6

I want to know the percentages of loanwords in Classical Latin (maybe including Old Latin but NOT post-classical Latin), including native terms and words, for example: 90% Native 7% Greek 1% Etruscan ?% etc...

d_e
  • 11,021
  • 2
  • 21
  • 40
ClassicalLatin4thewin
  • 61
  • 1
  • It doesn't include a percentage, but @Cerberus answered a similar question on the Linguistics SE about how many Latin words are of Greek origin. – Adam May 09 '21 at 20:40
  • @Adam Thank you, but I also want to know about non-greek loanwords. – ClassicalLatin4thewin May 09 '21 at 20:46
  • 1
    Wouldn't there be cases where we couldn't be able to tell whether a particular word was a Greek loan-word, a word that appeared in both Latin and Greek as a loan from some third language, or a PIE word? For instance, Troy is Τροία in Greek and Troia in Latin. How would we know whether this came to Latin through Greek or directly from Anatolia? I'm also not sure it makes sense to ask for a percentage. The percentage would be higher in fancy educated usage, lower in the vernacular. –  May 09 '21 at 21:18
  • Ben Crowell yes there are PIE words, but I still need those percentages. – ClassicalLatin4thewin May 09 '21 at 21:23
  • 2
    One way to try to answer this might be to go through a sample of Latin words and see how many have entries in de Vaan's etymological dictionary, since that dictionary only lists native Latin words. (@BenCrowell, cases of the kind you describe will be rare because most native Latin words will have undergone Latin-specific sound changes.) – TKR May 09 '21 at 22:18
  • Ok, I will wait for a professional to answer my question (thanks for your help :D). – ClassicalLatin4thewin May 09 '21 at 23:02
  • @TKR Contrary to his stated policy, De Vaan actually includes a good number of words that aren't (likely to be) native; he mainly excludes very transparent loans from Greek. I once went through a list of the 2000 most common words (based mainly on the Perseus corpus), and only about forty weren't in De Vaan, and of those, only persona and titulus weren't Greek. Some other high-profile non-Greek loans like ave and mappa (both from Punic) are also out, but e.g. harena is in, as well as any number of etymology-unknown certain loans. – Cairnarvon May 10 '21 at 01:20

1 Answers1

8

The following is from Loan-words in Latin (1888) by Edward R. Wharton.

He counted a total of 16,900 words from the following authors: Plautus, Terence, Cicero Caesar, Catullus, Lucretius, Sallust, Vergil, Horace, Livy, Tibullus, Propertius, Ovid, Persius, Tacitus, Juvenal.

Out of that total, he classified 92.43% of the words as natively Latin, and the breakdown of loan-words is as follows:

Words Group / Family Language Percentage
1080 Greek Greek 6.39%
21 Aryan Umbro-Sabellian 0.12%
43 Aryan Celtic 0.25%
5 Aryan Teutonic 0.03%
13 non-Aryan Etruscan 0.08%
1 non-Aryan Basque < 0.01%
13 non-Aryan Phoenician 0.08%
4 non-Aryan African 0.02%
1 non-Aryan Indian < 0.01%
90 Unknown Unknown 0.53%
Expedito Bipes
  • 10,740
  • 2
  • 29
  • 50
  • What definition is Wharton using for "Aryan" here? – Draconis May 10 '21 at 04:41
  • 1
    @Draconis Probably same from which the old name for IE, Indo-Aryan, comes, with Greek as a special category. Given that it's 1888, he should have made Indo-Aryan the category. I can't see why he would break it down this way. – cmw May 10 '21 at 04:45
  • 1
    @cmw I'm aware of early philologists using Aryan for IE, but I've never heard of them using Indo-Aryan that way (rather than for the family of languages still refered to as Indo-Aryan) – sgf May 10 '21 at 08:37
  • 1
    Terminology aside, it looks like Wharton is way too eager to call native Latin words Greek. This works as an upper bound (and it's probably the best answer possible without some serious legwork of our own, unless there's some other study out there), but it's a significant overestimate. – Cairnarvon May 10 '21 at 14:37