Where your Dictionary is Wrong
Latin dictionaries, at least good ones, do their best to mark which vowels are long by means of macrons: āēīōū. However this is an area where many dictionaries (even such as L&S and OLD) not seldom are misleading, if not outright wrong.
In what follows, I will list some of the primary areas where the vowel quantity was different from what most dictionaries indicate. Most information I have got from Vox Latina by Sidney Allen, which I whole-heartedly recommend to anyone who is at least a bit interested in classical Latin pronunciation.
Intervocalic, consonantal I
As most would know, an “i” which stands between two vowels in a word is usually consonantal, (a fact that is reflected in it sometimes being spelled as “j”). What is less known is the peculiarity that the sound is not only consonantal, but in fact double, i.e. a long consonant. Indeed, in poetry, the syllable which contains the vowel which preceeds the “j” always scan as heavy, but due to the double pronunciation of the consonant, that syllable is in fact long by position, not by nature, as would be expected from the spelling.
And indeed, that preceding vowel is regularly marked with a macron in dictionaries, either because the editors simply were ignorant of the double pronunciation of the consonantal “i”, or because the use of the macron was, though regrettably confusing, the only feasable way to indicate the length of the syllable due to typographic constraints, or, perhaps most likely, because they only wanted to indicate the heaviness in some way to help in the scanning of poetry, and not really caring about the actual pronunciation.
We then have that, what in dictionaries is listed as “māius”, was pronounced /măjjus/, and similarly /hŭjjus/, /ĕjjus/, /ăjjo/, /trŏjja/, etc. Exceptions are compounds such as “diiudico” = /dījūdicō/, and /trājectus/, etc. Also note the name “Gāius” = /gāius/, i.e. trisyllabic, with a vowel “i” (though in less careful speech supposedly /gājus/).
Hic, haec, hoc...
This common pronoun has a peculiarity similar to the one noted above, for, what in dictionaries (and grammars) are listed as “hīc”, “hōc”, really stands for /hĭcc/, /hŏcc/, with long consonant. (Historically, what happened was that the pronoun “hĭce”, neuter nominative “hŏcce” (from “hŏd”+“ce”), lost the trailing “e”, and the resulting /hŏcc/ in the neuter was spelled “hoc”. The consonantal length of “hoc” was then sometimes, but not always, borrowed by “hic”, which originally had a short “c”.)
Note that “hīc” and “hōc”, with long vowels, are the correct pronunciations of the adverb, ‘here’, and the pronoun in the ablative, respectively.
Artificial lengthening in poetry
Due to the nature of Latin poetry, some words are simply unusable in certain meters. For example, the word “arbŏrēs” can never be used in a hexameter due to the light middle syllable which is surrounded by two heavy, for in a hexameter, light syllables always come in pairs (except possibly on the end of the verse). Poets usually then have no choice but to use synonyms, or other circumlocutions.
Sometimes, though, they can be bold, and simply squeeze and stretch the word a bit so as to fit it into the meter. This regularly happens to “Ĭtălĭa”, which has (at least) three light syllables in a sequence, but which when used by e.g. Vergil always gets the first vowel prolonged so as to make the first syllable long by nature (albeit artificially so). Indeed “Ītălĭa” is what you will most likely find when you look up the word in your dictionary, but you should know that this is based on how the word is used in poetry, and does not reflect how the word was actually pronounced in daily speech. Evidence for this can be found e.g. in Quintilian (ca. 35 – ca. 100), Institutio Oratoria, 1.5.18:
... Praeterea quae fiunt spatio, sive cum syllaba correpta producitur, ut "Ītaliam fato profugus", seu longa corripitur, ut “unĭus ob noxam et furias”, extra carmen non deprendas, sed nec in carmine vitia dicenda sunt.
... In addition that which happen in regard to length, either when a short syllable is being prolonged, as in “Ītaliam fato profugus”, or when a long gets shortened, as in "unĭus ob noxam et furias", which you will not find outside of poetry, but which, when occuring in poetry, is not to be seen as an error.
While the vowel quantity in open syllables are readily apparent from poetry, the quantity of a vowel in a closed syllable is “hidden”: it can be long or short, but in either case the syllable will be long (heavy) in poetry. Clues as to the quantity of the vowel must then be searched for by other means, through etymology, Greek transcriptions, inscriptions, etc., and the conclusions are seldom as certain as the quantity of vowels in open syllables. Furthermore, the theories vary over time, so, while a certain class of words were thought to contain long vowels at one time, later research may show the contrary. This is of course also mirrored in dictionaries from different times. The OLD takes the safe route, and chooses to never mark hidden quantity, even in such cases where all the evidence agree and there is virtually no doubt.
For a (slightly dated) treaty on hidden quantity, see chapter 3 of The Latin Language by Charles E. Bennett.
Vowels before “gn”
Vowels before “gn” are regularly marked as long in some older dictionaries, but there is, to make the story short, no reason to assume that that was always the case. Unless there are reasons to assume that the vowel is long based on other derivations of the word stem, e.g. “rēgis” which points to “rēgnum”, it is more probable that the vowel is in fact short. (Allen indicates that list of words where the vowel is long before “gn” in fact may be restricted to “rēgnum”, “stāgnum”, “sēgnis”, and “abiēgnus”.)
Hence, for example, “ăgnus”, “măgnus”, “ĭgnis”, “dĭgnus”, “ĭgnotus”, etc., etc..
Errata for individual dictionaries
Here I collect errors in different dictionaries I have come across. Only such errors that are not covered above are included. The list is obviously in no way complete.
A Latin Dictionary, Lewis and Short
The following words are erroneously marked with long final e.
dēnĭquĕ, diffĭcĭlĕ, exindĕ, fĭdēlĕ, fortassĕ, fortĕ, futtĭlĕ, hūcĭnĕ, impūnĕ, măgĕ, magnŏpĕrĕ, mĕmŏrĕ, neutĭquĕ, perfăcĭlĕ, pĕrŭbīquĕ, plērumquĕ, prŏindĕ, quācumquĕ, quālĭtercumquĕ, quandōcumquĕ, quandōquĕ, quantŏpĕrĕ, quantŭlumcumquĕ (sub verbo quantŭluscumque), quāquĕ, quŏusquĕ, rĕpentĕ, segnĕ, sīcĭnĕ, sīcundĕ, sŭpernĕ, tĕmĕrĕ, tempŏrĕ (s.v. tempus), tētĕ, ŭbīquāquĕ, ŭbīquĕ
The adverbs hūc and illūc, and words related to those, like adhūc, are written without macrons.
Latinsk-svensk ordbok, Ahlberg, Lundqvist, Sörbom
expĕdĭtiō, expĕdĭtus, expĕdĭtē > expĕdītiō, expĕdītus, expĕdītē
2 prōfectus > 2 prŏfectus
Vox Latina – the pronunciation of classical Latin, W. Sidney Allen, 2nd ed. 1978.
The Latin Language – a historical outline of its sounds, inflections, and syntax, Charles E. Bennett, Allyn and Bacon, Boston, 1907.
Final -e in Lewis and Short, O. A. W. Dilke, Greece & Rome, 2nd Ser., Vol. 6, No. 2. (Oct., 1959), pp. 212-213.
Copyright 2007 by Johan Winge. Back to the index.