“The Craft So Long to Lerne”: Chaucer’s Invention of the Iambic Pentameter

In recent years two branches of mathematics, computation and statistical analysis, have helped settle some bitter and long-running disputes in the area of Chaucer’s metrics. First Gasparov developed a statistical technique, probability modelling, that compared accentual configurations in verse with those found in prose, and thereby established that the Italian endecasillabo should be classified as intermediate between syllabic and stress-syllabic. ¹ Since Chaucer borrowed so much from Boccaccio, this clearly has implications on the typology of the English poet’s long-line metre. Then the computer-based analysis of Barber and Barber established beyond doubt that Chaucer’s long-line is decasyllabic and that some word-final schwas count as syllables (however archaic that may have sounded at the end of the fourteenth century). ² My own contribution to this statistical work was to apply the principle of probability modelling to Chaucer’s long-line verse; this produced a conclusion on accentuation that was as clear as that of Barber and Barber on syllable count: Chaucer, unlike any previous poet before him, in any language, avoided placing prominent syllables in odd-numbered positions in the line (except the first). In other words, he invented the metre we call the iambic pentameter. ³

The iambic pentameter can be defined simply and economically in terms of the parameters of Hanson and Kiparsky. ⁴ Its template has ten positions and is right-strong; if we denote weak and strong positions by the binary digits 0 and 1, respectively, the iambic pentameter is: 0 1 0 1 0 1 0 1 0 1. Its correspondence rules are as follows: the maximum size of a position is one syllable, ⁵ although the final strong position in the line may also contain an extrametrical unprominent syllable; ⁶ no weak position in the line, other than the initial one, may contain a prominent syllable. ⁷ Prominence in this metre is defined as “having a lexically determined greater stress than either of its neighbours”; or, in the terms of Hanson and Kiparsky, “weak positions are constrained from containing the strong syllables of polysyllabic words.” ⁸

It should, of course, be no surprise that Chaucer decided to compose his mature works in decasyllables; the content of his verse makes it clear that Chaucer knew well both French and Italian decasyllabic poems. The remarkable fact is that Chaucer composed decasyllables such as no one previously had written (iambic pentameters), and it is important to trace the source of his innovation. The present article aims to do this by a qualitative analysis of decasyllable types, as defined by the interaction of metrical and grammatical structures within the line: caesurae (the traditional term) or the boundaries of cola (the term now used by linguisticians). ⁹ But, before conducting my analysis, I wish to recapitulate some important points in the history of both the decasyllable and Chaucer’s poetic apprenticeship (the “craft so long to lerne” of my title).

The earliest surviving decasyllables in a modern Romance language are found in the French Vie de Saint Alexis and the Occitan Boecis, and both date from the first quarter of the eleventh century. ¹⁰ The correspondence rules of this early vers de dix made it equivalent to two independent shorter lines: they prohibited the same word from supplying the syllables in positions 4 and 5, and stipulated that both positions 4 and 10 must contain a stressed syllable (and may also contain an extrametrical syllable). The shorthand used by French metrists to describe this line is 4M/F + 6M/F. ¹¹ Other variants of the vers de dix have survived from only slightly later that were 5M/F + 5M/F or 6M/F + 4M/F, but no French poem mixes the variants. It should be noted that an extrametrical syllable at the caesura (4F) made the actual number of syllables (to the last stress in the line) eleven, not ten. The transformation of the two-part (4 + 6) line into a unified line (of 10) was achieved by writers of lyrics to be set to music. Because having a regular number of syllables is more important to syllabic music than lexical stress, writers of French lyrics avoided lines with a word break after 4F (termed epic caesura), and instead wrenched the accentuation of words with feminine endings so that the first hemistich was 3F (lyric caesura). At the end of the twelfth century the primitive alexandrine (6M/F + 6M/F syllables) replaced the vers de dix as the metre of major epic and narrative poems in French ¹² and, when the decasyllable returned to favor in the fourteenth, it was the unified line of 10M/F syllables with a mandatory caesura after the fourth syllable (4M or 3F). This was the only type of decasyllable composed by Chaucer’s French contemporaries.

Provençal troubadours clearly perceived the Occitan vers de dix as a unified line of 10M/F syllables as early as the mid-twelfth century. While the overwhelming majority of their lines were 4 + 6, they occasionally included a line of 6 + 4; and, while they normally employed the lyric caesura (3F + 6M/F), they occasionally gave the first hemistich 4F syllables and reduced the second to 5M/F (see, for example, the verse of Peire Vidal). ¹³ The last practice is so alien to French poets that French metrists call this a coupe italienne. The earliest surviving Italian endecasillabi were composed in the middle of the thirteenth century at the court of the heretic emperor Frederick II (1197–1250) in Palermo and were influenced by Provençal poets seeking refuge there from the Albigensian crusades. ¹⁴ Although some early Italian poets composed poems with a fixed caesura, the most influential ones cultivated the variety allowed in the Provençal unified line of 10M/F. The caesura might fall after 4M, 4F, 6M, or 6F syllables; indeed, occasional lines have no caesura at all (that is, positions 4–8 may be occupied by a word of five syllables). This endecasillabo with no fixed caesura was employed by three poets whose work Chaucer knew well: Dante Alighieri (1265–1321), Francesco Petrarca (1304–74), and Giovanni Boccaccio (1317–75).

Chaucer was highly unusual for a fourteenth-century Englishman in speaking and reading Italian. As a boy his father worked in the London wine trade, which in the middle of the fourteenth century employed many Italians, and he had Italian friends throughout his life. We also know that he was sufficiently fluent in the language to make a very close translation of long sections of Boccaccio’s Filostrato, when he came to compose Troilus and Criseyde in the early 1380s. ¹⁵ His Italian is likely to have been revived and improved when he travelled to Italy, perhaps as early as 1368, certainly in 1372–73, but most significantly in 1378. From this last journey he brought back copies of Boccaccio’s Filostrato and Teseida, and Pearsall argues that Boccaccio provided the inspiration for the flowering of Chaucer’s poetic art in the latter’s middle age. ¹⁶ Only one of Chaucer’s long-line poems, the ABC, imitated from a French poem (dated 1331) by Guillaume de Deguileville, has been proposed to antedate Chaucer’s 1372 Italian visit. But Pearsall points out that this early dating depends entirely on Thomas Speght, who edited the 1602 edition of Chaucer’s works, and who wished to emphasize the connection between Chaucer and the House of Lancaster. ¹⁷ Pearsall, himself, proposes a date in the late 1370s for the ABC, and this is supported by my structural analysis (see below).

The other model that has been proposed for Chaucer’s longer line is French: the vers de dix, from which the Italian endecasillabo was derived. ¹⁸ The evidence in favor of this model is, however, almost all circumstantial, and not structural, as I shall demonstrate in this article. Chaucer was a courtier with royal patrons and the language of the royal court was French (it had been the official language of the country for two-and-half centuries after the Norman conquest). As long as the kings of England had realistic hopes of uniting France and England under their rule, it was a political necessity for courtiers to be bilingual, and Geoffrey Chaucer (Geoffroi Chausseur) was no exception to this rule. ¹⁹ The French-speaking English court was graced by two of the most famous French poets of the age: Jean Froissart (1337–1410) and Oton de Granson (?-1397), but the most brilliant French poet of the fourteenth century was Guillaume de Machaut (c.1300–1370), whose work was admired widely in Europe. We know that Chaucer was familiar with a great deal of French poetry (indeed, his early works are translations or imitations of French ones). We also know that young members of the English court, themselves, often composed French verse, and there is a manuscript in Pennsylvania that contains a number of such poems, including, perhaps, some composed by Chaucer. ²⁰ The evidence of Chaucer’s authorship is slight: fifteen poems have the letters “Ch” written against their titles, and no other candidate for these initials has been proposed. The “Ch” poems, with the exception of two lines, are structurally identical to French vers de dix of Machaut or Granson (see the analysis which follows). Chaucer thus had possible decasyllabic models in both Italian and French when he decided to forsake the octosyllable and compose iambic pentameters in the late 1370s. ²¹

Comparative Structural Analysis of French, Italian, and Chaucer’s Decasyllables

The analysis that follows presents eight types of decasyllable, as defined by the interaction between major syntactic boundaries (caesurae) and metrical positions. Each is illustrated by French, Italian, and Chaucerian examples, all of them earlier than the Canterbury Tales. The French examples are from the poems of “Ch,” the Vie de Saint Alexis (hereafter VSA), and the Cinkante Balades of John Gower (hereafter CB). ²² The Italian examples are from Boccaccio’s Filostrato (hereafter Fil.) and from the sonnets of Giacomo Lentini (fl. 1230–40). ²³ The English examples are from Chaucerian works composed before 1387: the Complaint unto Pity (hereafter CP), the ABC, Anelida and Arcite (hereafter AA), the Parliament of Fowles (hereafter PF), and Troilus and Criseyde (hereafter TC). ²⁴

In the (numbered) instances quoted below I have adopted a number of typographic conventions for scansional purposes: (1) lines are divided into hemistichs; (2) syllables in different positions are separated by hyphens; (3) stressed syllables are indicated by bold typeface; (4) in those parts of the line where prominence constraints operate, the syllables in strong positions are underlined; (5) in headless lines a void initial position is indicated by the symbol [V]; (6) whenever two hemistichs are fused by elision, apocope, or synaloepha, this is marked by the symbol ^ linking the two affected vowels.

A. Lines with French Masculine Caesura [4M + 6M/F]

Lines with masculine (in Italian, forte) caesura are the most common in vers de dix, because the French language has a large stock of oxytonic words and lexical monosyllables. Since the vers de dix had become a 4 + 6 metre by the fourteenth century, almost four in every five lines of the poems of “Ch” are Variant A; for example, the following lines:

1. 1. Car c’est de voir son corps gent et par-fait (“Ch,” XV. 27)

2. 2. Et se sa-voir vou-lez par au-cuns tours (“Ch,” XIII. 27)

3. 3. Et as-sez puet sa doul-ceur gra-ci-euse (“Ch,” II. 21)

Note that the only positions in which prominence is regulated are the last ones in each hemistich, 4 and 10. Other prominent syllables may occur anywhere: for example, positions 2, 6, and 7 in (1), 7 and 9 in (2), and 3 and 7 in (3). French verse, as C. S. Lewis pointed out, “has no rhythm in the English sense”, which is to say within the line or hemistich. ²⁵ The rhythm of French verse lies between lines. A regular succession of three undifferentiated syllables followed by a phrasal stress, then five undifferentiated syllables plus a phrasal stress. Cornulier argues that French syllabic meters are temporal because, in the syllable-timed delivery of that language, each syllable is perceived as occupying the same length of time. ²⁶ The phrasal stresses thus fall at regular time intervals. Such metres are clearly a good fit for modern French, where word stress has disappeared and only phrasal stress remains, but not so for Old French, which had word stress. ²⁷ Either medieval French poets and audiences had developed the modern Spanish and Italian double audition (hearing mid-line stresses but not hearing them), ²⁸ or else word stress was already perceptibly weaker than phrasal stress. ²⁹

This type of line is not uncommon in Italian, where it is called the endecasillabo a minore con forte cesura. Although Italian has relatively few oxytonic words, this type of line provides fifteen per cent of Boccaccio’s total. But, whereas Peire Vidal had included an occasional line with its caesura in a place other than after position 4, this variant is only one of four that Italian poets mixed with great freedom. ³⁰ The regularity of phrasal stress of the French metre was thus lost. In its place, because Italian is a language with strong word stress, poets were able to create regularity within the line and hemistich by the position of other accented syllables. From the first, Italian poets clearly preferred lines with accentual regularity, in particular those with no prominent syllables in odd-numbered positions, at least in the second half of the line, like instances (4) and (5) from the Filostrato: ³¹

1. 4. Tu, don-na, se’ la lu-ce chia-ra^e bella (Fil., I. 9)

2. 5. Ciò che di-rà ‘l mio ver-so la -gri-moso (Fil., I. 42)

These lines could be generated by Chaucer’s verse design: they have no strong syllables in weak positions, except the first of (5). But, although such duple time was the favorite Italian rhythm (the major rhythm), two other (minor) rhythms are also common in Boccaccio. One of them occurs in lines with a French masculine caesura, such as (6), below:

1. 6. Quan-do Cal-cás, la cui al- ta sci-enza (Fil., I. 57)

The regularity of this line arises from prominent syllables in positions 1, 4, 7, and 10, giving it a triple-time rhythm. ³² Such lines are very rare indeed in Chaucer, and it could be argued that Italian poets and audiences demand more variety than English ones. There are, however, other sources of variety that Italian poets did not employ; for example, the void position 1, or optionality of an extrametrical syllable at line ends. The first of these is unknown in endecasillabi and the second was abandoned by Italian poetic fashion very early in the history of the line. ³³ Virtually all Italian lines are piano (that is, with an extrametrical syllable in position 10) and the oxytonic words that they used at forte (masculine) caesurae were either avoided at line end, or resyllabified (for example, the word voi, a monosyllable mid-line, was pronounced as two syllables, vo-i, at the end of the line).

Variant A is Chaucer’s favorite structure, perhaps influenced by French verse he knew, but much more probably because English, like French, has a large stock of both lexical monosyllables and oxytonic disyllables to place before the caesura.

1. 7. My pur-pos was to Pi-te to com-pleyne (CP, 5)

2. 8. Help me that am the sorw-ful in-stru-ment (TC, I. 10)

3. 9. The lyf so short, the craft so long to lerne (PF. 1)

Chaucer’s line is fully stress-syllabic and there is only one accentual variant: the duple-time (iambic) one with no prominent syllables in weak positions. The two Italian minor rhythms that account for more than twenty-seven per cent of Boccaccio’s lines, between them, appear in barely three per cent of Chaucer’s. His decasyllable has become an iambic pentameter.

B. Lines with French Epic Caesura [4F + 6M/F]

By the fourteenth century the epic caesura had disappeared from French vers de dix, but in the long hagiographic poems and chansons de geste of earlier centuries it was almost the only alternative to Variant A; and a much needed one, because French has not only a large number of feminine nouns and adjectives, it also has verb forms that are paroxytonic, as in the followiing lines from the eleventh-century VSA:

1. 10. Puis con-ver-serent an-sam-ble lon-ga-ment (VSA, 21)

2. 11. I-loec a-rivet sai-ne-ment la na-cele (VSA, 82)

Once again, we can note how unregulated the position of word stress is in French lines, except when it coincides with phrase accent. In (10) unregulated prominent syllables occupy positions 1 and 6, and in (11) they occupy 2 and 7. Variety, and not regularity, rules within French hemistichs. ³⁴

It is perhaps not so surprising, given the mature Italian predilection for absolute isosyllabism (lines of 10F), that no epic caesurae are found in endecasillabi. But there are a number of lines in all Chaucer’s long-line works which have what he may have intended as epic caesura, and it is impossible to be sure. First of all there are lines in which the syllable in position 4 is followed by (unelidable) schwa and then a major syntactic break. We can not be certain whether such schwas should be syllabified (thus producing epic caesura) or not, just as we can not be sure whether line-final schwa should be pronounced. There is also another type of line that can be interpreted as having epic caesura; for example:

1. 12. Who-so thee loveth, he shal not lo-ve^in veyn (ABC, 71)

2. 13. And e-ver setteth De-sir myn hert on fire (CP, 101)

3. 14. As for to honour hir god-des ful de-voute (TC, I. 151)

It is possible, however, that the intended pronunciation of the word before the caesura in these lines is monosyllabic, with the final unstressed syllables slurred to something like “lov’th”, “sett’th”, and “hon’r”. There are good arguments for and against epic caesura being intended. For epic caesura are the facts that this device seems to fit well the rhythmic structure of the English language and that Lydgate almost certainly employs it. ³⁵ Against are the facts that epic caesura was already obsolete in French and that every other type of line employed by Chaucer can be found in Boccaccio (except for “headless” lines, see below). In my statistics in the table at the end of this article, I quantify all possible epic caesurae as Type B.

C. Lines with French Lyric Caesura [3F + 6M/F]

By the fourteenth century this had replaced epic caesura in French vers de dix as the principal method of incorporating a feminine word at the caesura. It is very common in Machaut, and other fourteenth-century French poets, including the “Ch” poet, from whom the following lines are taken.

1. 15. Tou-te bel-le de gra-ce droi-te plaine (“Ch,” I. 45)

2. 16. Et les heu-res fai-re leur com-mun cours (“Ch,” II. 3)

3. 17. A sim-ples- ce d’ex-cel-len-te va-lour (“Ch,” IV. 23)

Once again we can see that the French line has no one rhythm: other prominent syllables occur in positions 1, 6, and 8 in (15), thus making it a chance match for an iambic pentameter. But in (16) the other prominences are at 5 and 9, while in (17) there is only one in 7. I have marked the syllables at the caesura in these lines that would be prominent in normal speech, those in position 3. But in each case a schwa (in French, an e-muet) in position 4 must be wrenched into prominence in delivery so as to produce a line of 4 + 6. This phenomenon is called recession by English metrists.

Recession is found occasionally in English (but quite commonly in French) sung verse. It is also found in French poetry, probably because word stress is either weak (Old French) or non-existent (in the modern language). Because Italian, like English, has strong word stress, Italian poets rapidly rejected this type of variant. A few examples can be found in the earliest sonnets of the Sicilians, who were imitating Occitan models of this type of line; but by Dante’s time such ugly wrenching was avoided. The following Italian examples are therefore from Lentini’s sonnets:

1. 18. Se non quan- to ma-don-na mia vor-ría (IV. 7)

2. 19. E^a bon fi- ne de lo so re-gi-mento (XII. 6)

Probably because English word stress is even stronger than Italian, Chaucer seems to have eschewed the lyric caesura. All the examples that might be proposed have alternative explanations, and since the number of lines concerned is very few (compared with the many in vers de dix), it is probably correct to say that there are no lyric caesurae in Chaucer. Consider the following instances:

• *(20) I saw beu- te wi-thou-ten a-ny^a-tyr (PF, 225)

• *(21) On this la- dy , and now on that, lo-kynge (TC, I. 269)

Both “beute”/ “beauty” in (20) and “lady” in (21) had alternative (Frenchified) oxytonic pronunciations in Chaucer’s time, and these words should, I believe, be scanned: “beu-te” and “la-dy”, and not as above. This alternative scansion also has the merit of producing duple-time rhythm throughout both lines. All the possible lyric caesurae in Chaucer seem to involve words like “beauty” and “lady”, rather than accentually unambiguous words like “setteth”’ or “loved”. It is therefore reasonable to conclude that Chaucer shared the Italian distaste for the recessive lyric caesura.

D. Lines with Elided/Fused Caesura [4’ + 6M/F]

The unification of the vers de dix as a line of 10M/F gave fourteenth-century French poets another way of including a word ending in e muet at the caesura: this was to elide the -e before an initial vowel in the second hemistich. This produces a type of line that is relatively common in fourteenth-century vers de dix, including the poems of “Ch”.

1. 22. Pour nuit obs-cu-re^a son droit a-me-ner (“Ch,” III. 4)

2. 23. Se puet en gra-ce^a-mou-reu-se ve-oir (“Ch,” I. 10)

3. 24. Mo-rir m’est joi-e^et brief fi-ner doul-çour (“Ch,” II. 4)

Once again, note the difference between the positions of prominent syllables within second hemistichs; French poets were concerned to achieve variety of word stress, and regularity only of phrasal stress in positions 4 and 10. Variant D could be argued to be a more satisfactory way of incorporating paroxytonic words at the caesura, because, it involves violating no linguistic norms, unlike the recessive lyric caesura.

Poets composing in Italian deal with unwanted word-final vowels rather differently from French ones, although they use the same word, elisione, to describe the process (in order to confuse comparative metrists). It is called more correctly sinalefe (synaloepha) by Spanish metrists, who define this as a mixing of two vowels to form a sort of diphthong, as in “tutto^in”, below. ³⁶ I employ the same symbol (^) to indicate Italian synaloepha as I have used for French elision in these lines from Boccaccio, which have a caesura after 4’ syllables.

1. 25. Il mio cos-tu-me^an-ti-co^e u-si-tato (Fil., I. 7)

2. 26. Con-cor-di tut-to^in un pa- ri vo-lere (Fil., I. 54)

3. 27. Cas-tel-la^e vil-le^ar-den-do^e di -bru-ciando (Fil., I. 128)

It should be noted that instances (25) and (27) have Boccaccio’s major rhythm, and their first hemistichs are also in duple time. Individually, they are iambic pentameters, except that they are set among lines that are far from being so. Chaucer appreciated the accentual regularity of such lines and modelled his upon them. Instance (26) on the other hand is the minor variant with triple time (prominent syllables in positions 4, 7, and 10) which Chaucer rejected. In (27) we observe three synaloephae in a line; they glue the whole line together and defy any attempt to make a caesura into a pause in delivery. Multiple synaloepha is not uncommon in Boccaccio, but is employed even more frequently by Petrarch. ³⁷

English speech often prefers synaloepha to apocope/elision; we turn “the air” into one syllable by saying something like “thyair”; in contrast, “th’air” gives itself away immediately as artificial and poetic. Chaucer’s principal method of eliminating unwanted final vowels was the elision of word-final schwa (whenever it suited his metre). There are, however, still examples of synaloepha in his verse (see that of “many a” in n5, below). Chaucer employs the elided/fused caesura frequently.

1. 28. Thou art lar-ges-se^of pleyn fe-li-ci-tee (ABC, 13)

2. 29. With fyn-nes re-de^and ska-les syl-ver bryghte (PF, 189)

3. 30. And of-ter wol-de,^and it hadde been his wille (TC, I. 125)

It is reasonable to query whether it is likely that Chaucer syllabified word-final schwas and then elided most of them, when there is no sign that his contemporaries did so. But his contemporaries did not know Petrarch and Boccaccio, as he did, and so could not admire the way those two poets used final vowels, both to give their lines a duple-time rhythm, and to glue them together with synaloepha in one harmonious structure. As Windeatt points out, Chaucer frequently translates the lines with Boccaccio’s major rhythm not only word for word, but also stress for stress. ³⁸

E. Lines with Enjambed Caesura [4M > 6M/F]

In vers de dix, endecasillabo, and pentameter, alike, there may occur a word break after position 4 while the caesura in my definition (the most important syntactic boundary) occurs later in the line. This is called césure enjambante in French metrics, because it is the equivalent of enjambment at the line end. It clearly helps provide variety in a poem where all lines have a word break after position 4. In Italian the caesura is held to be after the main syntactic boundary in such lines: they are endecasillabi a maiore (meaning that their first hemistich is longer than their second) and may be fused by synaloepha. Examples of enjambed caesura in the three languages are:

1. 31. La est aus-si Sou-ve-nir qui ne cesse (“Ch,” I. 29)

2. 32. N’ains y per-çoy sa gra-ce, Dieu li mire (“Ch,” II. 18)

3. 33. Che t’a-ma più che sé d’a-mor per-fetto (Fil., I. 30)

4. 34. A-vea Cal-càs las-cia-to^in tan-to male (Fil., I. 81)

5. 35. And cer-tes yf ye wan-ten in these tweyne (CP, 76)

6. 36. ffor I, that god of lo-ves ser-vantz serve (TC, I. 15)

Note that in the layout of these lines I have shown the caesura as falling after the mandatory position 4 in the French examples, but after position 6 or 7 in the Italian and English ones.

F. Lines with Italian A Minore, Debole Caesura [4F + 5M/F]

This caesura is not permitted in French vers de dix, and in traditional French metrics it is termed coupe italienne. It is the most important result of Provençal/Italian unification of the line: the unprominent syllable that follows the final stress in the first hemistich is not extrametrical, but occupies position 5 and the syllable count continues (from 1–10, instead of from 1–4, followed by 1–6). Because the Italian lexicon contains a majority of paroxytonic words, this variant is very common in all Italian poets, and it is termed an endecasillabo a minore con debole cesura.

1. 37. Io de Par-na-so le Mu- se pre-gare (Fil., I. 3)

2. 38. E voi, a-man-ti, prie-go ch’as -col-tiate (Fil., I. 41)

3. 39. Ch’e-ra più bel-la ch’al-tra cre -a-tura (Fil., I. 99)

Note that (37) has a minor Italian rhythm (prominence in positions 1, 4, 7, and 10) and that the other two lines have the major rhythm (no strong syllables in weak positions in the second hemistich).

In all his pentameters, even the earliest, Chaucer uses Variant F very frequently.

1. 40. And fres-she Beau-te, Lust, and Jo-ly-te (CP, 39)

2. 41. Up by the bri-dil, at the sta-ves ende (AA, 184)

3. 42. That I were wur-thi my damp-na-ci-oun (ABC, 23)

Although no French poet employs this type of caesura in vers de dix, it was used extensively by Gower in CB. We have no way of knowing at what point in his life these poems were composed: Macaulay thinks late, but the content suggests early (they seem to be the conventionally immoral lyrics of a young man, certainly not the mature voice of the “moral Gower”). Their line structure is so similar to that of Chaucer’s iambic pentameter that it is probable that Gower was experimenting with this line in French at the same time as his friend Chaucer was doing so in English, in the late 1370s.

1. 43. A vous ma doul-ce da-me tre -shal-teine (CB, XXXIII. 22)

2. 44. Ma doul-ce da-me qui m’a-vetz ou-bli (CB, XXVIII. 22)

3. 45. Ma sa ba-ne-re quant mer-ci dis-plaire (CB, XXVII. 4)

These lines not only have an Italian caesura, they allow no prominent syllables in weak positions; they are endecasillabi/pentameters in Anglo-Norman. It is, however, possible that Gower composed CB over a number of years, or even towards the end of his life, because, although almost all his English verse is in octosyllables, his poem In Praise of Peace, ³⁹ dedicated to Henry IV on his accession in 1399, is in iambic pentameters; for example, lines 106–08:

1. 46. The wer-re is mo-dir of the wron-ges alle

2. 47. It sleth the prest in ho-li chir-che^at masse

3. 48. For-lith the maid and doth hire flour to falle

The first of these lines is one of many with an Italian caesura in Gower’s poem.

Even more surprising than Gower’s French endecasillabi/pentameters is a line in the poems of “Ch” with an Italian a minore, debole caesura.

1. 49. Se-cours con-for-te Dan-gier pres-te-ment (“Ch,” I. 16)

It might, of course, be argued that “Dang-i-er” is three syllables, but that would give this line an epic caesura, something not found in either the poems of “Ch”, nor in any fourteenth-century vers de dix by a French poet. Moreover, the word “Dan-gier” is always two syllables in the poems of “Ch” (for example, in IX. 13), as it normally is in French verse.

G. Lines With Italian A Maiore, Forte Caesura [6M + 4M/F]

This caesura is not permitted by the fourteenth-century rules of the vers de dix, but once again Gower employs it throughout CB.

1. 50. O gen-ti-le^En-gle-ter-re,^a toi j’es-crits (Traité, 25)

The prominent syllable in position 3 proves that Gower is here imitating an endecasillabo in French and not a Chaucerian pentameter. But some of Gower’s Balades do seem to have a constraint against prominent syllables appearing in weak positions that runs through all their lines; for example, XXXV. Another type of line with Italian caesura employed by Gower might be argued to be Variant F enjambed; because it has word breaks at two points mid-line; for example, this line with a word break at 4F, but a caesura at 6M.

1. 51. Sur-pris de vos-tre^a-mour et sus-pi-rant (CB, XXXVIII. 10)

This is further evidence that CB represents an experiment in composing endecasillabi in French.

The Italian endecasillabo a maiore con forte cesura is very common; as with its a minore equivalent, the first hemistich may end after 6M, or after 6’ followed by synaloepha.

1. 52. Gui-da la nos-tra man reg- gi lo’n-gegno (Fil., I. 31)

2. 53. Vo-len-do del fu-tu-ro^il ve-ro^u-dire (Fil., I. 60)

Instance (52) has the second minor rhythm commonly found in endecasillabi; it occurs when prominent syllables occur in positions 6, 7, and 10, and the line breaks from duple into triple time at the caesura. Because the second hemistich’s rhythm is called adonic in Classical metrics, this type of line is termed a fragmento adónico. ⁴⁰ Instance (53), on the other hand, has the major rhythm with no prominent syllables in weak positions in the second half of the line, and contains two synaloephae binding the words together.

Even in his earliest work in pentameters Chaucer uses Variant G caesura frequently; it is especially useful for accommodating Greek names early in the line.

1. 54. Of que-ne^A-ne-li-da and fals Ar-cite (AA, 11)

2. 55. For when Am-phi-o-rax and Ty-de-us (AA, 57)

3. 56. Up-on the cru-el-tee and ti-ran-nye (CP, 6)

H. Lines With Italian A Maiore, Debole Caesura [6F + 3M/F]

Once again, no French poet employs this caesura, and the only examples of it that I can find in French are in CB.

1. 57. Qe nul-le me-di-ci-ne m’est ver-raie (CB, XXVII. 4)

2. 58. A-mour est u-ne voi-e dan-ge-rouse (CB, XLVIII. 15)

3. 59. En re-sem-blan-ce d’ai-gle qui re-monte (CB, XLVI. 1)

These examples, taken in conjuction with those of Variant F, make it clear that Gower’s lines are not vers de dix but Italian endecasillabi imitated in the French language.

This type of line, called by Italian metrists the endecasillabo a maiore con debole cesura, is common in the work of all medieval Italian poets; the following examples are again from Boccaccio:

1. 60. Per cui nel te-ne-bro-so mon-do^ac-corto (Fil., I. 10)

2. 61. Per che se-gre-ta-men-te di par-tirsi (Fil., I. 65)

3. 62. E se-co^a ram-men-tar-si del pia-cere (Fil., I. 260)

These three lines have Boccaccio’s major rhythm. Lines with such a Variant H caesura appear in all Chaucer’s iambic pentameter works except CP, which is perhaps too short to contain one.

1. 63. Ne to no cre-a-tu-re made she chere (AA, 108)

2. 64. This no-ble em-pe-res-se, ful of grace (PF, 319)

3. 65. To herk-nen of Pal-la-dion the ser-uy ce (TC, I. 164)

Chaucer thus employs all three of the Italian caesurae that are never found in the work of French poets.

In addition to one Variant F line, the poems of “Ch” contain one other line with no word break after syllable 4. It has word breaks after 3M and 7M syllables.

1. 66. Et l’a-cort joi -eu-se-ment tra-ïc-tier (“Ch”, IX. 9)

This line does not conform to any Italian type; I suspect that it contains a scribal error: the word “pour” has been omitted before “l’accort”, and the dieresis in “traic-tier” is incorrect. If this is not the case, then this line is another piece of evidence that the author was beginning to chafe at the restrictions imposed by the mandatory word break in the vers de dix.

Note on “Headless” Lines

Headless lines in Chaucer are relatively rare. In my scansion I mark them by the symbol [V]. The fact that Chaucer’s most famous work opens with one is obviously of note.

1. 67. [V] Whan that A-prill with his shou-res soote

This is a Variant F Italian caesura line with a void in position 1, allowed by the closure rule. It is common in English verse for a metre to be indifferent as to whether or not there is an unprominent syllable before the first stress in the line. Thus we find in Browning’s The Lost Leader two types of line: one with an unprominent syllable before the first stress, and one without. ⁴¹ This is exemplified by the last two lines of the poem.

1. 68. Then let him re-ceive the new know-ledge and wait us,

2. 69. [V] Pard-oned in hea-ven, the first by the throne.

The second line clearly has the same (triple-time) rhythm as the first, but it is minus one syllable at its head. Browning’s pentameters do not allow this licence, probably because Milton and Pope had so much influence on the later pentameter, but some pentameters earlier than Milton seem to be missing their opening unprominent syllable, like the first line of the Canterbury Tales. An extra unprominent syllable at the start of the line is termed in Classical metrics anacrusic, a Greek term meaning “striking-up”. Music is similarly indifferent as to whether or not there are any anacrusic notes before the first beat in the first bar. When Chaucer decided to impose the yoke of syllable count on English long-line verse he seems to have allowed this traditional, musical, liberty of omitting the first unprominent syllable in the line.

The Relative Frequency with which Variants Occur

The table at the end of this article shows the frequency with which each type of line occurs in Chaucer’s work up to and including TC. ⁴² Remarkably none of the differences between the individual poems are significant, with one small exception. The proportion of possibly epic caesurae (Type B) in the Complaint unto Pity is double that of the other poems. This, I believe, is probably because it was his first composition in the metre to survive, and Chaucer had not yet perfected the technique whereby the majority of word-final schwas at the caesura are elided before a following initial vowel.

1. 70. But er I myghte with a-ny word out-breke (CP, 12)

2. 71. The world is lore; ther is no more to seyne (CP, 77)

The implications of the other figures in the table are made clear in the conclusions that follow.

Conclusions: The Development of Chaucer’s Iambic Pentameter

1. Chaucer’s model was clearly Boccaccio’s endecasillabo and not the French vers de dix. He employs all Boccaccio’s variants, while his only possible debt to France is the occasional epic caesura (and even this is doubtful). Chaucer also borrowed the Italians’ favorite rhythm, because French lines have no rhythm within the line, and because a duple-time rhythm (one of moraic trochees) happens to underlie both the Italian and the English languages.

2. In order to produce the iambic pentameter Chaucer had to restrict in his verse the two minor rhythms of Boccaccio’s (the first in triple time, the second the fragmento adónico). Such rhythms constitute barely three per cent of Chaucer’s lines. These rhythms are the chief source of variety in Italian endecasillabi, but in English, with its long accentual tradition, it was not acceptable to substitute a four-beat line for a five-beat one. To compensate for this loss, Chaucer introduced other, more traditionally English, sources of variety: the void position, the extra unprominent syllable within the hemistich, and possibly the epic caesura (with its extra unprominent syllable).

3. All of the poems by Chaucer in pentameters that have survived have the same mastery of the line and its many variants. None is likely to be significantly earlier than the others on metrical grounds, except the Complaint to Pity, which may show evidence that Chaucer’s technique still lacked one refinement: eliding most unwanted word-final schwas at the caesura. But, however long Chaucer may have studied the craft of composition, his first surviving pentameters are almost identical to his last. The iambic pentameter sprang forth in panoply; it was not a step-by step process or the result of fumbling trial and error, at least as far as the surviving poems are concerned.

4. Chaucer may be the author of the “Ch” poems: there are no other obvious candidates for the initials “Ch”, and these poems contain one line that is an endecasillabo in French, and another that may show impatience with the mandatory caesura of the vers de dix.

Epilogue: The Iambic Pentameter After Chaucer

John Gower (?1330–1408) may have been close enough to Chaucer (both in time and in their common interest in Italian versification) to compose a few iambic pentameters making use of word-final schwa, but the new Italianate verse did not immediately take hold. Because Thomas Hoccleve (?1369–1426) had no Italian, he imitated Chaucer in composing decasyllables, but he tended to let the rhythms within his lines look after themselves, as the French did. ⁴³ And because John Lydgate (?1370–1449) had no Italian, he thought that the possibility of iambic verse had died with the pronunciation of word-final schwa, and so he returned to the older English tradition of verse based on beats, while making his lines as Chaucerian (that is, iambic) as he could. ⁴⁴ Robert Henryson (?1424-?1506), who spoke a Scots dialect in which final schwa had not been pronounced for a century, clearly understood how Chaucer’s metre, with its use of convenient schwas, worked; and he composed the first iambic pentameters in schwa-deleted English. ⁴⁵ In the sixteenth century, Sir Thomas Wyatt (1503–42) made one last attempt to develop an English line that was not based on the foreign practice of syllable count: a balanced line with a central caesura and either two or three beats in each half. ⁴⁶ Eventually Henry Howard, Earl of Surrey (?1517–1547) and, a generation later, Sir Philip Sidney (1554–86) turned again to Italy to reinvent the iambic pentameter; and this is the metre that then dominated English poetry for more than four hundred years. ⁴⁷ But none of these poets can be compared for sheer inventiveness or metrical artistry with the poet who first “lerned the craft” of the iambic pentameter, Geoffrey Chaucer.

Table.

Percentages of Line Types Work/Sample

LINE TYPE	“Ch”	Fil.	CP	ABC	AA	PF	TC	Chaucer (pre-1387)
A (4M + 6)	78	15	46	46	39	41	44	42
B (4F + 6)	0	0	10	5	5	4	4	5
C (3F + 6)	10	0	0	0	0	0	0	0
D (4’ + 6)	5	23	7	9	11	9	11	10
E (4M > 6)	6	19	19	18	15	21	17	18
F (4F + 5)	0	29	11	14	19	13	15	15
G (6M + 4)	0	9	5	6	5	6	6	6
H (6F + 3)	0	3	0	2	4	5	3	3

Differences between the sums of the above percentages and 100 are the result of rounding and a very small number of irregular and defective lines.