Lerna I: 5000 × 3 ≠ 8500

I mentioned in the other place the Lernaean text, which fulminates that Greek has 90 million words, and English a mere 490,000. The text, of course, comes from people who have seen the Thesaurus Linguae Graecae’s word counts, and can’t tell between a word count and a word list. Still, it’s a wonder DARPANET could even come up with the internet, with English so obviously impoverished.

Nikos Sarantakos has long tracked the origins and extent of the Lernaean, and its appearance in a speech by a member of the Academy of Athens led to much discussion among the Fortiers. (That’s what I choose to call Nikos’ readership, since his surname is a patronymic referring to the Forty Martyrs of Sebaste, Άγιοι Σαράντα.) In the end, a couple of letters to the editor went out, first from a team of Fortiers, then just by himself.

Letters to the editor don’t really help, as the Fortiers found. The first letter got a response from a retired sports administrator, fulminating about “a team of known or unknown persons, whose scholarly knowledge of Greek is impossible to confirm”, before launching into all the greatest hits of Lerna, complete with Ibycus, Packard, and McDonald misspelled. The second got a response from an administrator in Athens Uni, with the obligatory conspiracy theory about Team Fortier (which now does exist), and reasserting that “linguists at the peak of their field and hellenists of international renown assert with evidence that the Greek language from Homeric times until today—unitary, undivided and uninterrupted in all its manifestations, is the richest in the world. This is obvious from the dictionaries (e.g. Stephanus, Liddell-Scott, the Adrdados dictionary with triple the lemmata than Liddell-Scott, etc.)”

Theodore Andreakos, you don’t even know what Imycus looks like, and Seraphim Dedousis, you don’t know what dictionaries do. DGE will be double to triple the size of LSJ (if and when it finishes), but that’s because they’re spending triple the pages going through the detailed definitions and citations of words. They’re going through the TLG’s word list, so they’re including lemmata LSJ didn’t, but noone has ever said it was going to be triple the lemma count: in the introduction to volume 1, they say that it contains 8500 lemmata to LSJ’s 5000. And 8500 is not triple 5000 in my understanding of arithmetic.

DGE are substantially overlapping with Lampe and Trapp; that’s not a bad thing, the more the merrier, especially if DGE has a more global outlook on the vocabulary than either LSJ or Lampe, because it has a larger purview than either. Still, DGE is not necessarily adding words no lexicographer has ever seen before. I’ve manually added all DGE lemmata for antiquity (up to ii AD) that show up in the TLG corpus that LSJ, Lampe and Trapp didn’t include. DGE is up to ἐκπελεκάω, by which point LSJ is up to 34,000 headwords. The number of missing DGE lemmata I found in the texts and added? 198.

That’s naughty of me: the TLG does not do non-literary papyri, and that’s where most of the new ancient words come from. But it’s not taking us to 350,000 lemmata (which you’d need to triple LSJ). To be fair, I note that they have published a supplement volume on the magical papyri. Since the TLG does include those texts, I went through the supplement as well; that brings us up to… 279 added lemmata.

I’ll let you know if the remaining 89.8 million lemmata turn up somewhere…