[Mar. Vict. Gaisford, I. v. 6.] Sunt longae naturaliter syllabae, cum duae vocales junguntur, quas syllabas Graeci diphthongos vocant; ut ae, oe, au, eu, ei: nam illae diphthongi non sunt quae fiunt per vocales loco consonantium positas; ut ia, ie, ii, io, iu, va, ve, vi, vo, vu.
Of these diphthongs eu occurs,—except in Greek words,—only in heus, heu, eheu; in seu, ceu, neu. In neuter and neutiquam the e is probably elided.
Diphthongs ending in i, viz., ei, oi, ui, occur only in a few interjections and in cases of contraction.
While in pronouncing the diphthong the sound of both vowels was to some extent preserved, there are many indications that (in accordance with the custom of making a vowel before another vowel short) the first vowel of the diphthong was hastened over and the second received the stress. As in modern Greek we find all diphthongs that end in iota pronounced as simple i, so in Latin there are numerous instances, before and during the classic period, of the use of e for ae or oe, and it is to be noted that in the latest spelling e generally prevails.
Munro says:
“In Lucilius’s time the rustics said Cecilius pretor for Caecilius praetor; in two Samothracian inscriptions older than B.C. 100 (the sound of ai by that time verging to an open e), we find muste piei and muste: in similar inscriptions μύσται piei, and mystae: Paeligni is reproduced in Strabo by Πελιγνόι: Cicero, Virgil, Festus, and Servius all alike give caestos for κεστός: by the first century, perhaps sooner, e was very frequently put for ae in words like taeter: we often find teter, erumna, mestus, presto and the like: soon inscriptions and MSS. began pertinaciously to offer ae for ĕ: praetum, praeces, quaerella, aegestas and the like, the ae representing a short and very open e: sometimes it stands for a long e, as often in plaenus, the liquid before and after making perhaps the e more open (σκηνή is always scaena): and it is from this form plaenus that in Italian, contrary to the usual law of long Latin e, we have pièno with open e. With such pedigree then, and with the genuine Latin ae always represented in Italian by open e, can we hesitate to pronounce the ae with this open e sound?”
The argument sometimes used, for pronouncing ae like ai, that in the poets we occasionally find ai in the genitive singular of the first declension, appears to have little weight in view of the following explanation:
[Mar. Vict, de Orthog. et de Metr. Rat., I. iii. 38.] Ae Syllabam quidam more Graecorum per ai scribunt, nec illud quidem custodient, quia omnes fere, qui de orthographia aliquid scriptum reliquerunt, praecipiunt, nomina femina casu nominativo a finita, numero plurali in ae exire, ut Aeliae: eadem per a et i scripta numerum singularem ostendere, ut hujus Aeliai: inducti a poetis, qui pictai vestis scripserunt: et quia Graeci per i potissimum hanc syllabam scribunt propter exilitatem litterae, η autem propter naturalem productionem jungere vocali alteri non possunt: iota vero, quae est brevis eademque longa, aptior ad hanc structuram visa est: quam potestatem apud nos habet et i, quae est longa et brevis. Vos igitur sine controversia ambiguitatis, et pluralem nominativum, et singularem genitivum per ae scribite: nam qui non potest dignoscere supra scriptarum vocum numeros et casum, valde est hebes.
Of oe Munro says:
“When hateful barbarisms like coelum, coena, moestus are eliminated, oe occurs very rarely in Latin: coepi, poena, moenia, coetus, proelia, besides archaisms coera, moerus, etc., where oe, coming from oi, passed into u. If we must have a simple sound, I should take the open e sound which I have given to ae: but I should prefer one like the German ö. Their rarity, however, makes the sound of oe, eu, ui of less importance.”