So I've come across this word βῡκάνη, ostensibly borrowed from Latin būcina ('an ox-horn trumpet'), from bou- ('ox') + canere ('to sing'). The lack of vowel reduction is immediately striking; additionally, Lat. V /ū/ corresponds to Gr. Υ /ʉ̄/. Three possibilities come to mind:

  1. it was borrowed from Oscan or another more southern Italic language where the reduction was absent (and /ū/ might have been at least allophonically centralised like in Greek). Note that bōs itself is often suspected to be of Osco-Umbrian origin, but I think it's predictably irregular to avoid an obvious homonymy with vōs ('you');
  2. it was borrowed before vowel reduction was operative in Latin - usually said to have been complete by the 3d century at the latest;
  3. it was actively undone during borrowing.

I'm wondering if there are further indications to help decide between these possibilities; and specifically I'm curious whether further examples can be found where A. Greek (or any other language, for that matter) shows lack of vowel reduction in loanwords from Latin.

  • 1
    Btw I've never seen the Greek vowel described as [ʉ̄] -- what's the evidence that it was centralized?
    – TKR
    Commented May 16, 2021 at 17:52
  • @TKR The fact that it underwent fronting from [u] and ended up as [i]; I simply don't know of evidence as to its exact quality in the period and varieties involved, so I picked the more conservative choice. In fact I've never seen evidence to decide whether it was [ʉ] or [y] in any sort of Greek, so I'm going by the general vowel space distribution wherein it further fronts to [y] when /oo/ raises to [u:]. Commented May 16, 2021 at 19:11

1 Answer 1


I looked up the Greek word in the etymological dictionaries of Chantraine and Beekes. They both say that your hypothesis #1 (an Oscan loan) was indeed proposed by Cuny in 1908, but that this was rejected by Niedermann in a 1917 Indogermanische Forschungen article, which can be viewed here. Niedermann says, if I understand correctly, that the specific meaning of βῡκάνη developed in Latin (though I don't see how we can possibly know this), and that the form of the word is analogous to pairs like machina-μηχάνη.

(Note that Oscan does actually show syncope, rather than reduction, of many medial short vowels; the conditions for this seem to have been complex or irregular and I don't know whether it should be expected to have occurred in a form like *būcana.)

Your hypothesis #2 seems perfectly possible given the history of Latin-Greek contact and I don't know why Niedermann and the dictionaries don't mention it. (Hypothesis #3 is obviously very implausible.)

On the vowel of the first syllable: in Attic and East Ionic (and later in the Koine) Υ fronted to [y], but in most other dialects, including those of Greek colonies in Italy, it remained a back vowel [u] as far as I know. If the word is an early borrowing then this may explain the Υ spelling. But note that there also exist derived variants with ΟΥ: βουκινίζω, βουκινάτωρ.

FWIW, the etymology from bou+canō seems a bit doubtful to me. There are other compounds in -cina, but those all refer to a female musician, where the first part of the compound is the name of the instrument (tībīcina, fidicina, sambūcina); by that analogy būcina would mean "female cow-player".

  • 2
    cf. Biville 1995 Les emprunts du latin au grec (t. 2, p. 106): "L’apparition tardive de βυκάνη (Polybe) et de ses dérivés, et les hésitations graphique entre βυ- et βου- dans les manuscrits, montrent que le grec a emprunté le terme au latin, en lui faisant subir une adapation suffixale sur le modèle des couples machina = μαχανά, patina = πατάνη, runcina = ῥύκανη, trutina = τρυτάνη."
    – Alex B.
    Commented May 16, 2021 at 12:58
  • Thanks for the answer; it's not the same analogy because of course an ox is not an instrument. It's a different type of compound with the same element (I'm totally lost with types of compounds). There's really nothing else that second element could be, nor the first one. I'm not quite satisfied with this answer though because Alex B.'s quotation provides just the type of "active reduction reversal" that I was looking for, ostensibly as a reversal of the way Greek loans are reflected in Latin. Commented May 16, 2021 at 19:21
  • possibly a typo in ῥύκανη - must be ῥυκάνη (?)
    – Alex B.
    Commented May 16, 2021 at 20:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.