A treatise on the complicated matter of Latin Versification is beyond the scope of this article. But a few basics on this complex area, as aid for beginning and intermediatestudents, can be simply put:

It should be noted that these rules, although correct and useful, will never help a person read the dactylic lines of Vergil as they should be read. Listening to someone with a good sense of musicality and interpretation reading Vergil, is the best, and perhaps the only way to learn to read Latin poetry. Marking out the "long and shorts" as is common in schools, is NOT the way to read poetry, and may well prevent you from ever reading Latin verse in a musically satisfactory manner.

If as a student, you are inclined to doubt my warnings, just ask your Latin teacher (HS or college) to read a page aloud so you can get the gist of it. If the teacher reads well, listen carefully, but if the teacher fudges, it means he or she does not understand the verse as poetry endowed with sound and rhythm. Then you have to go it on your own, learning it (as with all important things) by your own effort.

1) Vowels in Latin are either LONG or SHORT, this refers to the length of time they take to be pronounced. A long is about twice as long as a short. [However there is no evidence that Romans counted out their "one-and.. two-and.. " 's in the exact manner of a modern elementary music teacher.]

2) Vowels can be LONG BY NATURE, which means they are "genetically" long in the tradition of Latin and its Indo-European parentage.

3) Or they can be stated as LONG BY POSITION, which means that a vowel which is followed by two consonants is generally "considered" or made long, but the following stipulations must be noted:

a) A vowel before another vowel or -h- is short: e.g. via, nihil

b) A diphthong, representing two fused vowels, is long, and also vowels contracted from diphthongs (inclúdo from claudo)

c) Any result of contraction is quite naturally long: e.g. nil from nihil

d) A vowel before -ns, -nf, and sometimes -gn is long: infans

e) A vowel before -nd, -nt is short: e.g. amant

f) A vowel followed by a stop consonant (-p-,-b-, -ph-, -d-, -t-,-th-, -k-, -g-, -gh-) followed by a liquid (-l- or -r-) may be considered long OR short in Classical poetry, although it is always short in early verse. [Rather than wrestle with this complicated piece of legislation, assume that the highly sonorous -l- and -r- are felt to be so vowel-like as to invalidate the "two consonant rule".]

g) If a consonantal -i- (=-j-) is preceded by -a-, -e-, -o-, -u-, the vowel is considered long, because of the consonantal function of the vowel: e.g. aio, maior, peius..

h) Final -s after a short vowel is often suppressed, thus removing one of the causes for length. In Post-Augustan authors final long -o is regularly shorted, e.g. in cases like ego, amo etc., but not in the dat. and abl. of 2 decl.. nouns. Generally a vowel before another vowel is considered short if not part of a diphthong; but not in cases like cuius, and forms of fio.

i) By a well established process, called The Law of Iambic Shortening (or in the grammarian's jargon Brevis Brevians), a long vowel is shortened if it is preceded in the word by a short vowel, and if it is preceded OR followed by a syllable which has the (prose system's) word-accent. The detailed operation of this rule is too complex to discuss here; the interesting thing is the operation of word-accent to convert an "iambic" word to two-shorts, apparently as a concession to the rhythm of conversational parlance. In Plautus this process is universal, rarer in later poetry.

j) The vowel before an elided -m (which had disappeared from spoken Latin early) simply disappears if the next word starts with a vowel, as well as an initial vowel when backward elided, e.g. est,, giving 'st. We thus actually have a three-valued logic for the Latin vowels: Long, Short, and "gone".

h) "Long by position" means long for verse, hence in verse the vowel must be of double length. But this does not mean that in prose the vowel is long, or that it must affect the STRESS accent. This curious double-standard is one of the things Romans lived with, with some unease. On the other hand a master like Vergil can make the verse-rhythm and the prose-rhythm work with and even against each other, creating a subtly moiré effect in verse.

(Actually the word "position" is a mis-translating of the Greek grammatical term "thesis" which means "convention, agreement"; ignorant Roman schoolmasters thought it came from the verb "ti-THE-mi" which can mean "place, put in position", hence the error.)

Footnote: LONG means "long", not stressed or made loud. Hence a long "-a-" is actually pronounced "-a-a", not as two separate sounds, but one long one which occupies twice the space of a single "-a-", as if legato. It is like a half-note in music counted out as double, in a sequence where quarter notes are normal, although we should not count it out exactly, since we know so little about the exactness of Roman music. The first line of Vergil's Aeneid, with a little latitude, is to be pronounced thus:

A-arma viru-umque cano-o Troiae-ae qui-i pri-imus ab o-ori-is

Unfortunately many teachers substitute stress for length in reading Latin poetry, probably because English is a stress-oriented language and long vowels are not normally used. But if you are going to read Latin, you should read it as the Romans spoke it, and there is no questions about the fact that when they specified a vowel as LONG grammatically, they meant "L-O-N-G" acoustically. To read Latin with stress substituting for length is wrongheaded, it makes Vergil sound like something conjured up on a rocking-horse, and misses the real sound of Latin verse, which can be quite lovely. (For a full discussion of this problem, see Section 13):. Stress.)

The verse form most commonly used in Latin is the "dactylic hexameter".

This is the standard line for all epic poetry in Latin, following the example of Homer closely. The dactylic line is also the first line of the two-line "Elegaic Couplet", which embraces the whole output of Ovid, Propertius Tibullus and others. It can be said without hesitation that the dactyl has its foot firmly in the body of Latin literature.

Dactylic hexameters use only two basic rhythmic structures, or "feet":

1) The dactyl: A long followed by two shorts, e.g." long-short-short", and

2) The spondee: A long followed by a long, e.g." long-long"

[Note: Dactyl comes from Gr. dactylos "a finger",. which has one long bone and two shorts, while Spondee in Greek means "libations, sacral processioning at a service", with the measured rhythm of our traditional wedding processional music. From such humble origins the grand terms of the grammarians often come!]

Now for some purely practical rules drawn from observation::

1) Every dactylic hexameter line, and every foot within a line, must begin with a long.

2) This long CAN be followed by a long, in which case we are into the next foot, which must of course start with another long.

3) Or it CAN be followed by two shorts, in which case are we again in a new foot, which must start with a long.

4) MOST lines end with the tried-and-true "heroic" cadence of a dactyl followed by a spondee (the last syllable of which can be either long or short, by a kind of Latin poetic license). Only one out of fifty or so verses will end with two spondees, usually introduced for a somber effect.

5) Anticipating a next-to-final dactyl will help you get the line in order; if it doesn't come out with a dactyl, you have made an error. (But then again, you may be right, since there are occasional spondaic closings.)

6) Often we are so busy searching for LONGS that we forget the other side of the equation: Also look for SHORTS. Since a short must be followed by another short, and that always precedes a long, "searching for shorts" is just as good, and in some ways better, than "looking for longs". Latin, as against Greek, is heavy in the frequency of its long syllables. Since there are many fewer shorts in a typical dactylic line, why not find them, and assume as a rough rule that the rest are long? [One incidental way of finding some of the shorts is by appeal to the Prose Accentuation system; if the accent is third-from-end, the second from end must be a short, or it would have had the accent!]

And now for some notes drawn from observation of the stylistics of the Roman poets:

1) Even a skilled poet like Vergil can occasionally stage a line which is completely dactylic (see Aeneid 8, 596 for the sound of trampling hooves of horses), or wholly spondaic (as in 8, 452 demonstrating great bodily effort). Or a line can be made to mumble and murmur (3, 658) by means of piled up elisions. But these are the special effects of a master, and relatively rare.

2) The exact positioning of the two types of feet in the line is a technique which Vergil works with sensitively. Lines are shaped by the rhythms, and the artistic shaping of a line is among the things which make Latin poetry work.

3) Caesura, the separation of two words in the median portion of a dactylic line by "cutting" (caedo) the foot, is an accepted artistic practice derived from Greek epic poetry. A word-break after the first syllable of the third foot is the most common place, and is called the strong caesura,

the most common type. Variants of the location of caesura are noted in manuals on versification, but the most important thing is to read each line carefully and determine if the word-break (wherever it occurs) works with or against the metrical rhythm artistically. In many cases the effect which the writer is seeking is easier to comprehend than the discussion in manuals on Latin Versification. Put the other way around, the technical description is liable to be meaningless except as a statistic, unless it can be coupled with artistic meaning in a given line if poetry.

It may occur at this point, since one can spot the vowels which are LONG BY POSITION, and the DIPHTHONGS by eye, that it might be useful to know which vowels were LONG BY NATURE. If you have to ask this question, the simple answer is "Learn them from your high-school textbook where longs are marked" or "Look them up in the dictionary", which not going to be very helpful if you have to ask the question at this point. But how does the dictionary know which vowels are long?

In the late 16th c. a Dutch M.D. named Smetius published a dictionary of all commonly used Latin words, following each entry with a line of poetry taken from an author whose classical taste was beyond question. This book and its successors give the information which we now use: Long vowels are documented from the library of Augustan Latin poetry, which is the best source of authenticity, so we are dealing with a veritable circular argument. In a very few cases, Roman grammarians like Aulus Gellius (whose book neatly titled "Attic Nights" has disappointed generations of students expecting something racy) give additional information about lengths which the Romans were not sure about. [A full account of such grammatical cabalism is to be found in W.M. Lindsay"s remarkable book The Latin Language, Oxford l894, a short perusal of which will make it clear that even the Romans had a great many questions about the longs and shorts of their native language.]

One might note that the severe practice of making high-school students memorize all long vowels accurately, and then encourage them to pronounce these longs as LOUD or stressed, is not only unnecessary, but also unhistorical, unesthetic and unrewarding.


1) Be aware of the rules and observations above, but do not expect them to teach you everything you need to know about reading Latin verse.

2) Try to find a professor who reads Latin with dramatic (but not stagy) gusto and feeling. If he reads as if he were in an accelerated rocking-chair, go get someone else to read to you. Get the feeling of the line BY EAR. As in music, it is the ear which gets the measure and the up-beat, not the memorization of rules. Poetry is "language-with-rhythm", and if you can't get the ear into the process, you can't hear the poetry.

3) Read many lines OUT LOUD, taking a chance about correctness, and persevere until you hear lines coming out "right". If dactylic hexameters come out with that final resounding, epic dactyl cadence, you are probably on the right track. Continue, get the feel of the process, and later if you have a question about a vowel, ask someone or consult a dictionary for lengths. Recall that those who read Latin verse fluently, page after page with delight and feeling, got their technique for reading from the lines on the page and from their ear. There probably isn't enough time in life for most of us to look up each word for the LONGS, life is just too short. You can do it yourself by learning basic rules and jumping into reading with both feet, and taking a chance. Your ear will eventually correct you if something is terribly wrong.

Although standard Latin instruction teaches students how to puzzle out lines, charting them on paper with longs and shorts scribbled in over the text is clearly not the highroad to poetic appreciation. What you want is the ability to read page after page with an easy voice, so you can grasp the rhythms while you are imbibing the meaning of the Latin words. That is what reading Latin poetry is all about. It may take time, but it's worth it!

But all Latin poetry is not written in dactylic hexameters. The first variant to compare is an Ovidian elegaic couplet. The first line is a standard dactylic hexameter, the second line, called a "pentameter" for no especially cogent reason, starts off the same way but comes to an unexpected pause in the middle, often coinciding with some point of surprise or pathos. After this interruption or hiatus the line starts off again and resumes what feels like a normal dactylic line with a typical cadence of conclusion. If you have become familiar with the basic dactylic line, and when reading the second line of an elegaic couplet (which is conveniently indented in all printed texts) you expect a break or hiatus toward the middle of the line, you can probably make the verse form out for yourself by ear.

After the central stylically observed "break" or 'caesura' in the "pentameter", the line takes up again with a dactyl, followed by another dactyl, and then a spondee. This is a regular pattern, it concludes the line with a certain finality, and establishes the couplet as a self-contained unit, rarely reaching into the next couplet for its meaning. This is indeed a rudimentary description of the elegaic couplet, but it should be sufficient to work with initially; for a detailed account of the couplet in various authors and periods, complete with detailed statistics, consult any modern manual on Latin versification.

But dactyls are not all that there is to say about metre in Latin. The "Lyric poets" Catullus and Horace use a variety of metres drawn from the common practice of the Greek poets from the time of Sappho to the Alexandrians. Catullus and Horace reach into the library of Hellenistic metrification, and produce lovely little poems in very un-Roman metres, as Horace was quick to admit in print. The sharp and biting hendecasyllabic line with its uneven eleven syllables, the graceful stanzas of the Sapphic and Alcaic metres, even (if only once) the two longs pursued interminably by two shorts in the Sapphic ionic a minore.... these add spice and dimension to Latin poetry, but at the cost of having to study metre carefully before you can be at home in Lyric Poetry. Any one of the many monographs on ancient metrics will bail you out with the lyric metres if you have perseverance, but it is your ear which must eventually supply the musicality to make real the printed schematics of these complicated, Greek-based systems of versification.

Return to Latin Background index

William Harris
Prof. Em. Middlebury College