Do the Latin stress rules (antepenultimate if penultimate is light, penultimate if heavy) have any known exceptions? Also, sometimes the rule assigns antepenultimate stress to a syllable belonging to a prefix — does the rule apply regardless?
Here's a summary of most common exceptions (based on Belov 2007, Borovskii and Boldyrev 1975, Sihler 1995, Tronskii 2001):
- illic (from illice), istuc (from istuce), adhuc (from adhuce), addic (from addice), adduc (from adduce);
- NOM.SG.M. ending in -as or -is (originally, -atis and -itis respectively); e.g. nostras, Arpinas, Maecenas, Samnis etc.;
- audit (from audivit), fumat (from fumavit), irritat (from irritavit) etc.;
- interjections attat and papae;
- GEN. for nouns ending in -ius and -ium;
- enclitic -que, -ne, -ue. e.g. Musaque, uidesne, facisue. cf. itaque vs. itaque.