Skip to main content

Varro’s De Lingua Latina is now recognized as a milestone in the history of linguistics, and scholars have done much to show just how sophisticated Varro’s deployment of morphology and historical phonology can be (on the former, see Taylor 1974; for the latter, Pfaffel 1981). This paper shows that Varronian semantics can be launched into the same orbit of study. Although Varro downplays semantics in his discussions of etymology, scholars have indicated that there are significant philosophical and cultural reasons to suspect that Varro has a driving interest not just in how words are shaped, but in how they mean (Blank 2008; Hinds 2006). I suggest that a sophisticated semantic practice undergirds Varronian etymology.

Varro’s semantic interest is corroborated by three sets of passages from De Lingua Latina in which Varro faces multiple possible etymologies for a single word. In the first set, Varro prefers one alternative over the others. Sorting out why he thinks, say, caelum comes from cauum rather than caelatum or celatum (5.18-20) gives insight into what Varro thinks the most plausible etymologies should look like. In particular, Varro prefers to derive a word from another word (i) whose connection to the word in question is durable, specific, and complete; (ii) whose connection to the word in question is assured by usage; (iii) whose connection to the word in question is not metaphorical. The sky only looks engraved metaphorically. It only seems to be uncovered some of the time. But it always looks concave, and that people think of it as concave can be shown in both everyday usage and poetic quotation.

I suggest that these principles explain Varro’s practice in another set of passages. Sometimes Varro makes ‘unobvious’ etymological moves, as when he says the fratres in Fratres Arvales might come from fero, or Greek phratría, but ignores the seemingly obvious option that it just means ‘brothers’ (5.85). Varro finds more specific, durable and non-metaphorical reason to connect the Arval Brethren with ‘bearing’ and a voting association than with notional kinship.

There is, however, a final set in which Varro faces multiple alternatives and does not choose. Are these instances where Varro’s semantic practice fails? I suggest, instead, that if a single word cannot satisfy the principles of a good etymology Varro attempts to enmesh the word in question in a semantic web that cumulatively satisfies the principles. In these passages, Varro presents not options A or B, but options A and B harmonized. One instructive example is Varro’s derivation of ver (6.9), where he deploys both cumulative and disjunctive presentation of the options. The word for springtime might derive from virgulta and virere and vertere, or it might derive from Greek êr. Taken together, the three Latin options build out the meaning of ver, though none of them individually is captures completely what spring is. The Greek word, on the other hand, as a translation equivalent, can encompass all the ideas of spring, and therefore satisfies Varro’s semantic principles by itself.

Varro’s etymological practice, therefore, allows us to adumbrate some aspects of his historical semantics. Ancient etymology need not be a matter of just muddling through, as is sometimes suggested (e.g. Augustine Dialect. 6 p. 9-17 Crecilius), but the semantic drive of Varro’s etymology does resemble modern examples of ‘popular’ or ‘folk etymology’, and both his nearness to amateur thought-habits and his elaboration of them need to play a role in assessing Varro’s contribution to the history of linguistics.