cassius

there is a "Natural" word for a banana

Epíkouros says that he and Metrodorus, when they were younger, used to think there was no inherent connection between a word and an object -- particularly for an object of thought only conceived by analogy to the seen.

Yet over time they came to increasingly sympathize with the idea that there is an inherent connection (specific to race and location) for "the primary thought according to each word." (10.38a). For example, Epíkouros says "we do not use customary terms out of their bounds, nor while changing words for evident things." (Epíkouros, Peri Phýseōs, Book 28, P.Herc. 1479, fr. 13 [col. 5 sup.])