Iynx dixit:
Though I do not disagree with the substance of Mattheus' remarks here, I must take issue with his "completely" and "always". Apart from the exception he cites there are, as it seems to me, numerous other irregularities, even if one confines oneself to areas of general agreement. We might in illustration cite such words as Vergili (vocative of Vergilius), benefacit, calefacit, and venumdare, all of which are pronounced in violation of the usual rules.
If we allow ourselves to enter areas of controversy, there are still more exceptions (at least according to many). Consider: egomet, astronomia, theologia, and parietes.
None of this need trouble a beginner, it is true.
Sorry, I should not have so over-simplified. Other than addressing the wrong person, you are quite correct, of course.
2nd declension masculines in -ĭus have a vocative in -ī, but since it's actually a contraction of -ĭĕ, the accent manifests itself as if the word were not contracted. Hence
vergílī rather than
vérgĭlī. This is also true for the older contracted genitive -ī found in place of -ĭī.
Vēnumdăre "give for sale, sell",
bĕnĕfăcit "do good to, benefit",
călĕfăcit "heat, make warm" and other so-called "syntactic compounds" are accented as if they were separate words, hence:
vḗnumdắre,
bénĕfắcit,
cắlĕfắcit (but the more common form of the word for "sell" is a full compound:
vĕndo, vĕndĕre, like it's passive form
vēneo,
vēnīre "go for sale, be sold")
Ĕgomet is
ĕgo + the reflexive enclitic particle
-met. The enclitics in Latin, including
-quĕ,
-nĕ,
-vĕ,
-cĕ,
-met, and
-dum, are said to always shift the accent to the syllable immediately before the enclitic. Accordingly, it would be
egómet. Some authorities dispute this, however, saying that this is only the case when that syllable is long, either by position (
hŏmĭnísque) or by length of vowel (
mĭsĕrṓque), or if the accent was originally on the antepenult, in which case it is has a double accent (mísĕrắque, flúmĭnắve). So I think
égomet is still the more likely pronunciation, unless the 'o' is pronounced long, in which case it would be
ĕgṓmet. It's usually short in poetry, but there's some uncertainty about prose (it may have been long when the word is especially emphasized).
I suppose that words like
ăstrŏnŏmia and
thěŏlŏgia may be pronounced as their Greek equivalents (ἀστρονομία and θεολογία, respectively), but I believe the typical Latin accent was more common anyway.
I'm not sure what's unique about the accent of
păriĕtēs. I would assume it goes on the antepenult, according to the normal rules. Is that wrong?