Elision and Prodelision in Speech and Verse

When one Latin word ends in a vowel and the next begins in a vowel, the two written vowels do not necessarily make two spoken syllables. In connected Classical Latin, especially as verse reveals it, the boundary was commonly smoothed by elision; before es and est, the initial e- could instead be lost by prodelision. English speakers need to resist two opposite temptations: reading every written vowel as a separate beat, or mechanically deleting letters as though the spelling were a phonetic transcript.

The secure fact is syllabic: verse often counts the boundary as one syllable rather than two. The exact sound—complete loss, vowel coalescence, very short transition, or nasalized residue in some environments—is not recoverable with equal certainty in every case. Ancient spelling underdetermines that detail, so a responsible reading separates evidence from classroom convention.

The basic environments

Ordinary elision occurs when a word ending in a vowel, a diphthong, or a vowel plus final -m comes before a word beginning with a vowel or h- plus a vowel. Initial h did not normally protect the first vowel. A consonant, a real pause, or consonantal i normally blocks the process.

In a continuous reading, multa et normally joins at the boundary: the final -a of multa does not receive its own full syllable. The printed letters stay unchanged. Elision is first a fact about pronunciation and meter, not an instruction to respell prose.

Here ille hominum behaves as though h were absent for elision. That does not by itself prove that every speaker always omitted /h/ in every context; it shows that initial h did not normally prevent vocalic contact.

There is no elision in mea iuventūs: the i of iuventūs is consonantal /j/. The visible sequence a i therefore does not automatically mean “two adjacent vowels.” Parse the letters into sounds before applying the rule.

The minimal boundary contrast is visible in two ordinary reports:

The final -a of fēmina meets a vowel in adest and normally elides; it meets consonantal v /w/ in venit and remains a syllable. English spelling habits obscure this contrast because English readers tend to assign v its modern /v/ value without first analyzing the Latin onset.

Final -m: elide the preceding vocalic unit

A word ending in vowel + m also elides before a following vowel. Metrically, multum ille does not preserve the full final syllable of multum. Ancient descriptions and comparative evidence indicate that final -m was weak and that the preceding vowel could be nasalized, but meter alone cannot tell us exactly how much nasality survived at every period or speaking rate.

In connected Classical reading, treat multum ille as a single joined boundary, approximately mult(um) ille, rather than pronouncing a forceful English-style final /m/ followed by a new onset. The parentheses are an explanation, not a spelling to use in Latin text.

The end of bonum is the eliding unit. It is misleading to say simply that “the m disappears”: the metrical effect concerns the entire final syllabic nucleus, and the phonetic history of final -m is more complicated than deletion of one consonant.

This is where English spelling habits are especially unhelpful. English trains readers to give a written final consonant a firm release; Classical Latin evidence points instead to a weak final -m whose interaction with the preceding vowel mattered. See word-final m for the broader evidence.

Prodelision with es and est

With forms of sum beginning in e-, especially es and est, Latin often suppresses the second word’s initial vowel. This is called prodelision or aphaeresis. Thus factum est may be realized as factumst, and mea est as meast. The contracted spellings occur in early and informal evidence, but a normal edited prose text usually prints the words separately.

At the boundaries factum est and mea est, a close connected reading may suppress the initial e- of est. Do not move the stress: factum est does not acquire an English-style compound stress merely because the words join phonetically.

In beāta es, prodelision is possible in sufficiently close speech, but deliberate emphasis on es can preserve a clearer boundary. Syntax and information structure matter: reduction is less likely across a pause or when the copula is strongly contrasted. Verse supplies stronger evidence for where a syllable fails to count than prose spelling supplies for the exact conversational rate.

Reading verse without inventing a spelling

In hexameter and other quantitative meters, elision helps determine whether the line has the required sequence of heavy and light syllables. It does not make vowel length and syllable weight identical. A naturally long vowel can elide; a short vowel can stand in a heavy syllable before consonants; and an elided syllable does not occupy an independent metrical position.

Vergil opens the Aeneid with the following printed sequence:

This quotation is reproduced without pedagogical macrons. In the verse, Laviniaque venit has elision at -que v-? No: venit begins with consonantal v /w/, so it cannot trigger elision. By contrast, Vergil's multum ille et terris iactatus et alto contains elision at multum ille and ille et. The contrast shows why looking for written vowel letters is insufficient: consonantal v and consonantal i must first be recognized as consonants.

Here the meter securely shows compression at two adjacent boundaries. A classroom may realize them by deleting the first vocalic unit, by smoothly blending the vowels, or by retaining a trace. Those are performance conventions built on the same metrical fact.

Poets sometimes allow hiatus, keeping adjacent vowels in separate syllables, particularly at a strong pause, with interjections, Greek words, or for expressive effect. Such cases are (literary) licenses, not proof that the ordinary pattern did not exist. Conversely, a singer may preserve vowels for musical reasons; that is evidence about a performance tradition, not automatically about ordinary Classical conversation.

Speech, verse, and register

Elision was not merely a poetic trick. Comic spellings, contracted forms, grammatical testimony, and Romance developments support connected-speech reduction. Verse is simply our most systematic evidence because meter makes syllable count observable. The strength of reduction probably varied with tempo, emphasis, and register, just as English “want to” ranges from careful [wɒnt tu] to reduced forms without changing its spelling.

In (formal) deliberate prose, use smooth but intelligible juncture. In (literary) verse recitation, make every metrically required elision audible enough that the rhythm works. In reconstructed (informal) dialogue, stronger reductions are defensible, but they should not be advertised as uniquely certain phonetic facts. Do not add apostrophes to an edited Latin text unless the edition itself uses them.

Common Mistakes

1. Giving every written vowel a separate beat.

2. Treating initial h as a barrier.

3. Eliding before consonantal i or v.

4. Treating macrons as stress marks.

Key Takeaways

Elision normally joins final vowel, diphthong, or vowel + -m to a following vowel or h- + vowel. Prodelision chiefly removes the e- of following es or est. Meter proves the loss of an independent syllable more securely than it proves one exact phonetic realization. Parse consonantal i and v, respect pauses and exceptional (literary) hiatus, preserve the printed spelling, and let vowel quantity—not English spelling values—guide your underlying forms.