Chaucer’s Twelve “Long” and “Short” Vowels: The Evidence from the Rhymes in Troilus and Criseyde

For more than 125 years all accounts of Chaucer’s vowels in stressed syllables have assigned him the twelve monophthongs that are called “short a ” and “long a ”; “short e,” “long open e,” and “long close e ”; “short i ” and “long i ”; “short o,” “long open o,” and “long close o ”; and “short u ” and “long u.” ¹ That the monophthongs are called “long” or “short” means that they are thought to have quantity—the length of time that the vowel sound is held—as a phonemic feature. Table 1 gives some of the symbols that scholars have used to represent the twelve vowel sounds: symbols used by Bernhard ten Brink in 1884, Walter W. Skeat in 1894, F. N. Robinson in 1933, Norman Davis in 1987, and V. A. Kolve and Glending Olson in 2005. ²

But the rhymes in Troilus and Criseyde, Chaucer’s longest poem, show that all the accounts of Chaucer’s vowels are dead wrong: Chaucer did not have long and short vowels. What distinguishes Chaucer’s vowels from each other is quality, not quantity; any differences in vowel length are allophonic, not phonemic. In this regard Chaucer’s English is the same as modern English. As Kolve and Olson have said, “even though it takes longer to say the vowel of ‘bad’ than that of ‘bat’,” modern English does not “distinguish[ ] between long and short vowels” (page xv).

Table 1.

Symbols That Scholars Have Used to Represent Chaucer’s Twelve “Long” and “Short” Monophthongs (see notes 2 and 21)

Lines of poetry rhyme when the last stressed vowel sound is the same in each line and any sounds that come after the last stressed vowel sound are also the same. As Roman Jakobson put it, “rhyme by definition is based on a regular recurrence of equivalent phonemes or phonemic groups.” ³ So when in Troilus and Criseyde words that manuscripts often spell bitwene, contene, grene, kene, polixene (a name), quene, (y) sene, shene, sustene, tene, and wene constantly rhyme with each other and thereby can be said to sort themselves into one pile, while words that those same manuscripts often spell bene, clene, lene, and mene constantly rhyme with each other and thereby can be said to sort themselves into a second pile—and no word from the first pile rhymes with a word from the second pile—then the simplest explanation is that all the words in both piles have an n sound followed by an unstressed final e sound, but that the words in the first pile have one kind of stressed e sound and the words in the second pile have a different kind. ⁴

The piles of rhyming words in Troilus and Criseyde show how many different contrasting sounds—that is, how many phonemes—Chaucer had. The piles show that in stressed syllables, instead of the twelve monophthongs that all accounts assign him, Chaucer had just ten:

• • Instead of both “short a ” and “long a,” Chaucer had just one a sound, neither long nor short.

• • Instead of three e sounds—“short e,” “long open e,” and “long close e ” — Chaucer had two e sounds plus, before r, an æ sound. Chaucer’s two e sounds, “open e ” and “close e ” (ę and ẹ), differ over quality, not quantity. So instead of having the stressed vowels ę̄, ẹ̄, and ę, Chaucer had ę, ẹ, and (before r) æ: vowels that are neither long nor short.

• • Instead of three o sounds—“short o,” “long open o,” and “long close o ” — Chaucer had just two, “open o ” and “close o ” (ǫ and ọ), differing over quality, not quantity.

• • Instead of the sounds “short i ” and “long i,” differing over quantity, Chaucer had what can be called “open i ” and “close i ” (I and i), differing over quality.

• • Instead of the sounds “short u ” and “long u,” differing over quantity, Chaucer had what can be called “open u ” and “close u ” (v and u), differing over quality.

Skeat’s Rime-Index

Published in 1891 or 1892, Skeat’s Rime-Index...