Do-It-Yourself: How To Write Latin Verse

Harry C. Schnur

The Classical Journal, 52 (1957), 353-357

[Note of web editor: a bibliographical note listing some important tools for Latin verse composition has been posted at the end of this article.]

Writing Latin verse demands compliance with certain fixed rules: it therefore challenges both intellect and imagination, very much like a chess problem, and provides a similar kind of satisfaction, with the added bonus of a modest feeling of creative achievement. This pursuit immensely increases our understanding of the Roman poets' technique, difficulties and felicities: of meter, rhythm and sound. As a do-it-yourself hobby, moreover, it is both independent of the weather and much less costly than, say, collecting incunabula or Old Masters.

Our equipment, in addition to some familiarity with the Roman poets, is a dictionary and a Gradus ad Parnassum.1 Since hardly anyone nowadays is a native speaker of Latin, few of us can be sure of all Latin vocables and quantities; here, we rely on the dictionary, which will also tell us whether a given word is found in good classical usage. The Gradus, in addition to quantity, gives us synonyms and antonyms, cites relevant lines or phrases, and also lists adjectives grouped according to meter and declension.

The beginner should start with translation, not independent composition, bearing in mind the golden rule of all translation: reproduce ideas, not words, and in an idiomatic manner. Literal transposition, in prose or verse, from English into Latin will not result in Latin. "I could not care less, Caesar, whether you approve of me or not" - non minus curare possim utrum me probes necne will not do. Excessive reliance on the Gradus, on the other hand, will produce a mere cento; a middle course is indicated. And let us shun the unprecedented epithet: vegetabilis is found in late writers, but "our vegetable love" cannot be rendered as amor vegetabilis, nor "une nuit blanche" as nox alba.

I. The Hexameter

The instructor2 will begin by rearranging hexameter lines - first without, then with elision - in prose word order. Sed tamen da nobis, Tityre, qui sit iste deus; aspice, iuvenci referunt aratra suspensa iugo. The student will thus grasp the first essential: begin your hexameter at the end, and the rest of the line will fall into place. Suppose he ends the above lines (acceptable metrically) qui sit deus iste and aratra iuvenci: he will soon discover that the remaining words do not fit the line, and he will eventually, by trial and error, arrive at the solution. This exercise can be increased in difficulty by having sentences extending over several lines: Meliboee, hic vidi illum iuvenem, cui nostra altaria quotannis bis senos dies fumant. The student who has reached this stage is ready to compose his first hexameter.

What is a hexameter? It is emphatically not merely a line consisting of five dactyls or spondees, and a trochee (or spondee). If I may perpetrate two monstra: Juppiter optime maxime, dulcia munera dona, and: eheu! priscas nunc amisi libertates, it will be seen that they conform to the metrical definition. However, the first line is not merely an intolerable jingle but also a partipes: each word is a self-contained foot. As a consequence, the essential charm of the hexameter line, the clash between verse ictus and word accent required in the second and third feet, is absent. The second line is also largely a partipes (which could be partly remedied by placing nunc before priscas) and all spondees. Moreover, it terminates in a four-syllable word, which is normally undesirable and permissible chiefly in the case of proper names (Melicertes, Appenninus).

We shall, then, avoid all-spondee lines (and spondees in the fifth foot).

Rhythm should vary from line to line; this depends on the sequence of dactyls and spondees which, in turn, produces major and minor pauses (caesura, diaeresis) where sense and sound require them. Our first horrible example of partipes has, if you will, five breaks (after each foot!), which is too many. The pause should not follow an unimportant or monosyllabic word, as in: vir bonus et probus in/pacto nam foedere constat ("a gentleman keeps his word"). We afford some remedy to this line by recasting it: vir bonus in pacta/nam condicione manebit.

This pause after the first length in the third foot ("masculine" caesura), while good and very frequent, splits the line into two equal parts and therefore should not be employed in too many successive lines. A pleasant variation is the double break, after the first syllable of the second and the fourth foot. It imparts a graceful, tripping rhythm to the line: ferte simul/Faunique pedem /Dryadesque puellae. The "bucolic" pause (after the fourth foot) should be accompanied by a "masculine" caesura in the third foot: quo fugis, a, demens/habitarunt/di quoque silvas. The "feminine" caesura (after the trochee in the third foot) should be used only for special effect, as in this melodious line: spargens umida mella soporiferumque papaver.

How not to do it is shown by these lines of ours ("Examination Jitters"): papiliones in/stomachis volitant/hebetesque | mentes torpescunt, et solvit viscera terror. First, literal translation of un-Latin idiom; second, principal break after unimportant monosyllable; third, secondary break in fifth foot with line terminating in four-syllable word: in the second line, inversion of the first two words would be better. It will be seen, then, that pouring the requisite number of long and short syllables into a line does not by itself make a hexameter: the garment must be made to fit.

Square Pegs

Many words will not fit into the dactylic meter, because they cotain irremediable cretics or tribrachs. For instance, the short second syllable of angiportus is embedded between two longs; the second, third and fourth syllables of Alcibiades are short. We see now why the nom. and acc. pl. of arbor cannot be used - which is why the poets employ arbusta.

Here is one way of getting around this difficulty; there are others. Scipio was scanned as a cretic in Vergil's time, which is why he used Scipiades; but we may avail ourselves of "Silver Age" usage in shortening the final -o of such names as Curio (Lucan). Next, a succession of shorts may be remedied by final elision (ecthlipsis), or the last syllable may be lengthened by a following consonant. Thus, hominis must be followed by a word with initial consonant; homines (an anapaest) can be freely used, whereas hominibus is impossible, and something like humanis or mortalibus must be substituted. Nothing can be done, of course, about internal cretics: no form of latinitas can be used. Remember also that the end-syllables -um, -am, -em, -im, while short by nature, cannot be employed as shorts in the hexameter since, if followed by a consonant, they become long, while followed by a vowel they must be elided. Partial remedy exists in, the case of Greek words or names: where precedent can be found, Greek accusatives ending in -n or vowels may be substituted (Periclên, Anchisên, Athon; Delôn, Parin; Orphea). In other cases, the plural is used for the singular: precibusque oracula poscas; quaesitum oracula rebus.

This device also helps us to obtain short syllables, which are comparatively infrequent in Latin yet are badly needed for our dactyls. Another such device is the apostrophe, the direct address (even where not required by sense), which gives us the short syllable of the vocative. Lucan, for instance, uses the vocative in a simple enumeration of centaurs: te . . . frangentem, Monyche, saxa . . . teque . . . Rhoece ferox . . . et Phole, etc. However, this highly artificial subterfuge should be employed with great caution.

Quantity

All quantities should be checked, since even etymology can be deceptive: we have fîdus but (with short i) fides; pâr but pariter and parilis; statum but praestâtum; and so forth. Deceptive, too, is our habitual (mis)pronunciation of words like pius, amor and the prefix re-.

We have some leeway where quantity is optional (mute+liquid); we can also at a pinch employ synaeresis and diastole (abyetis; sil-u-a), but only where there is good precedent. Good precedent runs from Catullus and Vergil to Ausonius and Claudian but does not cover the archaic (Ennius, Lucretius), or the plain goshawful (Commodianus).

There are, however, some precedents we should not follow because quod licet Iovi non licet bovi. Under no circumstance may we lengthen a naturally short end vowel by two consonants beginning the following word - despite Vergil's lappaequê tribulique. Lengthening a short syllable before the caesura (desine plura, puêr, et quod nunc instat agamus) should be employed most sparingly.

There is authority for the converse procedure of retaining a short vowel before initial sp and st, so that we can defend phrases like, say, iam splendet plurima stella or tua munera sperno (tua munera sprevi is of course inadmissible). Totally verboten is the medieval usage of h as a position-making consonant.

Final -e must be closely watched (latê, but [short] bene, superne); some imperatives of the second conjugation may be used with short -e: cave, tace. Remember the imperative puta, but also the possibility of long vocative endings (Aeneâ , Anchisâ , and Anchisê besides Aenea and Anchisa).

Hiatus and Elision

The beginner will do well to avoid hiatus altogether, except in phrases like o utinam. Elision, too, should be used sparingly. Certain monosyllables (do, sto, spem, sim, etc.) are never elided, nor, as a rule, are iambic words. Interjections like vae, heu, a are never elided (elision before them is normal: Taygeta o qui me gelidis, etc.). Elision of iam, dum, nam can often be avoided by appending -que; for enim, etenim can be substituted. Despite respectable precedents, elision in the fifth foot should be avoided. Atque is often elided; hypermetric -que at end of line (if we want to employ this rather recherché device) must be elided by way of synapheia, i.e., the next line must begin with a vowel and there should be no break in the sense: et magnos membrorum artus magna ossa lacertosque | exuit.

Note that -que must not be appended to the active infinitive.

Fitting Words into the Line

We have seen that certain words will not fit into the dactylic hexameter. In addition to the above-mentioned internal cretics or three or more unchangeable shorts, we also have words beginning with three consonants and a short first syllable, followed by a long, like screator, strigosus (very few). When certain words or proper names cannot be used, synonym or paraphrase must be employed; the desperate device of tmesis (septemque triones) is better not attempted.

Some numerals are clumsy and harsh-sounding: quattuor, sedecim, duodecim fit the meter hardly, or not at all. Hence we employ the distributive: bis binos, or substitute tria lustra for quindecim annos. "A year had now passed" - iamque exactus erat bis senis mensibus annus.

For certain awkward word-forms we have useful contractions: noram for noveram, amasse for amavisse, tenuere for tenuerunt. Dehinc and some others may be scanned as one or two syllables; mihi, tibi, ubi as pyrrhics or iambi (so that crede mihi and mihi crede will fit equally well).

Now let us translate this phrase: "but when at the beginning of spring chill frost has fled." Sed simulatque, vere incipiente, frigidum gelu fugit. We attempt to recast this prose line in poetic form, first seeking for a line ending. Fugit pruina? - let's try fugere pruinae. Now we qualify the noun, not by the impossible cretic frigidae, but by the anapaestic gelidae - and behold! we have a hemistich: gelidae fugere pruinae. For the awkward vere incipiente we remember vere novo; now we have: (sed simulatque) vere novo gelidae fugere pruinae. Simulatque is too clumsy: ubi is shorter and can be scanned two ways. But now we are left with sed ubi vere novo, etc., and sed is short. So is at, and autem is unsuitable; but fortunately we remember the (exclusively poetic) ast. Now the line is complete: ast ubi vere novo gelidae fugere pruinae. We may, of course, play around with the line, recasting it with memories of diffugere nives, glaciale gelu, etc.

At an early stage the beginner should practice hyperbaton, separating nouns and adjectives with rhyming endings (as gelidae . . . pruinae above), since jingle must be avoided. He should also begin to shun too many self-contained lines, because they become monotonous. The word(s) carried over to begin the next line should be emphatic: by practicing this we learn the sweet uses of anaphora:
occidet et serpens, et fallax herba veneni
occidet; Assyrium volgo nascetur amomum.

Extension, Contraction, Padding

We try to render an English poem in the same number of lines; a sonnet will thus have 14 lines in Latin. Since Latin is very much more concise than English we may have to expand lines. This is best done by clothing nude nouns in appropriate (but not trite) adjectives. Here lies a pitfall. In describing an actual sea voyage I wanted to say: "the sea mew flies to and fro around the ship." For "ship" I picked a random synonym out of the two dozen or so available, puppis, rendered the fluttering of the bird by appropriate rhythm, and had

 (long, short, short, long)/circumvolitat/vaga gavia puppim.

Needing a choriambic qualifier for puppis, I thoughtlessly stuck in velivolam - until a critic asked me whether I had been on a windjammer. It was a steamer - consequently: fumivomam circumvolitat, vaga gavia puppim. Even then, puppis is not too happy, since only tankers emit smoke from their stern: but the jingle gavia navem had to be avoided.

The temptation to pad a line by putting in a word or syllable (such as en, heu, nunc) required by meter only, not by sense, should be resisted. As a rule, not more than one adjective should qualify a noun.

II. The Pentameter

In the elegiac distich, each two-line unit must be self-contained, i.e., the sentence may not carry on beyond the end of the pentameter. The pentameter is a highly restricted verse, since no substitution is permitted in the two dactyls of its second half. We likewise no longer have Catullus' freedom of elision before the caesura: sed tu cum Tappone/omnia monstra facis; cessarent tristique imbre madere genae. Elision in the second half of the pentameter should also be avoided, despite sentio et excrucior, and a line combining both these licences, like quam modo qui me unum atque/unicum amicum habuit, has been indefensible ever since Catullus.

Undesirable, too, is a monosyllable before the caesura: our made-up line vir bonus et probus in/condicione manet is bad, though two monosyllables are all right: quam mihi, non si se/Iuppiter ipse petat (Catullus).

In a series of distichs we shall do well not to couch the first half of every pentameter in nothing but dactyls, since this would become monotonous. Rhyme in the pentameter halves, if not overdone, is pleasant and permissible (in vento et rapidâ scribere oportet aquâ ).

Now to the question of ending the pentameter with a two-syllable word exclusively. This is the book rule; but here we venture to disagree, preferring Greek and pre-Ovidian practice. The deadly predictability of Ovid's pentameter ending makes for monotony. We shall at any rate try to avoid too many three-syllable words at the line endings, whereas choriambic words of four syllables are, in our view, thoroughly permissible. There is no objection either to ending the line with a monosyllabic form of sum (dictaque factaque sunt). We should avoid an end word terminating in a short vowel (regia bella pede).

Since the second half of the pentameter invariably requires four short syllables, we have to look for forms ending in short -a and -e (neuter plurals, vocatives, 3rd decl. ablatives), or words with initial vowel or h to enable us to employ short -us endings preceding them.

III. Conclusion

We sneer at the Alexandrian practice of writing a dozen versions of the same epigram (Ausonius loved this sort of thing), but it is excellent training for the incipient versifex. We cannot strive for poetic originality: if we can achieve a neatly turned phrase, some polished elegance, a few lines a Roman could have understood because they sound like Latin verse - then we have attained our aim, and upon our modest endeavors the Muse will have smiled.

 Non quivis poterit Pimpleum scandere montem:
  Praeceps a Musis plurimus eicitur.
 Ast "voluisse sat est" dicunt "cum magna petuntur:"
  Tu modicis coeptis, Musa, benigna fave!

New York University

BIBLIOGRAPHICAL NOTE

  1. Poetical Latin Dictionaries
  2. The works mentioned below are essentially identical in nature. As Schnur says, they are to be "found on secondhand shelves".

  3. Pedagogical Materials for Latin Poetry Composition
  4. A compilation of the works stated below can be ordered from Brad Walton.


    Footnotes

    1. I do not think that new Gradus have been published recently, but any number of them (whether of English or German origin does not matter, since they are usually all-Latin) are found on secondhand shelves.
    2. The student can do this himself. Helpful material is found in the usual handbooks on verse-writing, and in a very useful recent (n.d.) booklet, Rudimenta Poetica (Tirocinium Helveticum) , published in the series Editiones Helveticae by Orell Füssli, Zurich.
    3. All bad lines in this article are mine.
    4. Te libans, Lenaee, vocat (Verg.).
    5. However, many liberties taken by Catullus (see below under "The Pentameter") must be shunned by us.
    6. Likewise, medieval changes in quantity.
    7. Lucretius did it, but later convention banned it.
    8. But rhyme in the pentameter (see below) is good. Leonine hexameters, on the other hand, despite their occasional occurrence in the classic Poets (quot caelum stellas, tot habet tua Roma puellas) should not be sought. Besides, they are much too difficult to sustain.
    9. Like almost all our "rules," this is subject to common-sense exceptions.
    10. Think of the fine, rolling cadence of this distich in four words (reputed to be a joint effort of Oxford and Cambridge):
       Conturbabuntur Constantinopolitani
        Innumerabilibus sollicitudinibus!
    11. And, of course, other closed short syllables.
    12. Our two distichs at the end of this article were sired, of course, by Catullus'
       Mentula conatur Pimpleum scandere montem:
       Musae furcillis praecipitem eiciunt
      and the well-known tag: in magnis et voluisse sat est.