Some notes on how to read Latin verse

Note: This is not intended to give a complete survey of all aspects (whatever that means) of Latin poetry. Rather, I only intend to point out a couple of things which tend to be overlooked or misunderstood.

The Theory

A vowel in Latin may be either short or long. The terms long and short simply refer to the duration: a long vowel would be held a noticeable longer time than a short vowel. Vowel length is an innate quality of a word: changing the length of a vowel in a word will produce another word, just like what would happen if you change the vowel’s quality.

As far as the analysis of poetry is concerned, each verse is broken into syllables in a fairly straightforward manner, and without any regard for word divisions: each vowel sound (a single vowel or a diphthong) belongs to a separate syllable; if the vowels of two syllables are separated by one consonant sound, it is assigned to the following syllable (pa-ter); if there are two or more consonant sounds in between the vowels, at least the first goes to the first syllable, and the rest to the following (ar-ma, ter-ra) – unless the two consonants consist of a stop (/p, t, c, b, d, g/) followed by a liquid (/r, l/), in which case they usually count as one consonant and are moved to the following syllable (pa-tris). (Note that it is the number of sounds which is relevant, not the actual spelling; for example, “x” is two sounds (saxa = sac-sa), and so is in fact intervocalic consonantal “i” (Troia = Trŏj-ja), as well as intervocalic “z”, whereas “qu” is one sound (quo-que).)

A syllable that ends with a vowel is called open, while a syllable that ends with a consonant is closed. Based on this property of the syllable and on the length of the vowel, syllables can be sorted into two classes: light (or “short”), that is, those that end with a short vowel, and heavy (or “long”), which signify either those that have a long vowel (or a diphthong), or those that have a short vowel followed by a consonant. In the former case, the heavy syllable is traditionally called “long by nature” – that is, by the nature of the long vowel; in the later case, the syllable is named “long by position”, due to the position of the short vowel before a consonant within the same syllable; (or, according to a different interpretation of the original Greek term, “by convention”.) A structured presentation might make the matter more clear:

As an example, here follows the first line of the Aeneid, first in normal orthography, then broken up into syllables and with heavy syllables marked with bold text:

1. Arma virumque canō, Troiae quī prīmus ab ōrīs
2. ăr-mă-vĭ-rŭm-quĕ-că-nō-troi-jae-quī-prī-mŭ-să-bō-rīs

Both kinds of heavy syllables, those which are long by position and those which are long by nature, have that in common that they take longer time to pronounce than light syllables. This phonetic difference between light and heavy syllables is the very foundation of Latin poetry: the different meters are defined by the regular alternation between light and heavy syllables, and the resulting rhythmical variation which is achieved when reading the verse aloud is (part of) what makes it poetry.

The notion of “Long by Position”

There exists an all too common misconception, according to which a normally short vowel is “made long by position” when followed by two consonants. For example, the first two words of the Aeneid, “arma virumque” – which in natural speech have all vowels short –, would in poetry have to be pronounced as “aaaarma viruuuumque”, because the first “a” of “arma” and the first “u” of “virumque” both stand in closed syllables (“ar”, “um”), and are thus made long by position. Sometimes this is said to be the case in prose as well.

However, it should be stressed that there is simply no truth to this assumption, which seems to ultimately stem from a medieval confusion between vowel length and syllable quantity, and which continued to be taught all the way to the 19th century and beyond. For example, in The New American Latin Grammar from 1807, which is representative of the school grammars of the time, we are taught this charming rule:

A Vow’l before two Consonants is long,
In Latin, Prose, or Verse, if I'm not wrong.

Well, I’m sorry to say, but as it happens, you were wrong indeed.

When the ancient grammarians, who had first hand knowledge about the sound of the language, write about things that are “positione longa”, they invariably refer to the “syllaba”, not “vocalis”. Furthermore, there is incidently a passage by Aulus Gellius (ca. 125 AD – after 180) which deals with a related question, and which indirectly gives clear evidence against any such mode of pronunciation. Noctes Atticae 4.17:

Lucilii ex XI versus sunt:

Scipiadae magno imprŏbus obiciebat Asellus
lustrum illo censore malum infelixque fuisse.

“Obiciebat” “o” littera producta multos legere audio, idque eo facere dicunt, ut ratio numeri salva sit.

[Several similar examples follow, involving the words “conicere”, “subicit”, “conice” and “subicit”.]

Sed neque “ob” neque “sub” praepositio producendi habet naturam, neque item “con” [...] In his autem, quae supra posui, et metrum esse integrum potest et praepositiones istae possunt non barbare protendi; secunda enim littera in his verbis per duo “i”, non per unum scribenda est. Nam verbum ipsum, cui supradictae particulae praepositae sunt, non est “icio”, sed “iacio” [...] Id ubi compositum est, “a” littera in “i” mutatur, sicuti fit in verbis “insilio” et “incipio”, atque ita vim consonantis capit, et idcirco ea syllaba productius latiusque paulo pronuntiata priorem syllabam brevem esse non patitur, sed reddit eam positu longam, proptereaque et numerus in versu et ratio in pronuntiatu manet.

(A translation into English of the whole chapter can be found here.) In short, the placement of the word “obiciebat” in the hexameter reveals that the first syllable is heavy, even though the “o” of “ob” is naturally short. Gellius reacts to how some, who obviously think that the word is derived from the verb “icio” and hence that the first syllable is open, solve the problem by artificially lengthening the “o” (and thus, we can deduce, pronouncing the word as /o:bicie:bat/.) Instead, he remarks, “obiciebat” properly stands for “objiciebat” (from “ob” + “jacio”), which means that the first syllable is long by position, and hence there is no need to, as he puts it, barbarously prolong the vowel.

Despite such clear evidence, the idea of vowels becoming long by position continues to be perpetuated even today, online, in several otherwise eminent guides to Latin pronunciation. Amongst recent printed books which confuse syllables and vowels, we can note SAT Subject Test: Latin By Ronald B. Palma, 2006, (p. 508):

... Also long are diphthongs and vowels followed by two or more consonants in the same or different words; for example, the -o- in possumus is long by position because it is followed by two consonants, -ss, as is the second -u- in possumus legere, which is followed by s and l. ...

Another example is The Aeneid in translation by Patrick Dickinson, Signet Classics 2002, p. 382:

... A syllabic language has clear rules for the measurement of its syllables. They are long or short according to these rules, and this longness or shortness of a syllable scholars have called “quantity”. Of course, there are exceptions: the general rule is for a vowel [sic!] to be long if it is followed by two consonants, but in the line Sūnt lăcrĭ|māe rērum| ēt mēntēm mōr|tālĭă| tāngūnt, the first a is short.

This last quote illustrates a contributing factor of confusion: When it comes to scanning poetry, it is common to mark heavy syllables with a macron on top of the vowel. (The actual vowel quantities in the quoted verse are as follows: sŭnt lăcrĭmae rērum ĕt mĕntĕm mŏrtālĭă tăngŭnt.) Even if the author does not intend the vowels to be prolonged, if the reader is not sufficiently familiar with the distinction between vowel length and syllable quantity, this can be severely confusing, and I suspect that this is one of the main reasons why the “vowel-made-long-by-position” doctrine has been so regrettably persistent. Ideally, vowel length and syllable quantity should be marked in visually different ways. For example, a heavy syllable can be be marked with a macron extending over the whole syllable (and possibly leaving room for marks showing vowel length below):

Other methods that have been suggested is to place the macrons and breves below the relevant syllables, leaving room for diacritics above the vowels. The only drawback with methods such as these is possibly that they are not always easy to implement due to technical constraints of word processors.

How to Pronounce Syllable Quantity

In an idealised pronunciation, a heavy syllable should be pronounced twice as long as a light syllable. Statements to this effect can be found, e.g., in Quintilian (Inst. Orat. IX.4.47; an English translation by Rev. John Selby Watson exists.):

Rhythmos est aut par, ut dactylicus, una enim syllaba par est brevibus (est quidem vis eadem et aliis pedibus, sed nomen illud tenet: longam esse duorum temporum, brevem unius etiam pueri sciunt)...

In prose, it is of course not unrealistic to think that the exact ratio of the syllables would vary somewhat; in poetry, however, the constraints imposed by the meter would probably enforce a fairly exact 2:1 ratio between heavy and light syllables.

When learning to pronounce and appreciate the different lengths of heavy and light syllables, it might be a good idea, when practicing, to try to exaggerate the difference, and pronounce light syllables as quick as possible, while prolonging heavy syllables.

In the case of light syllables, this should not be any problems, as long as syllables are not inadvertently closed: special care should be taken not to lengthen stressed light syllables. For example, the word facere should be pronounced with three equally short staccato syllables: fă-cĕ-rĕ.

Heavy syllables with long vowels should not pose any problems either, since the long vowel easily can be prolonged indefinitely (or till the air runs out): eeeee-viiiii-taaaaa-baaaaa-tur.

Finally, we have heavy syllables with short vowels, i.e., syllables “long by position”. As we have seen, the vowel should not be lengthened. How then, should we prolong the syllable? The one element without which the syllable would be light is the final consonant, which closes the syllable, and it is this consonant, then, that will contribute to the length, and act as a sort of resting point for the voice. Example: cĭrrrr-cŭmmmm-spĕcccc-tum. (In the case of continuants, i.e. consonants which can be audibly prolonged – like /l, r, m, n, s, f/ –, this is done easily and intuitively. Stops – i.e. /p, b, t, d, c, g/ – can be prolonged as well, by simply postponing the release of the obstructed airflow. Thus, when I write specccc-tum, I do not mean that the long c should be pronounced in a stuttering fashion, but instead effectively as a prolonged silence.)

As an example of how the difference between the syllables can be performed in practice in accordance to this exaggerating method, I present this recording (again, bold text signifies heavy syllables):

1. ăr-mă-vĭ-rŭm-quĕ-că-nō-troi-jae-quī-prī-mŭ-să-bō-rīs
2. ī-tă-lĭ-ăm-fā-tō-prŏ-fŭ-gŭs-lā-vī-nă-quĕ-vē-nĭt
3. lī-tŏ-ră-mŭl-tvĭl-lĕt-tĕr-rīs-jăc-tā-tŭ-sĕ-tăl-tō
4. vī-sŭ-pĕ-rŭm-sae-vae-mĕ-mŏ-rĕm-jū-nō-nĭ-sŏ-bī-răm,
5. mŭl-tă-quŏ-quĕt-bĕl-lō-păs-sŭs-dŭm-cŏn-dĕ-rĕ-tŭr-bĕm
6. īn-fĕr-rĕt-quĕ-dĕ-ōs-lă-tĭ-ō-gĕ-nŭ-sŭn-dĕ-lă-tī-nŭm
7. ăl-bā-nī-quĕ-pă-trēs-ăt-quăl-tae-moe-nĭ-ă-rō-mae.

Note especially how the separation of syllables is performed without any regard for word boundaries: word-final consonants are moved to the following word in cases when that word begins with a vowel, so that the final syllable of the first word is opened.

In that recording, the time ratio is, on an average, 7:2, i.e., the long syllables were pronounced three and a half times as long as each of the short. So much for practicing. When read in a normal way, the long elements should of course not be overtly prolonged, but just enough to create a 2:1 ratio. In the case of syllables long by position, this means that the length of the closing consonant should be roughly equal to that of short vowel. To create the desired effect, in the mind of the speaker, the consonant should be felt as taking up a noticable time and as being disjoined, in some sense, from the following syllable.

How to realise the Ictus

As noted, the quantitative Latin verse is based on rhythm, that is, the meter is defined by a regular alternation of heavy and light syllables. Modern poetry, however, is typically based on the alternation of stressed and unstressed syllables. Hence, if Latin would be read out aloud just as if it were prose, it would typically, without training, not be immediately apparent for us that it is poetry at all. And if the reader furthermore does not carefully distinguish between long and short vowels, there is indeed very little of the meter left to discern.

For this reason, it is not uncommon to disregard the natural stress, and instead stress the syllables that bear the ictus, the beat of the verse, not dissimilar to how one may clap the hands to the rhythm of music. Listen to a scansion of the first seven lines of Vergil’s Aeneid, read with a (rather heavy) stress on the ictus, and with the natural word stress repressed. Its merit is of course that it serves to clarify the meter for us modern people, but it tends to feel rather monotone after a while, and I imagine ít would have sóunded sómewhat dísturbíng to a Róman.

How would our Roman have read it then? In fact, despite what some may think or teach, it is almost totally certain that poetry was not accented in any special way in antiquity, but instead that the natural word accents were used. To the best of my knowledge, there are no classical texts at all describing anything like a stressed ictus. On the contrary, the accounts we have where poetry and stress is discussed all support the natural stresses (with at the most a few adaptations.) For example, Quintilian, Inst. Orat. I.5.28:

Evenit ut metri quoque condicio mutet accentum: “pecudes pictaeque volucres”. Nam “volucres” media acuta legam, quia, etsi natura brevis, tamen positione longa est, ne faciat iambum, quem non recipit versus herous.

It indeed sometimes happens, we learn, that the meter affects the accent, for in the quoted verse (which is the end of a hexameter from Vergil's Georgics, 3.243), “volucres” is to be stressed on the middle syllable, not on the first, as would be the natural stress. The reason is that the meter forces the middle syllable to be heavy (long by position), instead of light as it would naturally be (due to the muta cum liquida.) But clearly, this is a special case, for, if the ictus should be stressed, this syllable would be accented anyway, and this explanation would have been superflous! The implication is instead that the rule that governs the placement of the word accent based on the length of the penultimate, rather than being ignored, as would be the case if the ictus were stressed, was so rigorously applied that the word accents could change when the meter enforced an artificial syllabification.

In general, our Roman would simply stress the words just as he would in daily speech. The meter would be apparent to him anyway, since it is made up by the innate variation of light and heavy syllables, which he would naturally and unconsciously make a clear distinction between. It should be noted, that, in Latin, the concept of syllable length is wholly separate from that of stress: a long syllable may perfectly well be unstressed while a short syllable is stressed, and vice versa. Listen to the same text read again, with exactly the same rhythm, but this time with natural word accents. Do you feel the meter?

Conclusion

Hitherto, in the sound examples you have heard, I have pronounced every line of verse separately, with a marked pause in between. Naturally, this is not the right way to recite a poem. Rather, the text should always be recited with the meaning of the words and clauses in mind, and a sentence that goes over several lines should be read without any disturbing pauses.

Please, listen to my attempt. Assuredly, there are other who can make a better performance, but I believe this at least is in the right directon.

1. Arma virumque canō, Troiae quī prīmus ab ōrīs
2. Ītaliam fātō profugus Lāvīnaque vēnit
3. lītora, multum ille et terrīs iactātus et altō
4. vī superum, saevae memorem Iūnōnis ob īram,
5. multa quoque et bellō passus, dum conderet urbem
6. inferretque deōs Latiō; genus unde Latīnum
7. Albānīque patrēs atque altae moenia Rōmae.

Another example, this time from book 6 of the Aeneid:

If you are interested in even more Latin poetry, you might also enjoy my recording of the first book of Ovid's Metamorphoses.

References

Latin Verse-Ictus and Multimodal Entrainment, Robert P. Sonkowsky and Franz Halberg, Electronic Antiquity, vol 8, number 2 (May 2005).

Vox Latina, W. S. Allen, 2nd edition 1978.

The Quantitative Pronunciation of Latin, and Its Meaning for Latin Versification, William Gardner Hale, The Classical Journal, Vol. 2, No. 3. (Jan., 1907), pp. 101-110.

A Graphic Device for Marking Syllable Quantity in Latin, Ellsworth David Wright, The Classical Journal, Vol. 11, No. 6. (Mar., 1916), p. 367.

Non oculis sed auribus: The ancient schoolroom and learning to hear the Latin hexameter, Andrew S. Becker, CPL Forum Online 1.1, Fall 2004.

Copyright 2007 by Johan Winge.