In his Letter to Herodotus (75f.), Epicurus offers a strikingly non-teleological theory of the origin of (spoken) names, the first phase of which (75) is emphatically and explicitly naturalistic:

The details are much contested, but the general picture seems to be the following: in early humans, involuntary vocal responses were produced indirectly by the external environment, directly by internal psychophysical states themselves caused by external objects. These vocalisations were shaped both by the different objects which came to be named by these vocalisations, and by the physiological and psychophysical idiosyncrasies of the populations of different regions. The notion that people may be differently constituted not only physiologically, but also psychologically, according to their physical environment, was, of course, a fairly common one in antiquity, and not at all a ‘curious idea’, as Bailey describes it (1926: 248, ad Ep. Hdt. 75.8). In brief, that different things have different names within languages, and that the same things have different names in different languages, are alike explained by differences in both the constitutions and the environments of different peoples, who come to make up different language communities.