12

In this answer to a question about the preciseness of Latin, there is a quote from from Frederic Taber Cooper's Word formation in the Roman Sermo Plebeius (1895):

There were, as Cicero himself has pointed out, three ways in which the deficiencies of the vocabulary could be supplied; either by the transfer of a Greek word bodily into the Latin, by the use of an existing Latin word in a new sense, or by the formation of a new word. But in the classic period the use of foreign words was felt to be contrary to good taste and was accordingly avoided as far as possible, while unusual expressions, either archaisms or neologisms, were severely discountenanced. Even Cicero, who did more than anyone else toward giving currency to new formations, introduced many excellent and sorely needed words with hesitation and apology. Quintilian, while admitting that new words must occasionally be risked, says frankly that even when received into the language they brought little credit to their author, and if rejected led only to ridicule; and Gellius, still more emphatic, declares that new and unknown words are worse than vulgarisms.

This extreme attitude, however, had become untenable long before the time of Gellius; a point had been reached where growth of vocabulary was essential to the life of the language. But it was a natural consequence of such conservatism that no process existed for forming a literary vocabulary possessing distinctive features which might stamp it as a cultured product; no scientific nomenclature corresponding to the -ologies, -isms, and -anas of our own language; there was not a single suffix which could be regarded as distinctly classic, and which was not comparatively more abundant in authors of inferior Latinity.

If Latin was more resistant to new formations and vocabulary as compared to other languages, did this mean that Latin was more mutually-intelligible over time as well? For the purposes of this question I'd prefer to limit this to written Latin and avoid sound changes.

Adam
  • 8,520
  • 3
  • 20
  • 76
  • 5
    To me, this quote sounds like it's talking specifically about the tradition based on imitating a fixed Classical style, rather than the language as a whole. Certainly "authors of inferior Latinity" were still speaking Latin, even if Cicero would disapprove of their word choices. – Draconis Jul 22 '23 at 16:30
  • True. I wonder if Adams' Anthology of Informal Latin has anything about this. – Adam Jul 22 '23 at 16:32
  • 2
    As Draconis said, Cooper is talking there specifically about a conservative literary register of Latin, which he believes was distinct from popular speech. The whole book can be read here, if further context is helpful: https://archive.org/details/cu31924021614775/ – Asteroides Jul 22 '23 at 16:55
  • Also worth noting that lexical change is just one aspect of language change, and that in terms of sound changes and attendant levelling processes Latin was notably much less conservative than many of its neighbours. – Cairnarvon Jul 22 '23 at 18:19
  • It seems to me that Latin did change over time, it's just that we call the modern, widely spoken version of Latin, Italian rather than Latin. – Jack Aidley Jul 24 '23 at 10:41

2 Answers2

15

Blame Cicero!

Note the last line of that quote, about "authors of inferior Latinity". In other words, this isn't about the Latin language in general, but is about Latinity—the quality of imitating Cicero (among others, but mostly Cicero). And Albrecht Norden being quoted and criticized by Albrecht, mea culpa makes note of Cicero's "extremely poor vocabulary". Cicero wrote that a key component of good oratory was consistency. If you have one perfectly good word or construction for something, why do you need others? And it seems he often practiced what he preached.

The following table gives the numbers of types (distinct words) and tokens (total words) in the most prolific authors from the PHI corpus (i.e. all authors with more than 100k tokens).

enter image description here

(Taken from here, page 12.)

Let's ignore Justinian for now. (Legal precedents are very formulaic which makes them a bad example for our purposes.) Instead, let's look at Cicero, and Livy. Cicero has an enormously high token count, which makes sense, because so much of his writing survived compared to anyone else's. But his ratio of types to tokens is 0.07, while Livy's is 0.17, over twice that! In other words, if you grab a thousand random words from Cicero, and a thousand random words from Livy, you can expect to find over twice as many distinct words in your Livy sample as in your Cicero sample.

Similarly, Plautus has a ratio of 0.16, Ovid 0.17, and Seneca 0.14. Cicero really does seem to be the outlier here. And if you look at figure 5 on page 14 of that paper, you can see that removing Cicero from consideration markedly increases the overall information density of the entire corpus!

So I would argue it's not that Latin changed less over time—it's that Latinity did. Even today, Neo-Latin writers try to imitate the style of Cicero, which (by definition) hasn't really changed in two thousand years. If you look at the "authors of inferior Latinity" you'll find greater variety and plenty of change. And as far as mutual intelligibility goes, I would wager that everyone using this site can find some Romance language in the world that's harder for them to understand than Cicero!

Draconis
  • 66,625
  • 6
  • 117
  • 269
  • 14
    It should be noted, though, that this "types per tokens" measure should be be taken with some sort of caution. Be sure that if Cicero's corpus was double the size the number of different types won't double - it would go up but not doubled. In other words, your suggested test of taking 1000 random words actually sounds better and solves this issue. – d_e Jul 22 '23 at 17:15
  • 6
    @d_e True! I should try doing that experiment. – Draconis Jul 22 '23 at 17:21
  • 2
    I love to see what you find if you do that experiment. Great answer, by the way! – Adam Jul 22 '23 at 19:32
  • 2
    Wouldn't you expect the ratio of types:tokens to decrease as tokens increase? Is there a logarithmic graph of different authors for a visual comparison? – CJ Dennis Jul 24 '23 at 03:42
  • [1/2] I think d_e's comment isn't just a caveat, but a pretty fatal flaw with this answer. The type:token ratio of the sentence, "I ate my hat" is 1.00, but that tells me absolutely nothing about my vocabulary. I'm also skeptical of Stelzer's standing as a judge of Ciceronian style: he wrongly attributes the quote about "extremely poor vocabulary" to Albrecht, when in fact Albrecht (pg. 136) is quoting and criticizing Norden's characterization of "literary Latin." – brianpck Aug 10 '23 at 05:23
  • [2/2] Even Stelzer makes the decision to include Cicero as a representative author, rather than exclude him as an outlier (as he does for Justinian). The point about finding twice as many tokens in "a thousand random words" isn't right: this would only be true if every word was uniformly distributed, which is obviously not the case. – brianpck Aug 10 '23 at 05:24
  • @brianpck Oops, the misquoting Albrecht is on me; the paper cites Albrecht for the claim that "some" call it that, and I missed that Albrecht wasn't actually the one espousing that viewpoint. I'll fix that. But you're right, this really should be using normalized type-token ratios of some sort. When I have some time I'll go through and test that, and probably replace this answer with a new one. I have to confess the problem didn't occur to me until d_e pointed it out. – Draconis Aug 10 '23 at 15:36
-2

Latin changed into the proto forms of the Romance languages shortly after Rome fell in the West. People speaking proto-French and proto-Portuguese etc. all would have said that they were “speaking Latin” but when these groups came together, they could not understand one another. Then over more time, the Romance languages became what they are today. Thus making Latin a “living language”. Latin is still being developed via “Neo-Latin”.

Stephen Darrenkamp
  • 103
  • 3