I wish that Latin dictionaries came with a warning label:
Use with Caution! Vowel length is only a guess.
“You mean the dictionary could be wrong?” you ask.
“And the high school textbooks, too?”
“How is that possible?” You ask.
The long and short of it is that the dictionaries and the textbooks we use were put together long after Latin died. No one knows anything about Latin pronunciation for sure, but before we get into the thick of it, let us start by saying this:
Latin has long and short syllables. The rhythm of Latin poetry is based on this simple fact.
A long syllable contains a long vowel: ōs [mouth].
A short syllable contains a short vowel: os [bone].
But a short syllable becomes long, if the short vowel is followed by two or more consonants: osseus [bony]. (Modern spelling doesn’t require us to write ōsseus.)
This rule is so powerful that it applies even if the second consonant is in the following word: os fractum [broken bone]. Compare that to fractum os, in which os remains short.
This rule is too powerful, however. It needs some qualification. The rule, it turns out, is optional when the first consonant is
p b t d qu or g
and the second consonant is
l r n or m
Thus there must have been two ways to pronounce a word like agricola. Some Romans may have said āgricola while others may have said agricola. No one can know for sure now.
These three rules—long vowels make long syllables, short vowels make short syllables, and double consonants often make long syllables—have another application. They determine whether the word is accented on the next to the last syllable or the one before that. Long Latin words are accented like this:
agri´ cola [farmer], agrā´ rius [relating to land]
But if the second to the last syllable contains a long vowel or a double consonant, the accent is:
Agrippī´na [a woman’s name] not Agri´ppīna
agrest´is [rustic] not a´ grestis
Of course the qualifying rule about double consonants applies here as well. Thus we have in Cassell’s dictionary:
mātrō´na [a married woman] as well as Mā´trona [the Marne].
What an easy system! The only thing you have to know is whether a vowel not followed by two consonants is long or not or whether the qualifying rule applies or not. This is where the dictionary comes in. It will tell you whether the word is ōs or os or how to pronounce matrona. Unfortunately the dictionary-makers and their accomplices, the textbook-makers, really messed things up.
Charles Jenny, Jr., who revised Smith and Thompson’s First Year Latin, decided that vox [voice], which is really voks, needed to be marked long: vōx. He could have left it unmarked, since the vowel is followed by two consonants. Here are more examples of his illogical and unnecessary marking: dēfēnsiō [defense], āctus [driven], Etrūscī [Etruscans], and fōrmo [I form].
To make matters worse, dictionaries and textbooks don’t always agree. The Lewis and Short dictionary has Matrona, not Mātrona. In Cassell’s Latin dictionary, we find dixi [I said]. In Lewis and Short, we have dīxi, and in Kennedy’s The Shorter Latin Primer, we see dixī.
Why all the discrepancies? It’s because Latin has been dead a long time. The only thing we have to go on is what they left behind. Some ancient grammarians wrote about vowel length. Sometimes the Romans used a tick mark called an apex to mark long vowels, and in the case of the long i, they simply wrote it taller:
Roman poets, too, left behind a wealth of information about vowel length, for the words they chose had to fit the meter. A word like amare [to love] was used when the meter called for ˘ ¯ ˘. Thus, we mark the word amāre. (Amāre was never used when the meter called for ¯ ˘ ¯ ). The poets also helped us figure out what happens to the vowel when it is followed by two consonants, as described in the rules above. Thus, amāre flōrem [to love the flower] can only fit the meter ˘ ¯ ¯ ¯ ˘
Sometime in the eighteenth century, someone decided to use the macrons [¯] and breves [˘] employed in scanning a line of poetry to mark vowel quantity. This coupling of marks used in scansion and marks used to indicate vowel length was brilliant. Too bad that some modern Latinists were a bit overzealous and destroyed the simplicity that was before them by unnecessarily marking a word like vox as vōx or by marking the first a in matrona as mātrōna. In this last case, was the lexicographer influenced by etymology, since mater is marked māter? This is the kind of thing that complicates spelling systems. Some pedant tries to show off. Just think of the spelling of ‘doubt.’ The b was inserted to show that the word had come from dubitum, as if we cared.
So, what of Horace? All I can say is that it has been incredibly difficult marking his odes up with macrons and breves for the new edition of my Carpe Diem. I never imagined what a can of worms I had opened. I never dreamed in my life that textbooks and dictionaries marked some words according to their own sets of rules, rules unknown to me.
As I mentioned above, I found three different ways to mark dixi. In the odes, we find Horace making no commitment as to the quantity of the first syllable but as for the second, he intends for it to be long. (Here the tilde (~) means either short or long according to the meter):
dĩxī sǎcrāmēntum: ībǐmǔs, ībǐmũs, [II:17:10]
In the next two examples, it looks as if the first syllable was either long or short. Only Lewis and Short give us mājor. As for Cassell’s, Jenner, and Kennedy, all of them write maior. Clearly they must be wrong, for in III:20:8, Horace intends for the first syllable to be long.
māiǒr ǎn īllī [III:20:8]
mãiōr Něrōnūm mōx grǎvě prōēlǐũm [IV:14:14]
Dictionaries and grammar books sometimes run contrary to Horace. Although it is possible that Horace may enjoy a bit of poetic licence from time to time, as far as I can tell, he does so very rarely. In fact, when he does seem to veer away from following strict metrical rules, scholars come to his rescue and cite arcane rules of vowel quantity as in these lines; for rīdēt and perrūpīt, they say, are ancient forms of the verb:
āngǔlūs rīdēt, ǔbǐ nōn Hymēttō [II:6:14]
pērrūpīt Ǎchěrōnta Hērcǔlěūs lǎbõr [I:3:36]
In the following examples, however, no scholar has come to the rescue to explain why Horace changes the quantity of the vowel. In all of the dictionaries I have seen, testudo ‘tortoise shell, lyre’ is marked testūdo, but in these two examples, Horace turns o into ō.
tūquě tēstūdō rěsǒnārě sēptẽm [III:11:3]
grātǎ tēstūdō Iǒvǐs, ō lǎbōrũm [I:32:14]
Jenner, however, to his credit, sides with Horace. But it is John Grant in his Institutes of Latin Grammar (London: Whittaker, 1823, pg. 333), who provides us with a bit of an answer. He says that a final o is common, i.e., either long or short and sensibly marks these “common” cases with a tilde: õ.
Finally, time and time again, marking the long vowels meant making sure that I understood what Horace was saying. Did he intend to use the ablative or not? Did he really mean to make the noun plural? In the following two examples I discovered that, according to Lewis and Short and Cassell’s there is a big difference between lavit and lāvit. Lavit is ‘he washes,’ and lāvit is ‘he washed.’ The meter makes it clear which tense Horace had in mind, or does it? Perhaps the dictionaries are wrong. Maybe it didn’t matter to Horace what the quantity of the a in lavit was.
vīllāquě, flāvūs quām Tǐběrīs lǎvǐt [II:3:18]
sǐmǔl ūnctōs Tǐběrīnīs ǔměrōs lāvǐt ǐn ūndīs [III:12:7]
Niall Rudd in the Loeb Series translates the first as “and your villa washed by the yellow Tiber.” The second, he translates as “as soon as he bathes his oiled shoulders in the waters of the Tiber.” Should Rudd have written “has bathed in the waters of the Tiber”? Maybe. As for my translation done in December 2009, I wrote: “washes his oiled arms in the waves of the Tiber.” Somehow the present tenses fits the meaning I think Horace had in mind. Let me look in Jenner’s highschool textbook. Ah ha! He gives the past tense as lavit. Somebody’s wrong, and I don’t think for a moment that it’s Horace.
As I said at the onset: I wish Latin dictionaries and textbooks came with a warning label.
:: Latin books by James Rumford ::
For all 102 odes purchase Carpe Diem, Horace De-Poetized, for $11.50 at
For a Latin translation of The Adventures of Tom Sawyer at $12, click here:
To find out more about Carpe Diem go to the blog of March 26, 2015;
for more about Pericla Thomae Sawyer, go to the blog of November 22, 2016.