Open Book Publishers logo Open Access logo
  • button
  • button
  • button
GO TO...
book cover

1. The life of Ovid

© William Turpin, CC BY

Publius Ovidius Naso was born in 43 BC, in Sulmo (modern Sulmona), in the rugged mountains of the Abruzzi about a hundred miles from Rome. His family, which must have been locally prominent and relatively wealthy, were Roman citizens of equestrian rank and seem to have intended Ovid for a political career in Rome. Ovid was a conspicuous success as a student of rhetoric at Rome, went on a tour of Greece, and held at least one minor magistracy in Rome before turning to poetry as a full-time occupation. He married at least three times, and had a daughter and two grandchildren.

Fig. 1 Statue of Ovid, in Constanta, Romania (ancient Tomis). Wikimedia,ţa,_Romania.jpg

In AD 8 he was banished by Augustus to the remote Greek city of Tomis (modern Constanta), on the Black Sea coast in what is now Romania. According to Ovid there were two reasons for his exile: his Ars Amatoria had given offense, and he had committed a mysterious error, perhaps connected with the imperial house (Augustus' granddaughter Julia was exiled for adultery in the same year). Despite much pleading Ovid was never allowed to return from Tomis, and died there in (probably) AD 17.

Ovid apparently began writing his Amores in 26 or 25 BC; he tells us that he wrote poems about the lover he calls Corinna as a young man of 17 or 18. These poems were originally published in five books, but were subsequently republished in the edition we now have, in three books, sometime after 16 BC. His other early works, all largely concerned with love affairs and/or women, are difficult to date precisely, and no doubt overlapped with the writing of the Amores: the Heroides is a collection of letters written by fictional heroines; the fragmentary Medicamina Faciei Femineae concerns female cosmetics; the Ars Amatoria is a didactic poem about how to conduct love affairs, and the Remedia Amoris is about how to end them.

Ovid’s greatest work is the Metamorphoses, an epic poem on mythological transformations. He also wrote the Fasti, concerned with the religious calendar, and the Ibis, an invective against an unnamed enemy. During his years of exile he wrote the Tristia and Epistulae ex Ponto. Two lost works are a drama, the Medea, and a translation of Aratus’ astronomical poem, the Phaenomena. All his surviving works except the Metamorphoses are in elegiac couplets.

2. The Amores

© William Turpin, CC BY

In writing poems in elegiac couplets about a love affair (or affairs) Ovid was firmly within an established tradition. The elegiac couplet (on which see the next section) was originally used, first by the Greeks and then by the Romans, for short epigrams, often on erotic subjects. Catullus (c. 84 to 54 BC), wrote not only epigrams, but longer poems in elegiac couplets; he also gave to many of his poems a unifying story, about a difficult love affair with the woman he called Lesbia.

He was followed by C. Cornelius Gallus (c. 70 to 27 or 26 BC), who seems to have written four books of love poetry exclusively in elegiac couplets, probably called Amores. Almost none of Gallus’ verses survive, but they depicted his affair with a famous actress of the day named Cytheris, whom he calls Lycoris. Gallus seems to have done much to establish the conventional figure of the poet as the broken-hearted lover; the allusions in Vergil’s Eclogue 10 suggest that in one poem he portrayed himself wandering in the woods and carving his and Lycoris' names onto tree-trunks.

Perhaps the most important of Ovid’s immediate predecessors was Sextus Propertius (born between 54 and 47 BC; died before 2 AD). Propertius published four books of elegies, the first appearing around 28 BC, a few years before Ovid’s first poems in the genre. Like his older contemporaries Vergil (born 70 BC) and Horace (born 65 BC), Propertius came to be a member of the circle of Maecenas, the political advisor of Augustus, but at least in his first three books he rebelled against Augustan values more than Vergil and Horace ever did. Most of Propertius’ poems concern a romantic affair with a woman he calls Cynthia, and in many of them the poet is portrayed as desperately, even morbidly, uncertain of her affections. Propertius wrote with self-conscious artifice (he claimed to be a Roman Callimachus), deploying mythological examples that are often obscure.

A near contemporary of Propertius was Albius Tibullus (born between 55 and 48 BC; died in 19 BC), who wrote two books of elegies, the first at about the time of Ovid’s first Amores. Tibullus wrote poems concerning three different love affairs, with women he calls Delia and Nemesis and with a young man he calls Marathus. Tibullus’ poems are much less mythological than those of Propertius, and the emotions he depicts are much less tortured. A second poet associated with Tibullus was Sulpicia, the niece of Messalla Corvinus. Six of her elegies are preserved in the third book of the Tibullan corpus, and describe (without many details) an affair with a man she calls Cerinthus.

Ultimately, perhaps, evaluating Ovid’s Amores requires a first-hand knowledge of the tradition in which he was working; it is a truism of Latin scholarship that Ovid plays with, even mocks, the conventions of his predecessors. But for practical reasons the Amores make a good introduction to the genre; Ovid's Latin is relatively straightforward, at least compared to that of Propertius, and he offers a livelier account of the traditional Latin elegist’s difficult love-life than either Tibullus or Sulpicia. The figure of the poet-lover that Ovid presents in the Amores is also a new departure in the western literary tradition, worth attention in its own right: we get our first hapless, light-hearted, insensitive, and selfish womanizer.

The reader approaching the Amores for the first time should be alert to at least three features of the poetry. Most important, though most elusive, is the question of tone, though it is not easy to develop a sense for the essential flavor of the Latin poetic idiom. We are so accustomed to high-flown language and ponderous allusions in our Latin that it is not easy to see when a poet is playing with the traditional language and mythology, but Ovid is (in my view) the best place to start: when he writes of abandoning the epic tradition (1.1), of the lover as a warrior (1.9), or the myth of Aurora and Tithonus (1.13), we get a clear sense of the playfulness possible in Latin poetry.

The second thing to be aware of in each poem is the structure of the “argument.” Ovid has traditionally been regarded as someone who wrote verse with such facility that he simply kept on going, making the same point over and over again with a kind of effusiveness that belies close analysis. But while it is certainly true that he does not write with the fanatical self-control of Vergil or Horace, it is also a mistake to ignore his careful attention to the construction of his poems. It is important to be aware of the way each poem develops: some thoughts lead naturally to others, and at some places the poet jumps to a new idea, but there is always a reasonable representation of coherent thinking. Moreover it is usually worth asking oneself how (or whether) the final couplet works as a satisfactory conclusion to each poem; there is often (perhaps always) a kind of punch-line at the end, and getting the point there is often the key to getting the poem as a whole.

Finally, it is worth remembering that poetry books in Ovid’s day were published with careful attention to their overall shape. Each poem was meant to be read, at least in part, as an element in the broader narrative of the poetry book, so it can be illuminating to ask how each poem relates to the ones preceding it, and serves as an introduction to the ones that follow.

3. The manuscript tradition
of Ovid’s Amores

Bart Huelsenbeck, with the assistance of Dan Plekhov

© Bart Huelsenbeck, CC BY

R Paris, BnF lat. 7311. 9th century. (Ars amatoria; Remedia amoris; Amores Epigr., 1.1.3–1.2.19, 1.2.25–50).1

P Paris, BnF lat. 8242. 9th century. (Heroides [incomplete], Amores 1.2.51–3.12.26, 3.14.3–3.15.8).2

S St. Gall, Stiftsbibliothek 864. 11th century (Amores Epigr. 1.6.45, 1.8.75–3.9.10).3

Y Berlin, Staatsbibliothek, Hamilton 471. 11th century (Ars amatoria, Remedia amoris, Amores).4

The manuscript witnesses to the Amores fall into two groups: the four earlier manuscripts (vetustiores) listed above, and an abundance of later manuscripts, referred to collectively as recentiores and dating to the 12th century and after. Franco Munari (1951) and E. J. Kenney (1961), who produced the first modern critical editions of the Amores, regarded these two groups of manuscripts (older and more recent) as representative of two independent lines of transmission. The dates of the manuscripts seemed to correspond closely with two separate pedigrees: the vetustiores were traced back to a now lost hyparchetype, called α, and the recentiores to a second lost hyparchetype, called β. (The independence of the β manuscripts is guaranteed by the presence of verses [1.13.11–14; 2.2.18–22, 25–27] that are absent from α manuscripts.)

This view of the data, current at the time of the first edition of Kenney’s Oxford Classical Text (1961), kept the textual transmission of the Amores relatively simple. The α branch was particularly straightforward since it consisted of only a few extant manuscripts. Making matters simpler still, within the α branch P was thought to be a copy of R. The perception that P descended directly from R could have eliminated its relevance in the eyes of textual critics, except for the fact that the portion of R (designated (R´)) containing the Amores is almost entirely lost. Consequently, P, which was believed to reflect α through the intermediary of R, needed to be used. S was, and still is, considered an inferior manuscript. It contains readings thought to be imported from the β branch, and, except for one passage (Am. 2.8.7), it does not offer good readings that cannot be found elsewhere. S, in the words of Kenney (1962:8), is “in an intermediate state of depravation”—textual critics frequently impute moral characteristics to manuscripts: the nature of its text shows that it belongs to the α family, but the later date of S (11th cent.) means that its text has “degenerated” and resembles in some particulars the text of β manuscripts.

A discovery and further research were soon to complicate this rather simple reconstruction of the tradition. It was not long after the appearance of the first edition of Kenney’s OCT that Munari (1965) first called attention to the Hamiltonensis (Y), which hitherto had been ignored. (Credit for the rediscovery belongs to Helmut Boese.) Because of a cataloguing error, which dated Y to the 14th century, neither Munari nor Kenney had taken Y into account in their editions.

The accession of Y to our knowledge about the tradition has had two important results. First, Y is a valuable independent textual witness—better than S and just as valuable as P. Second, and equally important, Y made obvious the “fog of unknowns” that still envelops a large mass of the tradition. With Y in the picture, there were now two manuscripts of the 11th century—but their texts were quite different from each other. Y was not a “depraved” representative of the α branch, as S was believed to be. Y had its own authority, offering readings now in agreement with α, now with β. It even had authoritative readings not found anywhere else (most notably, Am. 1.10.30, licenda).

The arrival of Y served as reminder of something else: the tradition is not bifid, though it has been represented as such. The division into two major branches of transmission (= a bifid stemma), corresponding to earlier and later groups of medieval manuscripts, is a convenient means to organize the tradition, but the reality is quite different. Y had shown that the dates of the manuscripts did not closely correspond with the nature of the texts that they offered. A good reading—a good reading attested nowhere else—could appear in a later manuscript. It had happened in the case of Y, and there was nothing to say it could not happen with a β manuscript. The β manuscripts vary widely in date and their relationships to each other have not been traced out, and for good reason: they are not a closely connected group. They are not one family, but amount to an intertanglement of many individual families. Kenney appreciated the complexity on the β side when he characterized β as a “convenient fiction” (1962:25). Later, he shrewdly observed (1974:134) that it would be more accurate to refer to the β branch as “non-α.” The idea is that, whereas the α manuscripts are manifestly related, the β manuscripts do not derive from a single ancestor; their membership to the same group is solely by virtue of the fact that they do not derive from α.

Reconsideration of P’s relationship to R has added a further wrinkle to the tradition’s history. D. S. McKie (1986) forcefully argued that, contrary to what had been thought, P does not derive from R. Belief that P was copied from R came about through a tentative suggestion made by Tafel (1910) that eventually was taken over as fact; decades later Goold (1965) corroborated the idea. Because P’s text of the Amores begins precisely where R’s text goes missing it was assumed that the scribes of P must have taken this portion of R to serve as their exemplar (model). Furthermore, the date of P, which had been set as “late 9th or 10th” century, is earlier than often supposed. As was demonstrated by B. Bischoff (1961), P belongs to a group of manuscripts copied at the French monastery of Corbie in the period 850–880 (see Huelsenbeck 2013). R, P, S, and Y are all independent witnesses to α.

Therefore, recent developments in the study of the textual transmission of Ovid’s Amores yield a more intricate, dynamic, and open-ended stemma. This stemma attempts to reflect the current state of our knowledge, showing what we know and do not know.

  • Non-α (formerly β) is not a single textual family: the tradition is not bifid.
  • In the stemma α and non-α are not shown to connect because non-α represents multiple families with intertangled connections. How and when α and non-α connect are unknowns.
  • R, P, Y, and S are independent witnesses to α, though S draws some of its text from non-α.
  • The copying dates of R and P are close.

4. Select bibliography

Adams, J. N. The Latin Sexual Vocabulary. Baltimore: The John Hopkins University Press, 1990.

Barsby, John A. Ovid's Amores: Book I. Oxford: Oxford University Press, 1973.

Booth, Joan. “The Amores: Ovid Making Love,” in A Companion to Ovid, ed. Peter E. Knox, 6177. Blackwell Companions to the Ancient World. Oxford: Blackwell, 2009.

Boyd, Barbara Weiden. Ovid’s Literary Loves: Influence and Innovation in the Amores. Ann Arbor: University of Michigan Press, 1997.

Huelsenbeck, Bart. “A Nexus of Manuscripts Copied at Corbie, ca. 850–880: A Typology of Script-style and Copying Procedure,” Segno e testo 11 (2013): 287-309.

Jestin, Charbra Adams and Phyllis B. Katz. Ovid: Amores/Metamorphoses, selections. Wauconda: Bolchazy-Carducci, 2000, 2nd edn.

Kenney, E. J. “The manuscript tradition of Ovid’s Amores, Ars amatoria, and Remedia Amoris,” Classical Quarterly 12 (1962): 1–31.

Kenney, E. J. P. Ouidi Nasonis Amores, Medicamina faciei femineae, Ars amatoria, Remedia amoris. Oxford: Oxford University Press, 1995 (orig. 1961), 2nd edn, reprinted with corrections.

McKeown, J. C. Ovid, Amores: Text, Prolegomena, and Commentary in Four Volumes. Liverpool and Wolfeboro: Francis Cairns. 1987–.

McKie, D. S. “Ovid’s Amores: The Prime Sources for the Text,” Classical Quarterly 36 (1986): 219–238.

Munari, Franco. Il Codice Hamilton 471 di Ovidio: Ars amatoria, Remedia amoris, Amores. Rome: Edizioni di storia e letteratura, 1965.

Pasco-Pranger, Molly Claire. “Duplicitous Simplicity in Ovid, Amores 1,” Classical Quarterly 62 (2012): 721–730.

Raven, D. S. Latin Metre: An Introduction. London: Faber and Faber, 1965.

Ryan, Maureen B. and Caroline A. Perkins. Ovid’s Amores Book One, a Commentary. Norman: University of Oklahoma Press, 2011.

Showerman, Grant. Ovid, Heroides, Amores. Revised by G. P. Gould. Loeb Classical Library. Cambridge, MA: Harvard University Press, 1977.

Stapleton, M. L. Harmful Eloquence: Ovid’s Amores from Antiquity to Shakespeare. Ann Arbor: University of Michigan Press, 1996.

Tarrant, R. J. “Ovid, Amores, Ars amatoria, Remedia amoris,” in Texts and Transmission, ed. L. D. Reynolds, 259–262. Oxford: Clarendon Press, 1983.

Volk, Katharina. Ovid. Malden and Chichester: Wiley-Blackwell, 2010.

West, David Alexander. “Amores 1.1-5,” in Ancient Historiography and its Contexts: Studies in Honour of A. J. Woodman, eds. Christina S. Kraus, John Marincola, and Christopher Pelling, 139154. Oxford: Oxford University Press, 2010.

Wilkinson, L. P. Ovid Recalled. Cambridge: Cambridge University Press, 1955.

5. Scansion

© William Turpin, CC BY

Since the Amores may well be among the first Latin poems a student encounters, it may be helpful to provide a brief introduction to the rules of Latin prosody (the quantity of individual syllables) and to the reading aloud of elegiac couplets. For fuller discussion see D. S. Raven, Latin Metre: an Introduction (London: Faber and Faber, 1965).


Whereas English meters are based on a word’s accent (“Múch have I trávelled in the reálms of góld”), Latin meters are based on quantity; what matters most is whether syllables are long or short.

For most of us the obstacle to reading Latin verse aloud is that we have not learned the quantities of Latin very well. All diphthongs are normally long by nature, but individual vowels can be either long or short, though a vowel followed by another vowel not in a diphthong is normally short. Ideally we would all know, say, that the first syllable of miles was long and the second one short, but in practice we are often uncertain, or even wrong, and it sometimes necessary to consult a dictionary solely to ascertain the quantities of a word.

An additional problem is that it is often necessary to know the meaning of a Latin word before one can know its prosody. Latin has a number of virtual homonyms, distinguished only by their quantity, such as lĕvis (“light”) and lēvis (“smooth”). Much more common are the words whose form is identified only by their quantity: puella can be nominative singular or ablative singular, cīvis can be nominative or genitive singular or accusative plural, and manus can be nominative singular or nominative or accusative plural, etc. In such cases it is almost impossible to scan the line without also establishing its sense.

On the other hand the endings of Latin words provide us with a large collection of easily learned quantities: with a review of the basic declensions and conjugations it is not difficult to learn that the o of amō is long, and that the i of trādit is short, or that the ō and īs of puerō and puerīs are long.

Other syllables with easily identifiable quantities are those which, though short by nature, become long by position because of the consonants that follow them. The most obvious instances are when vowels are followed by double consonants (ll, mm, nn, pp, ss etc.), and such words are also the easiest for a reader to speak correctly; in Latin there was a clear difference between the L-sounds in malus and bellum, and it is easy to make this distinction aloud once alerted to it (MAL-us vs. BEHL-Lum). More generally, a short syllable can be long by position when followed by any two (or more) consonants together (except h), or by x and z, which were each the equivalent of two consonants.

But before the following combinations of consonants the preceding short syllable can remain short:

  • bl, br;
  • cl, chl, cr, chr;
  • dr;
  • fl, fr;
  • gl, gr;
  • pl, pr;
  • tr, thr.

However, a syllable cannot remain short when the two consonants following it belong to different parts of a compound abrumpo), or to different words (et refer).


A further complication in reading aloud is the fact that a vowel or a vowel + m at the end of a word is usually suppressed (“elided”) when the next word begins with a vowel, or h + a vowel. This occurs even if the elided vowel would have been long.

āstĭtĭt īll(a) āmēns ālb(o) ēt sĭnĕ sānguĭnĕ vūltū (Am. 1.7.51)

nēc tē dēcĭpĭānt vĕtěrēs cīrc(um) ātrĭă cērae (Am. 1.8.65)

A failure to elide (hiatus) is rare.

The elegiac couplet

The Amores are all written in elegiac couplets. This meter consists of a line of dactylic hexameter, the meter of epic poetry, i.e. six dactyls (― U U) or spondees (― ―), followed by a line of dactylic pentameter, i.e. five dactyls or spondees (with one of the spondees divided into two). The basic scheme is as follows:





U U |

U U |

U U |

U U |

U U |

― x



U U |

U U |

― // U U |

U U |

In the hexameter line the fifth and sixth feet are almost always a dactyl and a spondee (the last syllable of each line is technically anceps, i.e. it can be either long or short, but for practical purposes the lines can all be read as if the last syllable is long); thus each line can be expected to end ― U U / ― ―. The first four feet can be any combination of dactyls and spondees, and it is here that a knowledge of prosody becomes important.

In addition, the hexameter line almost always has a break between words in the third foot, most commonly after the first beat (whether of dactyl or spondee). This is called a strong caesura, e.g.

Iam super oceanum ‖ vĕnit a seniore marito (Am. 1.13.1)

Sometimes the break occurs after the second beat of the third foot (which must be a dactyl), giving a kind of syncopated feel to the line. This is the so-called “weak” caesura, e.g.

quo properās, Aurōră? ‖ mănē: sic Memnŏnis umbris (Am. 1.13.3)

The first half of the pentameter line can be thought of as the first part of a hexameter line extending to a strong caesura. As in the hexameter line spondees can be substituted for dactyls in the first two feet. The second half of the pentameter essentially repeats the first, but here there are no spondees. As with the hexameter line, the last syllable of the pentameter is anceps, i.e. it can be either long or short, but for practical purposes each pentameter line can all be read as if the last syllable is long. (I cannot find this explicitly stated in the reference books).

Reading aloud

Despite the apparent complexities, elegiac couplets are reasonably easy to read aloud. The key, in my view, is to become thoroughly at home with the basic unit of ― U U | ― U U | ―, which in its pure form provides the second half of the pentameter line, and which with spondaic variation provides the first half of the pentameter line and begins the vast majority of the hexameter lines. This, combined with the near certainty that the last two feet of the hexameter lines will be ― U U | ― ―, makes it possible to guess how most of Ovid’s couplets should be scanned, even if one’s grasp of basic Latin prosody is weak. It is important, of course, to be alert to those quantities which can be known in advance, such as diphthongs, certain word endings, vowels followed by double consonants, and vowels followed by more than one consonant, while remaining alert to the exceptions mentioned above.

I suggest practicing by beginning with the easiest section to scan, reading the second halves of all the pentameter lines in a poem; here there are no variations from ― U U | ― U U | ― and it is usually easy enough to see where the second halves of the lines begin. Follow this by reading the pentameter lines complete; the first two feet will offer some variation, but there are only four possible combinations for the first half of a pentameter:

― ― |

― ― |

U U |

U U |

― ― |

U U |

U U |

― ― |

Practicing the pentameter lines should make the hexameter lines much easier. Most lines will have a strong caesura, and will thus offer exactly the same four possibilities as the first half of the pentameter line. Following the strong caesura there will be either one long beat or two short ones to complete the third foot. The fourth foot will be either a dactyl or a spondee, and is thus usually the hardest foot to scan, but the fifth and sixth feet will almost certainly be a dactyl and a spondee. Lines with a weak caesura of course work slightly differently: the third foot will be a dactyl, with the caesura coming between the two short beats.

To introduce this approach to reading aloud, I print here a modified text of Amores 1.1. I have introduced gaps in the text to identify caesurae, all of which are strong caesurae. I have also put elided syllables in parentheses. In theory this should make it possible to follow the procedure suggested above with relative ease, so that unknown quantities can be deduced rather than looked up.

Arma gravī numerō violentaque bella parābam
ēdere, māteriā conveniente modīs.

pār erat inferior versus; rīsisse Cupīdō
dīcitur atqu(e) ūnum surripuisse pedem.

“quis tibi, saeve puer, dedit hōc in carmina iūris? 5
Pīeridum vātēs, nōn tua, turba sumus.

quid, sī praeripiat flāvae Venus arma Minervae,
ventilet accensās flāva Minerva facēs?

quis probet in silvīs Cererem regnāre iugōsīs,
lēge pharetrātae virginis arva colī? 10

crīnibus insignem quis acūtā cuspide Phoebum
instruat, Āoniam Marte movente lyram?

sunt tibi magna, puer, nimiumque potentia regna:
cūr opus adfectās ambitiōse novum?

an, quod ubīque, tuum (e)st? tua sunt Helicōnia tempē? 15
vix etiam Phoebō iam lyra tūta su(a) est?

cum bene surrexit versū nova pāgina prīmō,
attenuat nervōs proximus ille meōs.

nec mihi māteri(a) est numerīs leviōribus apta,
aut puer aut longās compta puella comās.” 20

questus eram, pharetrā cum prōtinus ille solūtā
lēgit in exitium spīcula facta meum

lūnāvitque genū sinuōsum fortiter arcum
“quod” que “canās, vātēs, accipe” dixit “opus.”

mē miserum! certās habuit puer ille sagittās: 25
ūror, et in vacuō pectore regnat Amor.

sex mihi surgat opus numerīs, in quinque resīdat;
ferrea cum vestrīs bella valēte modīs.

cingere lītoreā flāventia tempora myrtō,
Mūsa per undēnōs ēmodulanda pedēs.