Free Verse and the Constraints of Metre in English Poetry

FACULTY OF HUMANITIES UNIVERSITY OF COPENHAGEN

PhD thesis Jesper Kruse

Free Verse and the Constraints of Metre in English Poetry

Name of department:

Department of English, Germanic and Romance Studies

Author:

Jesper Kruse

Title:

Free Verse and the Constraints of Metre in English Poetry

Supervisor:

Professor Charles Lock

Submitted:

31 January 2012

Acknowledgements The faults of this study would have been greater than they are and its merits smaller without the kindness and help from a large number of friends and colleagues. I wish to thank: Jonas Holm Aagaard, Dorte Albrechtsen, Susan Ang, Søren Staal Balslev, Martyn Bone, Clare Brant, Dorrit Einersen, Anastasia Gremm, Greg Hewett, Adam Hyllested, Annemarie Jensen, Line B. Kristensen, Josephine Lehaff, Arianna Maiorani, Winfried Menninghaus, Andrew Miller, Jens Erik Mogensen, Jimmi Nielsen, Toke Nordbo, Dennis Omø, Rajeev S. Patke, Siff Pors, Robert Rix, Ruben Schachtenhaufen, Karsten Schou, Steen Schousboe, Inge Birgitte Siegumfeldt, Annelise Siversen, Jørgen Staun, Anna Wegener and Elzbieta Jolanta Wójcik-Leese. I am particularly indebted to Henrik Gottlieb, Jessica Ortner and Bill Overton for their generosity and hospitality. Furthermore, for their enduring patience and encouragement, I am grateful to my parents as well as to my friends, Christel and Lars. Most of all, however, I am grateful to Cindy for, well, everything: this dissertation is dedicated to her. Finally, I wish to thank my supervisor, Professor Charles Lock, who also supervised my BA on exclamation marks in English verse and my MA on the relationship between phonotactics and metre. Both of these projects owed their conception to an informal reading group headed by Charles Lock, and of which I was fortunate enough to be a member. This reading group provided the setting for my first serious introduction to the intricacies and joys of poetry: how does one acknowledge such a debt? In his capacity as supervisor on this dissertation, Charles Lock has been an inexhaustible well of fiercely erudite inspiration, as well as a good friend: walking in his company to Steep Church on a splendid June day in 2010 (in commemoration of Edward Thomas) easily stands out as the single most memorable experience during the process of which this dissertation is the result.

2

Contents Introductory Remarks

3

Chapter 1. Beginnings

10

Chapter 2. The Thawing of an Icicle: Metre and Theory

32

Chapter 3. Unseen Roots in Tongue-Tied Springs: The English Pentameter

64

Chapter 4. The Chinese Wall of Milton: Towards Blank Verse

95

Chapter 5. On Golden Hinges Turning: Enjambment in Blank Verse

131

Chapter 6. To that Sweet Yoke Where Lasting Freedoms Be: Free Verse in English

167

Chapter 7. To Enumerate the Muses: The Territory of Neither-Prose-Nor-Verse

203

Concluding Remarks

232

Appendices

234

Bibliography

247

3

Introductory Remarks This dissertation was supposed to be about free verse, and in a sense it still is. Accordingly, it assumes (against the explicit admonition of T.S. Eliot) that free verse exists and that it evolved in response to a mode of versification from which it is itself distinguishable, namely metrical versification. In the context of English literature, the metrical tradition that free verse reacted against goes back to the late fourteenth century when Geoffrey Chaucer abandoned alliterative verse and introduced the iambic pentameter. Over the following centuries, this particular form of accentual-syllabic verse became established as the norm for serious English poetry to such an extent that, by the early twentieth century, the iambic pentameter had come to be viewed as an intolerable constraint by a large number of poets. As a result, English poetry in the early decades of the twentieth century underwent a remarkable formal transition, the most conspicuous result of which was the introduction of verse that was self-consciously free of metrical structure. Therefore, this dissertation is also about metrical structure and about the ways in which metre affects language. The advent of free verse coincides with the advent of modernism, and there has been a strong tendency to use the willingness with which poets of the early twentieth century broke away from traditional metrics as a criterion for inclusion in the modernist canon. Ezra Pound, T.S. Eliot and William Carlos Williams, all of whom employed free verse for their most important work, are by common consent regarded as central figures of modernism, whereas the position of poets who did not abandon traditional metre, say Robert Frost or Edward Thomas, is more ambiguous. The most notable exception to this general tendency is W.B. Yeats, whose mature work is usually considered part of the modernist canon, in spite of Yeats’s staunch retention of traditional metrics in his own work. However, when Yeats in 1936 undertook the editorship of The Oxford Book of Modern Verse, he included as its opening text a lengthy sentence from Walter Pater’s essay on Leonardo da Vinci, arranged as a poem in free verse: She is older than the rocks among which she sits; Like the Vampire, She has been dead many times, And learned the secrets of the grave; And has been a diver in deep seas, And keeps their fallen day about her;

4

And trafficked for strange webs with Eastern merchants; And, as Leda, Was the mother of Helen of Troy, And, as St Anne, Was the mother of Mary; And all this has been to her but as the sound of lyres and flutes, And lives Only in the delicacy With which it has moulded the changing lineaments, And tinged the eyelids and the hands.1 This rather remarkable editorial choice on the part of Yeats is a testimony of (and suggests Yeats’s position on) the most immediate critical issue that the emergence of free verse raised: is free verse really verse, or is it merely prose chopped into lines? This question has not yet been satisfactorily answered, in spite of the fact that free verse went on to become the dominant compositional mode of English poetry in the course of the twentieth century. Some early commentators on free verse – e.g. Ford Madox Ford, F.S. Flint and Llewellyn Jones – were content to accept free verse as occupying a borderland between prose and verse. However, with the rise of structuralism, the question of free verse was brought to the fore again. Benjamin Hrushovski’s paper ‘On Free Rhythms in Modern Poetry’ marks in many respects the beginning of modern scholarship devoted to the problems of free verse. Hrushovski begins by declaring: There are two possible ways of facing the fact of the existence of free verse: one is to exclude free rhythms from poetry ... the alternative, if we cannot afford simply to dismiss important parts of modern poetry ... is to revise thoroughly our old notions of poetic rhythm ... and then to come back to a structural and meaningful description of free-rhythmic phenomena.2

1

W.B. Yeats (ed.), The Oxford Book of Modern Verse (Oxford: At the Clarendon Press, 1936): 1. Benjamin Hrushovski, ‘On Free Rhythms in Modern Poetry’, in Thomas A. Sebeok (ed.), Style in Language (Cambridge, MA: M.I.T. Press, 1960), 173-190: 173. 2

5

Hrushovski’s own solution to the vexed problem of herding together both free and metrical verse under the same poetic umbrella is both simple and radical. Having discarded the notion that free verse represents a ‘border area between prose and poetry’ as a ‘misunderstanding ... going back to the Greeks’,3 Hrushovski posits that numerical metricality is both insufficient and unnecessary as a criterion for defining poetry and that the only meaningful differentia of poetry therefore is the verse line. To bestow upon the line the honour of being the defining constituent of verse has become so commonplace in modern poetics that one can easily forget just how far removed from earlier conceptions of verse such a definition is. However, the idea that lineation rather than metre is what distinguishes verse from prose did not originate with Hrushovski. One could indeed argue that it is implicit already in the whimsical distinction between prose and poetry that Jeremy Bentham draws in a letter to Lord Holland as early as 1808: ‘But, sir, oh, yes, my Lord I know the difference. Prose is where all the lines but the last go on to the margin—poetry is where some of them fall short of it.’4 However, the roots of the constitutive status accorded to lineation in modern poetics – and the roots of Hrushovski’s use of the concept – should not be sought after in nineteenth-century letter exchanges, but rather in the groundbreaking work performed by the Russian Formalists at the beginning of the twentieth century. Boris Tomashevsky’s contention, dating back to 1925, that the ‘breaking up of poetic language into lines, into sound units of similar and possibly equal force, is clearly the distinctive feature of poetic language’5 is generally accepted, with very few modifications, by most metrists today.6 Hrushovski’s challenge was taken up by Donald Wesling in his 1971 article ‘The Prosodies of Free Verse’.7 Drawing on Michael Halliday’s grammatical theories, Wesling proposes a rank scale of interconnected poetic elements consisting of poem, stanza, line, word and syllable, each of which influences our conception of the others. Of these elements Wesling singles out the line as the most crucial, but at the same time also most volatile, element of poetry in that it is ‘equally liable to 3

Hrushovski (1960): 185. Cited in Stephen Adams, Poetic Designs: An Introduction to Meters, Verse Forms and Figures of Speech (Ontario: Broadview Press, 1997): 152, n. 5. 5 Cited in Anthony Easthope, Poetry as Discourse (London: Methuen, 1983): 51. 6 Cp. recent statements such as ‘Poetry is the sound of language organized in lines’ (James Longenbach, The Art of the Poetic Line (St. Paul, MN: Graywolf, 2007): vii) and ‘What distinguishes all poetry from prose is that poetry is made up of lines ... It is the organization of the text into lines that defines poetry in all languages and literary traditions’ (Nigel Fabb & Morris Halle, Meter in Poetry: A New Theory (Cambridge: Cambridge University Press, 2008): 1). 7 Donald Wesling, ‘The Prosodies of Free Verse’, in Reuben A. Brower (ed.), Twentieth-Century Literature in Retrospect (Cambridge, MA: Harvard University Press, 1971): 155-187; reprinted in a revised edition in Donald Wesling, The New Poetries: Poetic form since Coleridge and Wordsworth (Lewisburg, PA: Bucknell University Press, 1985: 145-171). 4

6

be taken downwards to syntax and accent or upwards in the direction of the whole poem’. Wesling then applies to the area of free verse P.J. Wexler’s notion of ‘the grammetrical, a hybridization of grammar and metrics “whose key hypothesis is that the interplay of sentence-structure and linestructure can be accounted for more economically by simultaneous than by successive analysis”’. This leads to a lengthy discussion of enjambment, the distribution and effect of which are at the heart of Wesling’s contention that ‘prosody is meaning’.8 The modest title of Charles O. Hartman’s Free Verse: An Essay on Prosody (1980) downplays the fact that it is the first monograph published in English to address exclusively the prosodic problems created by the emergence of free verse. Disavowing any influence from Hrushovski,9 Hartman begins his study with a set of sensible definitions, the most basic of which is a familiar one: ‘Verse is language in lines’. Much more daring is his definition of metre: ‘A meter is a prosody whose mode of organization is numerical; consequently ‘the prosody of free verse is rhythmic organization by other than numerical modes’.10 This is an extremely useful distinction in that it makes clear what is usually only hinted at vaguely: for poetic rhythm to be metrical, it must be structured in accordance with some objective parameter, the basis of which can only be numerical. Hartman – like Hrushovski – also argues against the notion that the difference between verse and prose is one of degree: ‘There is a dichotomy—not a spectrum—between verse and prose. Lineation distinguishes them’.11 According to Hartman, the effect of lineation, regardless of whether it is determined by numerical parameters or free, is to promote and control the attention of the reader. This effect is explained as the result of a set of shared conventions between poet and reader, among which Hartman counts the habit of pausing at line endings. Without any reference to Wesling, Hartman posits for free verse a ‘minimal prosody’ by which rhythm is controlled by ‘the counterpoint of lineation and syntax alone’.12 As ought to be evident there are significant theoretical similarities between Hrushovski (1960), Wesling (1971) and Hartman (1980). The affinity between the three texts is openly acknowledged in a revised version of Wesling’s essay and finds an obvious successor in Enikö Bollobás’s 1986 study Tradition and Innovation in American Free Verse. All of these studies are poised between phenomenology and structuralism and tend to seek, in the words of Donald

8

Wesling (1971): 162-164. Charles O. Hartman, Free Verse: An Essay on Prosody (Princeton, NJ: Princeton University Press, 1980): 184, n. 6.3. 10 Hartman (1980): 11-25. 11 Hartman (1980): 52. 12 Hartman (1980): 72. 9

7

Wesling, to ‘understand [poetry], as Modern linguists understand language, as a system of relations rather than an addition of particles’.13 Whereas in the first three decades after 1960 scholarship devoted to free verse was mostly concerned with establishing its theoretical foundations, the period after 1990 saw an increased interest in historicizing and contextualizing free verse. This tendency underlies three important monographs devoted to our subject: Timothy Steele’s Missing Measures: Modern Poetry and the Revolt against Meter (1990), H.T. Kirby-Smith’s The Origins of Free Verse (1996) and Chris Beyers’s A History of Free Verse (2001). Of these Kirby-Smith may be said to represent the most purely historical approach to his subject. Positing a contrapuntal relationship between free verse and metrical verse, Kirby-Smith notes that ‘Free verse is in itself epiphenomenal to older meters [for which reason] the relationship to its origins cannot be ignored.’14 Consequently, Kirby-Smith is audacious enough to look for progenitors of free verse in centuries not usually explored for such purposes and cites remarkably ‘free’ poems by Barnabe Barnes, William Drummond and John Milton in support of his case. Kirby-Smith is well aware that ‘to argue in this fashion is to invite denunciations from formalists and organicists alike’,15 and these two labels may well be applied to Steele and Beyers respectively, although not without qualification. Whereas Kirby-Smith’s aim is to establish plausible historical lineages between metrical verse and free verse, Steele takes a more formal approach and views modernist free verse as an unprecedented break away from a stable metrical tradition. Steele’s work ‘is not centrally concerned with analyzing different species of free verse and with speculating about the ways in which these do or do not work’;16 rather it focuses on explaining how and why free verse became the dominant form of poetry at the beginning of the twentieth century. Chief among the reasons for this shift, Steele lists the identification of metre with dated diction that is implicit in Eliot’s second Milton lecture and in his essay on ‘The Music of Poetry’. This identification sets apart the modernist revolution of poetry from earlier poetic revolutions, such as those led by Dryden and Wordsworth:

13

Wesling (1985): 153. For a work which differs radically from the line-oriented approach to free verse, see Winifred Crombie’s Free Verse and Prose Style: An Operational Definition and Description (London: Croon Helm, 1987). Crombie diverges fundamentally from the previously reviewed works by stating that the ‘only defining characteristic of verse is that it has metre’ and that ‘organization into lines is frequently, though by no means always, little more than a typographical convention’ (11-12). See also Richard D. Cureton, Rhythmic Phrasing in English Verse (London & New York: Longman, 1992), 277-323. 14 H.T. Kirby-Smith, The Origins of Free Verse (Ann Arbor, MI: University of Michigan Press, 1996): 37. 15 Kirby-Smith (1996): 37. 16 Timothy Steele, Missing Measures: Modern Poetry and the Revolt against Meter (Fayetteville, AR & London: University of Arkansas Press, 1990): 26.

8

these poets also objected to the mannerisms of a stilted idiom, but found no reason to reject metre; the modernists, on the other hand, threw out the metrical baby with the idiomatic bath water. Beyers’s work may be said to represent a more organicist approach to free verse, as long as by ‘organicist’ we understand the Coleridgian appreciation of ‘form as proceeding’ over ‘shape as superinduced’. Already in his Introduction, Beyers criticizes Kirby-Smith’s unequivocal rejection of organicism as a valid framework for assessing free verse, and states that the purpose of his own study is to examine ‘the enabling myths, including organicism, that Modern writers used without judging these ideas according to whether or not they are true.’17 Beyers insists on the necessity of distinguishing between different genres of free verse: some genres do enter into a fruitful relationship with older metrical forms, whereas other genres are characterized by consistently and self-consciously subverting metricality. Ultimately, however, even as far as the latter genres of free verse go, ‘asserting that a poem avoids prosody brings up the question of whether any poem can truly do so’.18 Form, in other words, establishes itself, regardless of the poet’s aspiration to escape it. This study shares with the work of Steele, Kirby-Smith and Beyers the basic conviction that the phenomenon of free verse is best studied in a historical and contextualizing manner. KirbySmith’s contention that ‘the effects of free verse depend on an implied contrast, or difference, with what it is not, or rather, on what it came from and what it no longer is’19 also underlies the approach taken in this dissertation. But whereas Kirby-Smith is mainly concerned with finding precursors to free verse on the fringes of the canon of English poetry, my focus is on the kind of metrical verse which the modernist proponents of free verse saw their own work as a reaction against. I thus share with Steele the view that modernist free verse represents an unprecendented revolt against a stable metrical tradition, and it is with this tradition that the present study is primarily concerned. This view is also reflected in the structure of the dissertation: it is divided into seven chapters that fall in three major sections. The first section is concerned with metre in general: Chapter 1 outlines the various metrical systems of European versification, and Chapter 2 is dedicated to the ways in which these metrical systems have been approached and appropriated by theorists throughout the history of poetics. The second section is devoted to metre within the context of English versification, with special emphasis on the iambic pentameter. Chapter 3 recounts the decidedly mixed heritage of English metrics and its relationship to Romance versification; Chapter 4 discusses the invention of 17

Chris Beyers, A History of Free Verse (Fayetteville, AR: University of Arkansas Press, 2001): 3. Beyers (2001): 223. 19 Kirby-Smith (1996): 37. 18

9

blank verse; and in Chapter 5 the properties of epic blank verse are pitted against the properties of dramatic blank verse. The third section examines the emergence of modernist free verse: Chapter 6 investigates the role that translations, imitations and adaptations of foreign verse designs have played as pretexts for experiments with free verse, whereas Chapter 7 engages with the relationship between prose and verse. The dissertation concludes with a quantitative study of certain phonological characteristics of metrical verse, free verse and novelistic prose: this part of the dissertation finds, at least methodologically, its most obvious predecessor in G. Burns Cooper’s Mysterious Music: Rhythm and Free Verse (1998). In addition to the literature covered here and in the bibliography, my work has been aided greatly by the resources offered by The Modernist Journals Project. A joint venture between Brown University and the University of Tulsa, The Modernist Journals Project provides free access to impeccably scanned versions of all the most important literary magazines in which so much modernist poetry first emerged.

10

Chapter 1 Beginnings (But our beginnings never know our ends!) T.S. Eliot, ‘Portrait of a Lady’

Western literature began, not with a bang, but with the sententious whimper of a young man, sulking over the loss of an even younger girl whom he had claimed as a spoil of war. The wrath of Achilles, brought on by Agamemnon’s illicit seizure of Briseïs, is what sets in motion the events recounted in the 24 books which make up Homer’s Iliad. As pivotal as these events have been to the imagination of subsequent generations of writers and artists, it must – at least in the context of this dissertation – be conceded that the true merit of Homer’s epic rests with the fact that it is also our earliest major work of metrical language. Because of its inherent redundancy, the collocation metrical verse was until quite recently likely to have been frowned upon by conscientious language users. Advocates of linguistic economy would have been justified in claiming that the phrase metrical verse is about as informative as the phrase white milk, which Aristotle draws on to exemplify the sort of unseasonable pleonasm which should be allowed to occur only in poetry, and not in prose.1 After all, the Oxford English Dictionary defines the word verse, when used without an article, as ‘Metrical composition, form, or structure; language or literary work written or spoken in metre; poetry, esp. with reference to metrical form’. Tellingly, the OED gives as its most recent example of this usage a quotation from 1883: after this date, the particular sense of verse as designating metrical compositions exclusively became destabilized by the rapid emergence of a new type of verse that was, if not entirely, then at least partially free from metrical concerns. 1886 thus marks the first recorded instance of the collocation free verse in the OED, to be found in an article by the American philologist Francis B. Gummere, entitled ‘The Translation of Beowulf, and the Relations of Ancient and Modern English Verse’. In that article the term free verse is, somewhat surprisingly, applied to certain parts of Edmund Spenser’s Shepheardes Calender. These parts – among them the 1

Aristotle, Rhetoric III, iii.

11

eclogue for September2 – hardly qualify as specimens of what we have now come to associate with free verse; if anything, they exemplify in a fairly moderate manner the type of ‘tumbling’ metrical freedom which is alluded to in one of T.S. Eliot’s famous debunkings of free verse: ‘freedom is only truly freedom when it appears against the background of an artificial limitation’.3 The artificial limitation against which the alleged metrical freedom of Spenser’s ‘September’ is set is clearly that decasyllabic line-form which in English metrical theory is most commonly referred to as iambic pentameter. It is this norm which Spenser constantly allows himself to withdraw from and approximate in the passages that Gummere describes as free verse: Eliot, we must assume, would have approved. But there is another, more restricted, sense of the term free verse in English, and it is with this other sense that the present dissertation is mostly concerned. The year after Gummere’s article was published, the Parisian poet and magazine editor Gustave Kahn put out his Les Palais nomades, the first collection of French poetry to contain a significant number of poems written in vers libre.4 Vers libre is, of course, the French term on which the English term free verse was calqued at the turn of the twentieth century, and it is ostensibly this calque which is being cited in the OED’s second quotation of the term free verse. The citation in question has been culled from F.M. Warren’s review of Philipp August Becker’s doctoral dissertation Zur Geschichte der Vers libres in der neu-französischen Poesie dating from 1888. Warren’s review was published in the February edition of Modern Language Notes in 1890, and the excerpt singled out for quotation by the editors of the OED reads: ‘The author examines the origin and development of free verse in modern French poetry’.5 A closer reading of Becker’s dissertation reveals, however, that Becker’s treatment of his subject had not avoided the notion of metre, as is evident from the wording of his opening paragraph: ‘Vers libres sind ein astrophisches metrisches Gebilde, welches aus beliebig geordneten Versen ungleicher Silbenzahl mit freier Reimstellung besteht’.6 Exactly how a ‘construction of randomly ordered lines of unequal syllable count’ qualifies as ‘metrical’ is not explained, but the fact that Becker retains metricality in his definition of vers libre is a symptom of the reluctance with which the notion of metre was abandoned by scholars in poetics.

2

‘The third and last eclogue in free verse is September’ (Francis B. Gummere, ‘The Translation of Beowulf, and the Relations of Ancient and Modern English Verse’, in The American Journal of Philology, Vol. 7, No. 1 (1886): 68). 3 ‘Reflections on Vers Libre’, in T.S. Eliot, To Criticize the Critic (London: Faber and Faber, 1969), 183-189: 187. 4 Anne Holmes, Jules Laforgue and Poetic Innovation (Oxford: Clarendon Press, 1993): 99, n. 19. 5 F.M. Warren, review of Ph. Aug. Becker’s Zur Geschichte der Vers libres in der neu-französischen Poesie, in Modern Language Notes, No. 2 (February, 1890): 58. 6 Ph. Aug. Becker, Zur Geschichte der Vers libres in der neu-französischen Poesie (Halle: Erhard Karras: 1888): 1.

12

Although it lingered in critical writings on the excellences of verse, metre was in fact abandoned as the sine qua non of verse by a very large number of poets in the decades that followed Becker’s dissertation. Claims to the contrary are either polemical or disingenuous because they purposely ignore what even a conservative metricist like George Saintsbury was willing to admit, in the preface to the second edition of his History of English Prosody (1923): ‘The most prominent feature of poetic practice today is, of course, the preference shown to “free” verse’. 7 The scare quotes surrounding free are significant, of course, in that they reveal a certain reluctance on the part of Saintsbury to embrace fully what is actually signified by this ‘free verse’ chimera. In this Saintsbury was not alone. In fact, the very existence of free verse was denounced on a regular basis by some of that form’s most celebrated practitioners in English – among them T.S. Eliot, Ezra Pound and William Carlos Williams8 – but their objections to free verse were first and foremost aimed at the apparent absurdity of the term itself, which they considered to be a misnomer and a contradiction in terms. Dissatisfaction with free verse as a critical term is still voiced today on the same grounds,9 for which reason its unabashed usage in this dissertation requires qualification. The term free verse is used here as a byword for verse which is not metrical, partly to ensure continuity with the French term from which it originated, and partly in acknowledgment of the fact that free verse, regardless of the objections that are habitually raised against it, has gained currency as the most commonly used critical term in English for non-metrical verse. What free verse is free from – as far as this dissertation is concerned – is thus metrical organisation of line structure, and consequently we should first examine what exactly is entailed by calling certain types of verse metrical.10 We began this chapter by invoking the wrath of Achilles on the grounds that it is with the repercussions of this wrath that our earliest major work of metrical verse is concerned. It therefore seems reasonable to turn to the Iliad once again as the most natural starting point for a discussion of metricality in verse. In preliminary terms, it may be said that what characterizes a work of metrical 7

George Saintsbury, A History of English Prosody from the Twelfth Century to the Present Day, Second edition (London: Macmillan, 1923), Vol. I: vi. The year before the second edition of Saintsbury’s History, Thomas Hardy had asked ‘if one may quote Tennyson in this century of free verse’ in his Late Lyrics and Earlier (1922); on this, see Charles Lock, ‘Inhibiting the Voice: Thomas Hardy and Modern Poetics’, in Keith Wilson (ed.), A Companion to Thomas Hardy (Oxford: Blackwell, 2009), 450-464: 463, n. 1. 8 With reference to Eliot’s and Williams’s denunciation of free verse, Chris Beyers readily admits that he ‘can think of no other formal movement for which two of its most important practitioners disclaim their chosen form’s very existence’ (Beyers (2001): 14). 9 E.g. Lewis Putnam Turco, ‘Verse vs. Prose/Prosody vs Meter’, in David Baker (ed.), Meter in English: A Critical Engagement (Fayetteville, AR: The University of Arkansas, 1996), 249-263. 10 This in response to Joseph Brodsky’s plea that ‘with free verse, the first question should be, free from what, free from what?’ (Cynthia L. Haven, Joseph Brodsky: Conversations (Jackson, MS: University Press of Mississippi, 2002): 28).

13

verse is that a salient phonological has been singled out as sufficiently prominent to be used as a building block in the construction of that work’s verse lines. The length of such verse lines is determined by a preordained number of occurrences of the phonological feature in question; when that number is reached, the line is complete and a new line must begin. Poetic metre, in other words, requires counting: it works by selecting a salient phonological feature of a language and arranging this feature in patterns that can be explicated and summarized numerically. What qualifies as a salient phonological feature suited for metrical enumeration differs from language to language, depending on the structure of those languages’ phonological systems. A good rule of thumb to keep in mind, however, is that in all viable systems of versification that we know of, the salient phonological features which have been selected for numerical organisation have invariably been tied to the syllable. We shall have more to say about the vexed status of the syllable in contemporary phonology later; for now, it will suffice to say that the syllable has proven to be an indispensable unit of measurement in the description of every extant metrical verse system. By far the most successful phonological features to be used as the basis for metrical systematizations are syllable length, syllable stress, syllable pitch and syllable count. In the case of Homer’s Greek, the salient phonological feature singled out for metrical organisation was syllable length. Metrical systems based on syllable duration are called quantitative and such systems can, in addition to Greek, be found with slight alterations in for instance Sanskrit, Urdu and Persian poetry. If a Greek syllable contained a long vowel, a diphthong, or a short vowel followed by two or more consonants (even across word boundaries), the syllable was considered long; if a syllable did not live up to these requirements, it was considered to be short. Long syllables formed the back-bone of Homer’s preferred metre, and every line of the Iliad is thus structured around a row of six long syllables, here signalled by macrons: 1 2 3 4 5 6 ‒ ___ ‒ ___ ‒ ___ ‒ ___ ‒ ___ ‒ ___ In the spaces after each of these long syllables, the poet was at liberty to place either two short syllables or a single long syllable. Two short syllables, in other words, were regarded as equivalent to one long syllable, and short syllables are traditionally signified by breves: ᴗᴗ = ‒

14

Each of the six long syllables that formed the back-bone of the metre, together with the long syllable or the two short ones which followed it, were conceived of as constituting a metrical foot. Metrical feet consisting of one long syllable followed by two short syllables (‒ᴗᴗ) are called dactyls; those consisting of two long syllables (‒ ‒) are called spondees. And because it takes six metrical feet to make a line of Homer, and because dactyls occur more frequently in Greek verse than do spondees, the metre of the Iliad is called dactylic hexameter. But dactyls and spondees were not allowed to mix completely at random in any single line of Homeric verse: the sixth foot of a dactylic hexameter was invariably a spondee,11 and the fifth foot was almost always a dactyl; spondees, in other words, could be substituted freely for dactyls only in the first four feet of any given line. A prototypical Greek dactylic hexameter therefore adheres to a matrix which may be formalized like this: │‒ᴗᴗ or ‒ ‒│‒ᴗᴗ or ‒ ‒│‒ᴗᴗ or ‒ ‒│‒ᴗᴗ or ‒ ‒│ ‒ᴗᴗ │ ‒ ‒ │ Three things are worth noting at this point, all of which will have important bearings on what we have to say about English verse further on. Firstly, it should be noted that Greek metrics makes ample use of substitution, both at the level of the foot and at the level of the syllable: a dactyl can be substituted by a spondee and vice versa because two short syllables in certain positions can be substituted by a long one. Secondly, that it is the number of feet in each line rather than the number of syllables which is being counted in classical Greek metrics: a completely uncontroversial line of dactylic hexameter can therefore consist of anything between thirteen and seventeen syllables, depending on the number of spondaic substitutions. Thirdly – and this is perhaps the most important point – in spite of the great freedom gained by the possibility of foot substitution, the preordained structure of the dactylic hexameter effectively bars the occurrence of certain words and collocations in this particular metre. No combination of dactyls and spondees will allow for three consecutive short syllables, just as any combination of the syllabic template short-long-short is precluded from occurring. Words such as μενετός (‘patient’ (adj.), short-short-short) and κοινότης (‘community’, long-short-long) are thus excluded from the lexicon of the dactylic hexameter: not on account of

11

In point of fact, the last syllable of the final spondee in a dactylic hexameter is very often occupied by a phonologically short syllable. This phenomenon – which Paul Maas has termed brevis in longo, i.e. short [syllable] in place of long [syllable] – has, however, no bearing on the line’s metricality, for which reason the final short syllable in such cases is counted as metrically long, and therefore the foot is considered to be equivalent to a spondee. The point is that the sixth foot of a dactylic hexameter is never a dactyl.

15

their semantic properties, but solely because of their phonological structure.12 This last point – that poetic diction is intimately connected with, and delimited by, the phonological requirements of the chosen metre – is one which we shall return to in some detail in Chapters 5 and 7. In addition to dactyls and spondees, Greek versifiers made use of a host of other metrical feet, all of which were based on the binary distinction between short and long syllables. Some of the names of these metrical feet are well-known to any reader of English verse, e.g. iamb (ᴗ‒) and anapaest (ᴗᴗ‒), while others belong to the province of scholars specializing in Greek verse, e.g. amphimacer (‒ᴗ‒) and molossus (‒ ‒ ‒). The exact inventory of Greek metrical feet is of limited relevance to our immediate purpose, which is to trace the ways in which Greek metrics was imitated and reinterpreted in the literatures of other languages. Before embarking on that enterprise, we must, however, pay due attention to another aspect of Greek versification, which is too rarely touched upon in critical engagements with that discipline, namely the alphabet in which it is embedded. The scholarly description of the intricacies of Greek metrics requires a vast and somewhat alien critical nomenclature, and it is perhaps this nomenclature which has imbued the field with a certain air of arcane fussiness which is not always appreciated by the modern student. Yet it must be admitted that an even greater obstacle to the enjoyment of Greek verse presents itself as soon as our eyes happen to meet something like this: μῆνιν ἄειδε θεὰ Πηληϊάδεω Ἀχιλῆος οὐλομένην, ἣ μυρί’ Ἀχαιοῖς ἄλγε’ ἔθηκε, This is the Iliad’s opening invocation of the goddess, imploring her to sing of the wrath that seized Peleus’ son Achilles and the countless woes it brought down on the Achaeans. However, to reading eyes accustomed only to the code of the Roman alphabet, the sheer visual otherness of the Greek lettering is likely to suppress this recognition and effectively prevent the goddess from singing. As familiar as we are with the content of these lines, most of us are completely in the dark as to how they should be sounded, a deficiency which is intimately linked to the use of the Greek alphabet. As Charles Lock notes, ‘The effect of an alien script is to silence the reader ... reduc[ing] us, however 12

It should be noted that there is a way to accommodate words consisting of three short syllables such as μενετός in the dactylic hexameter: since the rule by which a short vowel followed by two or more consonants becomes long also applies across word boundaries, a short syllable such as the last one in μενετός can be made long by beginning the next word with a consonant. No such tricks can be performed to include words of the long-short-long structure in the dactylic hexameter, for which reason the general argument – that certain metres preclude certain words from occurring – is still valid.

16

briefly, to baffled looking’.13 In order to bridge the gap between the iconic and the phonetic which an alien script invariably induces on the reader, recourse is habitually taken to transliteration into the Roman alphabet in contemporary typographical representations of Greek texts. Exposed to the most common standard of transliterating Greek into English lettering, Homer’s opening lines from the Iliad would look something like this: menin aeide thea Peleiadeo Achileos oulomenen, he muri’ Achaiois alge’ etheke. While this typographical transformation renders the sequence less alien to the eye, the ear remains positively baffled, not only as to the probable sounding of each word, but especially as to the correct scansion of the sequence. Very little help in determining which syllables are long and which are short is provided by such a transliteration, and the catalogue of diacritical marks that is sometimes called upon to remedy this deficiency is likely to render the transliteration almost as confusing to the eye as was the original Greek lettering. Even more demanding on the eye, but at least helpful to the scanning ear, is a genuine phonetic transcription into IPA, such as this one of the first line of the Iliad: [mɛː.nin a.eː.de tʰe.aː pɛː.lɛː.i.a.dɔː akʰ.i.lɛː.os]14 I have left out of this transcription any indication of stress since this prosodic feature plays only a marginal role in Greek metrics.15 I have, however, tentatively marked word-internal syllable boundaries by means of a baseline dot ; this practice allows us to perceive that the opening line of the Iliad contains sixteen syllables. If we parse the transcribed line for occurrences of the triangular colon , which indicates that the preceding vowel is long, we can see that our line contains seven such vowels: they occur in syllable positions 1, 4, 7, 8, 9, 12 and 15. And since none of the remaining syllables contains a diphthong or a short vowel followed by two or more

13

Charles Lock, ‘Heterographics: Towards a History and Theory of Other Lettering’, in Angles on the English-Speaking World, New Series, Vol. 6 (2006), edited by Ida Klitgård, 97-112: 100-101. 14 I have based my transcriptions of Greek on the principles laid down in W. Sidney Allen’s Vox Graeca: A Guide to the Pronunciation of Classical Greek (Cambridge: Cambridge University Press, 1968): 155-157 et passim. 15 M.L. Gasparov, A History of European Versification, translated by G.S. Smith & Marina Tarlinskaja, edited by G.S. Smith with Leofranc Holford-Strevens (Oxford: Clarendon Press, 1996): 85-87.

17

consonants, we may safely conclude that they are all short. In terms of syllable length our line thus scans like this: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ‒ᴗᴗ‒ᴗᴗ‒‒‒ ᴗ ᴗ ‒ ᴗ ᴗ ‒ ᴗ The only way in which the sequence ‒ᴗᴗ‒ᴗᴗ‒ ‒ ‒ᴗᴗ‒ᴗᴗ‒ᴗ can be split into dactyls and spondees in such a manner that no syllable is left unaccounted for is to introduce foot boundaries in these positions: │‒ᴗᴗ│‒ᴗᴗ│‒ ‒│‒ᴗᴗ│‒ᴗᴗ│‒ᴗ│ This gives us the expected count of six metrical feet, of which the third and the sixth are spondees (even though the latter of these is an example of brevis in longo) and the rest are dactyls. With this in place, we are now in a position to map onto the Greek line its correct scansion pattern: │ ‒ ᴗ ᴗ │ ‒ ᴗ ᴗ │‒ ‒ │ ‒ᴗᴗ │ ‒ ᴗ ᴗ │ ‒ ᴗ│ │μῆνιν ἄ-│ειδε θε-│ὰ Πη-│ληϊά-│δεω Ἀχι-│λῆος│

One observation immediately forces itself upon the parsing eye, namely that foot boundaries are not in any way determined by word boundaries. In the line above, each line-internal foot boundary thus falls inside a polysyllabic word. While this technique of consistently counterpointing foot divisions with word divisions is usually considered a hallmark of foot-based prosodies, it is not a requisite in Homer’s versification, as can be seen from this scansion of the second line of the Iliad: │ ‒ ᴗ ᴗ│ ‒

‒ │ ‒ ᴗ ᴗ │ ‒ ‒ │‒ ᴗ ᴗ│ ‒ ᴗ │

│οὐλομέ│νην, ἣ │μυρί’ Ἀ│χαιοῖς│ἄλγε’ ἔ│θηκε,│

In this line, word boundaries coincide with foot boundaries after the second, fourth and sixth foot. Even so, the ideal of counterpointing foot division and word division turned out to be a remarkably viable feature of subsequent adaptations of the Greek metrical system in other languages, English

18

included. For now, however, we should simply note that metrical scansion and lexical parsing are two different levels of analysis that do not in any significant way influence one another. The short introduction to Homer’s versification provided here does not purport to grant the reader anything but a faint glimpse into the intricate system of Greek quantitative metrics. Even such a glimpse ought, however, to make it clear that the type of verse which we find in Homer, with its highly sophisticated rules for composition, is indicative of a fully fledged metrical system, the roots of which must be sought significantly further back in time.16 Inquiries into the origins of metrical verse in a larger Indo-European context began, not surprisingly, in Germany in the nineteenth century when scholars such as Adalbert Kuhn and Theodor Bergk sought to apply the comparative method as developed by Rasmus Rask, Jakob Grimm, Franz Bopp and August Schleicher to the province of metre. However, these early attempts to reconstruct the sort of IndoEuropean Urvers from which Greek and other ancient systems of versification had sprung largely fell on stony ground. As M.L. Gasparov, the greatest authority on comparative historical metrics, notes, the German philologists responsible for these initial inquiries were all classical scholars, for which reason their endeavours were mostly directed towards Greek verse: But Greek versification is a very bad starting-point for historical research, since in Homer and the lyric poets we find it already in full flower, while practically no records remain from its formative period. Things had to be reconstructed by analogy, and the analogies that came to mind were mainly their own Germanic ones.17 Since the structural basis of Germanic metrics is syllable stress rather than syllable length, scholars such as Kuhn and Bergk made little headway in establishing a plausible common origin that could account for the development of both of these systems, and it was not until the twentieth century that a convincing theory of Indo-European metrics was put forward. When this theory finally did emerge, it came from a French rather than a German scholar, namely from Antoine Meillet. A student of Ferdinand de Saussure, Meillet served as a professor at the Collège de France from 1906 to 1932 where he not only supervised Milman Parry’s groundbreaking dissertation on the oral-formulaic origin of Homer’s epics, but also came to exercise a considerable influence on the next generation of eminent French linguists, among them Émile Benveniste, André Martinet and 16

For a thorough survey of the pre-historical development of Greek quantitative metre, see M.L. West, ‘Greek Poetry 2000-700 B.C.’, in The Classical Quarterly, New Series, Vol. 23, No. 2 (Nov., 1973), 179-192. 17 Gasparov (1996): 5.

19

Lucien Tesnière. In 1923, Meillet published a short treatise entitled Les origines indo-européennes des mètres grecs, in which he put forward the theory that neither stress nor syllable length were likely to have played a constitutive part in the oldest species of Indo-European versification. Based on a comparison between Aeolic Greek and Sanskrit metres, Meillet instead proposed that isosyllabism – i.e. the principle of observing a fixed number of syllables in each line – was the most likely candidate for the original metrical principle from which subsequent metrical systems had evolved. Meillet’s notion of isosyllabism does not tally well with what we observed earlier about the dactylic hexameter – that the number of syllables in this metre may vary between thirteen and seventeen syllables – but here it should be remembered that Homer’s work represents a comparatively late development of the Urvers with which Meillet was concerned.18 Isosyllabism is, however, a prominent feature of early Sanskrit (and Aeolic) versification (as it indeed is in the modern species of French versification, with which Meillet would have been intimately familiar) and it was these types of verse which Meillet focused on in his comparative investigations. In doing so, he diverged significantly from his German predecessors whose focus had been on bringing into alignment stress and syllable length. At the time when they were first presented, Meillet’s findings were, of course, merely hypothetical (as he himself admits),19 but they were largely corroborated some thirty years later by Roman Jakobson’s important article ‘Slavic Epic Verse: Studies in Comparative Metrics’ (1952), in which Slavonic verse was shown to have originated from a common Indo-European form of syllabic verse. Further evidence for Meillet’s conception of Indo-European verse was provided by studies in Celtic verse by Calvert Watkins20 and in Latin Saturnian verse by Thomas Cole;21 this body of work was summed up and expanded in West (1973b). This is obviously not the place to go into details about the substance and problems of reconstructing Indo-European metrics, but we have allowed ourselves to outline that discipline’s history here in order to make clear that the beginning of western literature – the texts in which this conglomerate is preserved and handed down to us – cannot reasonably be viewed as the beginning of metrical verse. Metrical verse, whether in written or in oral form, predates our earliest recorded 18

In point of fact, Meillet excluded the dactylic hexameter from his theory on account of the fact that no apparent Sanskrit cognate for this metre could be found; instead he proposed an origin of this metre outside of the Indo-European area. This interpretation has, however, been shown to be groundless by Gregory Nagy, on which see Richard Stoll Shannon, III, The Arms of Achilles and Homeric Compositional Technique (Leiden: E.J. Brill, 1975): 12-13. See also M.L. West, ‘Indo-European Metre’, in Glotta, 51, Bd., 3./4. H. (1973), 161-187: 169, n. 10. 19 See Gasparov (1996): 7. 20 Calvert Watkins, ‘Indo-European Metrics and Archaic Irish Verse’, in Celtica, 6 (1963): 194-209. 21 Thomas Cole, ‘The Saturnian Verse’, in Studies in Latin Poetry, Yale Classical Studies, 21 (London: Cambridge University Press: 1969), 1-73.

20

instances of literary scraps by such a vast span of time that our decision to begin the present discussion of metricality in verse with Homer is, in fact, a beginning in medias res. In order to account for the fact that metre appears to be as deeply embedded in the very fabric of our linguistic behaviour as syntax and morphology, a number of explanations have been offered: all of them, alas, more appealing to imaginative speculation than to scholarly reasoning. Consequently, the question of why humans have found it worth their while to subject the naturally occurring phonological features of language to numerical segmentation is one to which we shall not attempt to provide an answer here. Rather, the one insistent question in this context is why this ancient trait of human behaviour was suddenly deemed trivial at the beginning of the twentieth century when free verse emerged as the dominant form of versification on a global scale. Before we can venture an answer to that question, we must, however, pay due attention to that early species of verse which occupied the minds of German scholars like Kuhn and Bergk to such an extent that they overlooked the syllabic component in Indo-European versification, namely Germanic accentual verse.22 Germanic accentual verse exemplifies even more saliently than does Greek quantitative verse the point made above: that our textual attestations of a given verse form usually represent rather late stages of that form’s development. Whereas even the earliest extant attestations of Greek verse are commonly seen as indicative of a metrical system in full flower, the text corpus of Germanic accentual verse is generally viewed as the remnants of a largely unattested golden age of that particular system of versification. This is true especially for the geographical area which is now southern Germany and Austria: less than 200 lines of Old High German accentual verse have been preserved, and they occur in only four works: the Hildebrandslied, the Muspilli, the Merseburg Incantations and the Wessobrunn Prayer. These texts represent a fascinating interim between the older, more purely accentual verse form, which is our immediate concern, and the subsequent developments in German and Anglo-Saxon systems of versification, to which we shall return. The salient phonological feature used for metrical enumeration in Germanic accentual verse is syllable stress rather than syllable length: what is regulated in this type of verse is the number of stressed syllables in each line rather than the duration of each single syllable. Given the naturalness with which stress is perceived by a native speaker of a modern Germanic language, it is curious to observe the caution with which phoneticians seek to define what exactly is entailed by this complex 22

In fairness to Bergk and Kuhn and their contemporaries, it should be noted that Germanic accentual verse is inadequately accounted for by Meillet’s isosyllabic hypothesis (Gasparov (1996): 7); for elucidation, see Martin J. Duffell, A New History of English Metre (London: Modern Humanities Research Association and Maney Publishing, 2008): 57-58.

21

prosodic phenomenon. Most modern textbooks thus describe stress as a combination of increased loudness, pitch and duration,23 a description which hardly meets the degree of exactitude usually strived for in this academic field. However, for our purposes such a definition is not only adequate but also illuminating, for reasons which will be made clear later. For now, let us simply revel in the fact that the notion of syllable stress is grasped as easily by us as the notion of syllable length was grasped by a literate Greek some 2,500 years ago. The ease with which stress is still perceived by modern speakers of Germanic languages greatly facilitates our comprehension of the Germanic accentual verse system, whose primary characteristics can seem almost trivial in comparison with the intricate technicalities of the Greek quantitative verse system: in general, lines of Germanic verse are split into two hemistichs, each of which contain two major stresses tied together by means of structural alliteration. Unstressed syllables, in other words, are not regulated in Germanic accentual versification; only strongly stressed syllables count, whereas the number of unstressed syllables between them is metrically insignificant. The earliest example of this metrical system to be preserved dates from the early fifth century CE and consists of a single line that was carved round the rim of a golden horn, found at Gallehus in Southern Jutland in 1734. It reads:

‒ which transliterates into: / / / / ekhlewagastiR⁞holtijaR⁞horna⁞tawido⁞ In translation: ‘I, Hlewagast, Holt’s son, made the horn’.24 Not, perhaps, such stuff as epics are made on, but a fine line of accentual Germanic verse nonetheless. It comes with the expected number of four stressed syllables, the first three of which 23

E.g. John Laver, Principles of Phonetics (Cambridge: Cambridge University Press, 1994): 511; R.L. Trask, A Dictionary of Phonetics and Phonology (London: Routledge, 1996): 269; Peter Roach, Phonetics (Oxford: Oxford University Press, 2001): 32. 24 Geoffrey Russom, Beowulf and Old Germanic Metre (Cambridge: Cambridge University Press, 1998): 1. The same source also informs us that the rune transliterated as ‘represents a sound derived from Germanic z that had not yet merged with r’.

22

alliterate, and it will furthermore be observed that between the stressed syllables an unregulated number of unstressed syllables, ranging from one to three, is allowed to occur. No special graphic indication of the structural segmentation into hemistichs is provided by the original inscription of this line; even so, it follows a very common pattern by which the first hemistich (‘ekhlewagastiR holitjaR’) tends to be slightly longer than the second hemistich (‘horna tawido’). In addition, the alliterative pattern of the Gallehus horn inscription exemplifies the most commonly found version, in which the two stressed syllables of the first hemistich alliterates with the first stressed syllable of the second hemistich. All of these features will, of course, be familiar to readers of Beowulf, which poem is clearly wrought on the same metrical trellis as Hlewagast’s inscription: sceaþena þrēatum,

Oft Scyld Scēfing monegum mǣgþum, egsode Eorl[e].

meodo-setla oftēah; Syððan ǣrest wearð

fēasceaft funden;

hē þæs frōfre gebād:

wēox under wolcnum, oðþæt him ǣghwylc ofer hron-rāde gomban gyldan:

weorð-myndum þāh, þāra ymb-sittendra

hȳran scolde, þæt wæs gōd cyning! (Beowulf, Introduction, 4-11)25

The earliest successful attempt to summarize the basic structure of the different permissible permutations of Germanic alliterative verse was made by the Junggrammatiker Eduard Sievers, whose division of Anglo-Saxon accentual hemistichs into five separate categories is still a cornerstone in modern treatments of that verse system.26 Sievers later abandoned his own original system of classification in favour of a highly complex and idiosyncratic construct which he called Schallanalyse, but it was his early work on Germanic metrics which would be the more influential, even outside the narrow field of historical metrics. According to Christine Brooke-Rose, it was Sievers’s original analysis of Anglo-Saxon metre which most likely served as a guide for Ezra

25

Cited from C.L. Wrenn, Beowulf and the Finnesburg Fragment, 1973 Edition, fully revised by W.F. Bolton (London: Harrap, 1973). 26 Eduard Sievers, ’Zur Rhythmik des Germanischen Alliterationverses’, in Pauls und Braunes Beiträge, 10 (1885): 209-314, 451-545.

23

Pound’s experiments with that metre in The Seafarer and in the Cantos.27 Sievers’s fivefold classification scheme was taken up and radically expanded by Alan J. Bliss, who identified no fewer than 130 different sub-divisions of Sievers’s five basic types,28 and by Thomas Cable, whose work focuses on comparing and contrasting metrical and linguistic norms in Anglo-Saxon for the purpose of explaining exactly why Sievers’s five types were permissible and other permutations not.29 More recent attempts to account for the intricacies of Germanic accentual versification include Geoffrey Russom’s work, by which hemistich structure is explained by reference to permissible permutations of word structure,30 and that by Paul Getty, who has employed Optimality Theory to identify the hierarchy of constraints by which Germanic alliterative verse is structured. 31 Duffell (2008) argues that the scholarly attention which has been bestowed upon Germanic accentual verse in recent decades is a welcome addition to the corpus of literature on historical metrics on the grounds that this type of versification seems to be singularly at odds with Meillet’s syllabic theory of Indo-European metre. Using Hlewagast’s inscription as evidence, Duffell makes the case that Germanic accentual verse in its origin appears to have been based on the principle of regulating the number of words rather than of syllables in each line: ‘Hlewagast’s metrical intention is not thirteen syllables, but four words, as the colons (i.e. ) and the inclusion of the pronoun “ek” in the first word clearly show’.32 It is an interesting theory, and one which merits exploration in more detail than can be permitted within the limits of this investigation. It should, however, be noted that Duffell’s point about the missing word boundary symbol on the Gallehus horn could just as easily be construed as evidence for the opposite interpretation, namely that metrical junctures in Old Germanic versification were not concurrent with word boundaries. This objection does not invalidate Duffell’s more general point – that Germanic verse is poorly accounted for by Meillet’s hypothesis – but it does make it prudent for us to stick to Gasparov’s formulation of the basic metrical principle of Germanic verse as: ‘any group of syllables united by one stress is equivalent to any other group of syllables also united by one stress’.33

27

Christine Brooke-Rose, A ZBC of Ezra Pound (Berkeley & Los Angeles, CA: University of California Press, 1971): 88, n. 1. 28 Alan J. Bliss, The Metre of Beowulf (Oxford: Blackwell, 1958) & An Introduction to Old English Metre (Oxford: Blackwell, 1962). 29 Thomas M. Cable, The Meter and Melody of ‘Beowulf’ (Urbana, IL: University of Illinois Press, 1974). 30 Geoffrey Russom, Old English Meter and Linguistic Theory (Cambridge: Cambridge University Press, 1987) & Russom (1998). 31 Paul Getty, The Metre of ‘Beowulf’: A Constraint-Based Approach (Berlin & New York: Mouton de Gruyter, 2002). 32 Duffell (2008): 58. 33 Gasparov (1996): 37.

24

This simple principle adequately accounts for the metrical structure of Scandinavian, Icelandic and early Anglo-Saxon (Caedmon, Cynewulf, Beowulf) modes of versification, but it begins to break down when we look at High German and Middle English species of accentual verse. While Scandinavian and Icelandic versifiers tended to uphold the accentual principle with purist rigour at least up until the fourteenth century, their colleagues further south went in the opposite direction by allowing more and more syllables to creep into their hemistichs, eventually abandoning the metrical trellis of Germanic alliterative versification altogether.34 This process first affected German verse: in fact, as early as the ninth century, the first German poet whose name has been handed down to us through his work, Otfrid of Weissenburg, employed end-rhyme rather than alliteration in his Evangelienbuch (a work which would later influence Samuel Taylor Coleridge, as we shall see in the next chapter). In Britain similar developments began to take place from the eleventh century onwards in for instance Layamon’s Brut. The explanation for the gradual erosion and subsequent abandonment of purely Germanic versification in these geographical areas ought to be evident: it must have been the result of prolonged contact with speakers of Romance languages and influence from their modes of versification. As a monk, Otfrid was not only well-versed in Latin,35 but the abbey in which he composed his rhymed gospel harmony was furthermore located in present-day Alsace, a region which even today remains a contested ground between the two most widespread ethno-linguistic groups in Western Europe, the Germanic and the Romance. In Britain, on the other hand, these two ethno-linguistic groups came to merge fruitfully with one another in the centuries which followed the Norman Conquest in 1066. The long-term metrical repercussions of this merging for English poetry are treated in more detail in Chapter 3; our present concern here is to outline the process by which Romance versification acquired the characteristics which were taken over and modified by English versifiers in the aftermath of 1066. The Romance languages are conventionally defined as those languages which can be traced back to Latin, for which reason it seems reasonable to begin our present quest by uncovering the roots of this language’s metrical practice. The attested history of Latin verse begins, somewhat pathetically, with the all but complete annihilation of that Italic verse form whose demise is brazenly condoned in Horace’s Epistle to Augustus in the following manner:

34

Gasparov (1996): 40-43. As can be witnessed in Otfrid’s perceptive comments about his Evangelienbuch, penned in impeccable Latin prose and directed to Liutbert, the Archbishop of Mainz. 35

25

sic horridus ille defluxit numerus Saturnius, et grave virus munditiae pepulere. (Horace, Epistulae, 2, 1, 157-159) Of ‘this horrid Saturnian measure’ only fragments – some 125 lines in total – have been preserved to our day, and the fundamental metrical principle on which they are structured remains deplorably sketchy.36 All that is known is that the ‘munditiae’ (i.e. ‘worldly sophistication’) to which Saturnian verse eventually succumbed was the unabashed adaptation of Greek quantitative metrics in Latin. Credit for having introduced quantitative metrics based on Greek models in Latin is usually given to Ennius; indeed, a few lines before Horace’s disdainful denunciation of the ‘numerus Saturnius’ quoted above, we are informed that Ennius was commonly named ‘alter Homerus’, ‘the other Homer’, by earlier Latin critics.37 This epithet acknowledges and pays tribute to the fact that Ennius’ Annales (composed between 184 and 169 BCE) was the first poem to employ Homer’s preferred measure – the dactylic hexameter – in Latin. The Annales, of which some 600 lines have survived, is an epic poem tracing Roman history all the way back to that pivotal event which the wrath of Achilles had ultimately effected, namely the sack of Troy. The related story of how a certain survivor from the fall of Troy escaped and subsequently made Rome a Trojan colony was, of course, taken up by Virgil some 150 years later; that poet, too, employed the dactylic hexameter for his great epic, the Aeneid, whose opening line reads: Arma virumque cano Troiae qui primus ab oris Given that the basic Greek rules for determining syllable quantity also apply to Latin, Virgil’s line must be scanned like this:

36

Cole (1969) thus defines Saturnian verse as ‘any archaic Latin verse that is used stichically and divided by a caesura into two parts, the first of which contains five to nine syllables, the second (usually one to three syllables shorter than the first) five to eight syllables’ (10). 37 Horace: Epistulae, 2, 1, 50. The unnamed critics who describe Ennius as alter Homerus may well have taken their cue from Ennius himself: in the Proem to the Annales, Ennius thus relates how upon falling asleep he is transported to Mount Helicon where he is met by Homer. The Greek bard then assures his Latin heir that he, Ennius, is in fact a reincarnation of him, Homer.

26

│ ‒ ᴗ ᴗ│ ‒

ᴗ ᴗ│‒

‒│ ‒ ‒│ ‒

ᴗ ᴗ│‒ᴗ│

│Arma vi│rumque ca│no Troi│ae qui│primus ab│oris│ Even with the scant treatment of Homer’s versification earlier in this chapter as our only guide, we are able to see that this line by Virgil unambiguously fits the same trellis as the lines we looked at in the Iliad. However, the ease with which we recognize the dactylic hexameter – even when it has been transplanted from its native Greek habitat to Latin soil – should not be allowed to obfuscate the very real problems that this transplant entailed. Chief among these problems was the ‘embarrassingly high proportion’38 of long syllables in the Latin language compared to Greek, as a result of which Latin hexameters in general came to exhibit a more spondaic profile than did Greek ones.39 Another crucial difference between Latin and Greek was the role played by stress in the two languages: whereas stress for all metrical purposes was insignificant in Greek, it played a prominent part in spoken Latin. Latin word stress falls on the first syllable of disyllabic words (e.g. móntes) and on the penultimate syllable of polysyllabic words (e.g. magíster) unless that syllable is short, in which case stress moves to the antepenultimate syllable (e.g. dóminus).40 This, however, is not to say that word stress played a constitutive role in classical Latin versification: quantity and quantity alone is what determines the segmentation into metrical feet in the line by Virgil above, as can be seen from the fact that its two medial spondees begin with an unstressed syllable. Even so, there is evidence to suggest that the prominence of word stress in Latin did influence the way in which quantitative verse was composed, especially in less orotund genres than the epic. An instance of this is the law of brevis brevians, by which a long syllable following a short syllable could be shortened when it was either preceded or followed by a stressed syllable (e.g. vŏlūptátem (ᴗ‒ ‒ᴗ) → vŏlŭptátem (ᴗᴗ‒ᴗ)).41 This device not only helped to boost the proportion of short syllables in Latin; it presumably also brought metrical forms of Latin

38

D.S. Raven, Latin Metre: An Introduction (London: Faber and Faber, 1965): 17. Gasparov (1996): 72. 40 Raven (1965): 32. These principles are laid out in Quintilian’s Institutio Oratorio (1.30-31): ‘[T]he rule [for assigning stress/acute accent] is simplicity itself. For in every word the acute accent is restricted to three syllables, whether these be the only syllables in the word or the three last, and will fall either on the penultimate or the antepenultimate. The middle of the three syllables of which I speak will be either acute or circumflex, if long, while if it be short, it will have a grave accent and the acute will be thrown back to the preceding syllable, that is to say to the antepenultimate. Every word has an acute accent, but never more than one. Further, the acute never falls on the last syllable and therefore in disyllabic words marks the first syllable.’ Cited in Harry van der Hulst (ed.), Word Prosodic Systems in the Languages of Europe (Berlin & New York: Mouton de Gruyter, 1999): 477. 41 Raven (1965): 36-37. 39

27

into closer alignment with the intonational patterns of everyday speech, for which reason it was especially favoured in drama. Marginal as the phenomenon of brevis brevians may seem, its reliance on word stress in determining syllable quantity was an important indication of things to come. As the Roman Empire continued to expand its borders, especially during the first century CE, the Latin language came into contact with a host of other languages, and Rome herself became home to an ever increasing number of non-native speakers. This spawned a bundle of Latin dialects and sociolects that we now refer to by the blanket term Vulgar Latin. In the process from Classical to Vulgar Latin, word stress showed itself remarkably resilient to change, whereas the old vowel system was reorganized in such a way that long syllables gradually became indistinguishable from short syllables by ear alone, even among educated speakers. That did not mean, however, that Latin quantitative verse ceased to be composed; in fact, as Gasparov points out, ‘The Middle Ages and the Renaissance produced more poems written impeccably according to the rules of classical metrics than we have inherited from genuine antiquity’.42 The principles of classical quantitative metrics thus survived in educational systems all over Europe where students were taught to distinguish between long and short syllables according to an elaborate set of rules rather than by ear. In addition, they were taught to compose verse on the basis of these ossified principles, a practice which lasted well into the twentieth century.43 If the sack of Troy can be viewed as the ancestral cataclysm from which classical quantitative metrics proliferated, then the end of that metrical system as a living matrix for composing verse may conveniently be said to have coincided with the sack of Rome in 410 by the Visigoths. By this point in time, the distinction between long and short syllables had disappeared, stress reigned supreme, Germanic barbarians controlled the streets of Rome, and the beliefs of an obscure Judean sect had been promoted to the empire’s official religion. That the fall of Rome to the Visigoths in many respects marks the ending of that epoch which had begun with the fall of Troy is a view which was shared by many a contemporary commentator. Thus, in a letter to Principia from 412 (CXXVII), St. Jerome pithily summarizes the significance of Rome’s fall in the sentence ‘capitur Urbs quae totem cepit orbum’ (‘The City to which the whole world fell, has 42

Gasparov (1996): 89. An indication of how recently this practice was abandoned in Britain is that Ainger & Wintle’s An English-Latin Gradus or Verse Dictionary (1890) saw its seventeenth and – so far (after all, hope costs nothing) – final imprint as late as 1963; in the preface the authors present verse composition in Latin as ‘the proof and the flower of that scholarship which loves the old writers with an unselfish love, and delights to clothe modern thoughts and modern expressions in the dress of ancient metre and rhythm’ (cited in Nicholas Ostler, Ad Infinitum: A Biography of Latin (London: Harper Press, 2007): 303). 43

28

fallen’), before quoting the following portion from Virgil’s account of Troy’s fall as a framework for his own observations: quis cladem illius noctis, quis funera fando explicet aut possit lacrimis aequare labores? urbs antiqua ruit multos dominata per annos; (Book 2, 361-363) In Theodore C. Williams’s translation: But who the bloodshed of that night can tell? What tongue its deaths shall number, or what eyes Find meed of tears to equal all its woe? The ancient City fell, whose throne had stood Age after age.44 At the beginning of De civitate Dei, a work whose conception was directly motivated by the fall of Rome, St. Augustine likewise quotes copiously from the Aeneid, if only to make the point that, unlike the Achaeans responsible for the sack of Troy, the recent conquerors of Rome had shown mercy on account of their Christian faith. This was in response to the widespread sentiment among Romans at the time that it was the abandonment of the old religion in favour of Christianity which had paved the way for the Gothic invasion. There would, perhaps, have been more validity to the claim that the adoption of Christianity had helped pave the way for the definitive demise of Classical Latin. After Christianity was made the state religion of the Roman Empire under Theodosius I, the Church Fathers rejected the rhetorically refined Latin of yore in favour of the contemporary idiom of sermo humilis – the humble speech of ordinary people – as the means for propagating their doctrines. The driving force behind this strategy was Augustine, whose defence against allegations of linguistic barbarism in Christian writings can be found in his De doctrina Christiana:

44

Cited from The Æneid of Virgil, translated into English Verse by Theodore C. Williams (Boston & New York: Houghton Mifflin Company, 1908).

29

And likewise, what is a barbarism but pronouncing a word differently from those who spoke Latin before us? For whether the word ignoscere [‘pardon’] should be pronounced with the third syllable long or short is indifferent to the man who is praying to God....45 Such pesky details are, however, a matter of great concern to the man who is attempting to write quantitative verse in Latin, though this was not Augustine’s aspiration. In fact, Augustine’s only work in verse, his ‘Psalm against the Donatists’ (393), should be credited as the first known Latin hymn based on a strictly syllabic principle with no regard to quantity.46 Augustine’s hymn consists of 288 lines of sixteen syllables divided into two hemistichs of eight syllables each, and it even employs a sort of end-rhyme in that every line terminates in an unstressed -e (occasionally -ae): Propter hoc dominus noster ║ voluit nos praemonere Comparans regnum caelorum ║ reticulo misso in mare Congreganti multos pisces ║ omne genus hinc et inde. (Augustine, ‘Psalm Against the Donatists’, ll. 3-5)47 The ‘Psalm Against the Donatists’ represents a very early outcome of that process which M.L. Gasparov has termed ‘the Great Resyllabization of south European Verse’.48 As the distinction between long and short syllables in Latin became increasingly blurred, Latin versifiers found themselves at a loss as to what they should be counting instead, and the choice first fell on the number of syllables in each line. Later on, they would combine this principle with the principle of also observing a regular stress pattern, especially towards the end of each line, but the earliest stage of post-quantitative verse in Latin – of which Augustine’s psalm is an example – can best be described as isosyllabic. The reason that we allow ourselves to dwell on this intermediary stage of Latin versification is that it was during this period of isosyllabism that the dactylic hexameter fell out of favour. We have already noted that the dactylic hexameter was characterized by its ability to hold as many as seventeen and as few as thirteen syllables: for this reason alone the dactylic hexameter provided an exceptionally poor model for the new ideal of maintaining a fixed number of syllables in each line. 45

Cited in Ostler (2007): 117. Gasparov (1996): 89. See also William Beare, Latin Verse and European Song. A Study in Accent and Rhythm (London: Methuen, 1957): 248-250. 47 Cited from Carolinne White, Early Christian Latin Poets (London: Routledge, 2000): 53. 48 Gasparov (1996): 88. 46

30

Other classical verse forms were better suited for this kind of metrical recasting, prime among them the iambic trimeter and the iambic dimeter: the resyllabized forms of these metres would evolve into the dodecasyllable and the octosyllable respectively.49 It was thus these two metres that came to be passed on from the classical to the medieval world, leaving the dactylic hexameter behind as an obsolete remnant of former glory.50 This did not happen overnight; throughout the course of the fifth and sixth centuries Christian writers continued to make use of the quantitative dactylic hexameter. The most common use of this metre was for metrical recastings of select books from the Bible, such as Sedulius’s Carmen paschale (based on the gospels), Marius Victor’s Alethia (based on Genesis) and Arator’s De actibus apostolorum.51 Like Ambrose, whom Augustine credits with having introduced in Latin ‘hymns and psalms sung after the custom of the eastern Churches’,52 Arator spent time in Milan, but he eventually ended up in Rome as a subdeacon of the Roman Church, in which capacity he composed the Actibus as part of his service.53 Arator’s work is singled out here because it is the latest work to be featured copiously in two immensely important medieval treatises on metrics: Aldhelm’s De metris and Bede’s De arte metrica. Significantly, both of these treatises were composed in Britain rather than Italy. As William Beare succinctly points out, ‘Prosody, Latin prosody, was the first science studied by Englishmen’,54 and Aldhelm’s treatise should be credited as the first proper manual of quantitative versification composed anywhere. Before this time, the only treatises on metrics available to aspiring poets – such as those by Festus Aphthonius and Mallius Theodorus – were mostly concerned with compiling examples chosen for their antiquarian value and thus offered little in the way of helping the newcomer. This deficiency was remedied by Aldhelm’s De metris, which provides practical instruction on how to compose lines of dactylic hexameter in Latin for non-native speakers of that language. Being the first speaker of a Germanic language to have composed extant works in Latin quantitative verse,55 Aldhelm was eminently qualified to write just such a treatise, but it was Bede’s De arte metrica which was to have the most lasting influence on medieval metrics. Composed at the beginning of the eighth century, Bede’s De 49

The subsequent fate of the dodecasyllable and the octosyllable in Byzantine, Romance and Germanic poetry forms the subject of Chapter 3. 50 The only other classical metre to make the transition into medieval versification was the trochaic trimeter, which came to be codified as a fifteen-syllable line divided into hemistichs of seven and eight syllables. 51 See Marcia L. Colish, The Stoic Tradition from Antiquity to the Early Middle Ages (Leiden: E.J. Brill, 1990): 93-94. 52 Saint Augustine, Confessions, translated by Henry Chadwick (Oxford: Oxford University Press, 1991): IX, vii. 53 White (2000): 159-160. 54 Beare (1957): 32. 55 Michael Lapidge, ‘Aldhelm’s Latin Poetry and Old English Verse’, Comparative Literature, Vol. 31, No. 3 (Summer, 1979), 209-231: 209.

31

arte metrica soon became the most authoritative treatise on metrics in the Middle Ages, and it is in that work that we find the earliest attempt to distinguish between quantitative metrics and the new species of ‘rhythm’ which had been developing in the centuries after Augustine’s psalm: It appears that rhythmus is similar to metric, for it is a composition with modulation of words, without metrical measure, but arranged by the number of syllables to please the ears, as are the songs of vernacular poets. And thus rhythmus can be without meter, although meter cannot be without rhythm; which is why it is said, metrum is measure with modulation, rhythmus is modulation without measure.56

56

Bede, De arte metrica, xxiv; in James J. Murphy, Rhetoric in the Middle Ages: A History of Rhetorical Theory from Saint Augustine to the Renaissance (Berkeley, Los Angeles, London: University of California Press: 1974): 78.

32

Chapter 2 The Thawing of an Icicle: Metre and Theory What an elaborate theory have we here, Ingeniously nursed up, pretentiously Brought forth, pushed forward amid trumpet-blast, To account for the thawing of an icicle Robert Browning, The Ring and the Book

The theoretical distinction between rhythm and metre, which Bede touches upon toward the end of his De arte metrica, is a crucial one, and one which has occupied metrists ever since antiquity. It is, however, important to realize that metrists in different ages have used each concept with drastically different, and sometimes conflicting, meanings. Bede – as we saw in the preceding chapter – clearly identifies metre with quantitative versification and reserves the term rhythm to describe that new form of composition which merely regulates the number of syllables in each line: by modern standards such syllable-counting strategies would, of course, be regarded as metrical as well. In antiquity, on the other hand, the terms rhythm and metre were used to distinguish the guiding principles of two different theories of quantitative verse: the rhythmikoi held that poetry was so closely aligned to music that the structure of verse was identical to the rhythmic structure of the melody to which it was sung (for which reason the actual duration of syllables could vary greatly), 1 whereas the metrikoi held that a formal binary distinction between long and short syllables was sufficient to account for quantitative verse structure (for which reason long syllables could be substituted by two short ones).2 The former view is associated with Aristotle’s pupil Aristoxenus of Tarentum, a renowned music theorist; the latter with the grammarian Hephaestion of Alexandria. Needless to say, it is the linguistically tinted view of the metrikoi which informed the preceding chapter’s overview of metrical systems, and this choice of theory justifies in and by itself the

1

Rudolf Westphal, Die Fragmente und die Lehrsätze der Griechischen Rhythmiker (Leipzig: Druck und Verlag von B.G. Teubner, 1861): 6. 2 John Hollander, ‘The Music of Poetry’, in The Journal of Aesthetics and Art Criticism, Vol. 15, No. 2 (Dec., 1956), 232-244: 235; see also Alex Preminger & T.V.F. Brogan (eds), The New Princeton Encyclopedia of Poetry and Poetics (Princeton, NJ: Princeton University Press: 1993): 787, s.v. ‘Metrici and rhythmici’.

33

purpose of this chapter, which is to outline the various theoretical constructs through which poetic metre, as well as the absence thereof, has been approached in the western world. From a historical perspective it must be conceded that the approach of the metrikoi has had a far more constant and significant impact on our conception of verse structure than has that of the rhythmikoi, especially in recent centuries. The standard view today is that versification is more adequately accounted for by reference to its linguistic properties than to its musical properties, and that formalized binary oppositions rather than precise temporal spans are what ultimately structure metrical verse. That this is so is in no small part contingent upon the troubles that Renaissance metrists encountered when they sought to bring into alignment the theory of classical quantitative verse with the doings of poets working in the vernaculars, particularly those vernacular languages in which stress was a prominent feature.3 The task of these early vernacular metrists was not an easy one: unaided by the insights of modern phonetics, they habitually conflated stress, length and pitch in their endeavours to distil the basic principles of the new vernacular poetry. This is evident in our earliest treatise devoted entirely to English versification, George Gascoigne’s ‘Certayne Notes of Instruction concerning the making of verse or ryme in English’ (1575), in which poets are admonished to ‘place every word in his natural emphasis or sound ... in such wise, and with such length or shortness, elevation or depression of syllables, as it is commonly pronounced or used’.4 This is good and sound advice to any aspiring poet, and Gascoigne’s terminology for designating what makes certain syllables stand out as more prominent than others is, it should be noted, only slightly more precarious than what is used today: ‘Other things being equal,’ John Laver explains in his standard work on phonetics, ‘one syllable is more prominent than another to the extent that its constituent segments display higher pitch, greater loudness, longer duration or greater articulatory excursion from the neutral disposition of the vocal tract’. 5 Syllabic prominence, in other words, is a complex and multilayered phenomenon, the terminological explication of which should not be used to disqualify en bloc the general insights of the pioneers of Renaissance metrics. T.V.F. Brogan, the formidable modern compiler of all things pertaining to metrical theory, advices readers of early metrical treatises to ‘attend not to what a metrist says but to

3

The principal texts of English Renaissance prosody are collected in Brian Vickers (ed.), English Renaissance Literary Criticism (Oxford: Oxford University Press, 1999). 4 George Gascoigne, ‘Certayne Notes of Instruction concerning the making of verse or ryme in English’ [1575], in Vickers (ed.) (1999), 162-171: 164. 5 Laver (1994): 511.

34

what he does’ by examining first the scansions, ‘[which] is where the real principles appear’.6 In the case of Gascoigne’s treatise, scansions are few and non-extant, but the author does invite his reader to consider the following two examples in order to illustrate his principles: I understand your meaning by your eye vs. Your meaning I understand by your eye Having detailed (and lamented) how his contemporary English versifiers make use of ‘none other but a foot of two syllables, whereof the first is depressed or made short, and the second is made long’, Gascoigne proceeds to explain: In these two verses [quoted above] there seemeth no difference at all, since the one hath the very selfsame words that the other hath, and yet the latter verse is neither true nor pleasant, and the first verse may pass the musters. The fault of the latter verse is that this word ‘understand’ is therein so placed as the grave accent falleth upon ‘der’, and thereby maketh ‘der’ in this word ‘understand’ to be elevated; which is contrary to the natural pronunciation, for we say ‘ùndèrstánd’, and not ‘ùndérstànd’.7 This is as fine an explication of lexical stress and its relation to metre as any: Gascoigne’s inconsistent use of ‘elevated’ and ‘long’ to designate syllabic prominence clearly corresponds to that indeterminate conglomerate of loudness, pitch and duration which in modern phonetics is covered by the adjective ‘stressed’. Furthermore, Gascoigne’s perceptive comments on the proper pronunciation of a polysyllable such as understand and its permissible positions in a line of metrical verse also reveal that Gascoigne views metre as a delimited and highly fixed pattern of alternating stress-positions, the requirements of which take precedence over the actual wording of any given line. Thus, in the second verse line quoted above it is the grave accent of the metre which causes the wrenched pronunciation of understand, and not the lexical stress of understand which wrenches the

6

, T.V.F. Brogan, English Versification, 1570-1980: A Reference Guide with a Global Appendix (Baltimore: Johns Hopkins University Press, 1981): 143. 7 Gascoigne [1575], in Vickers (ed.) (1999): 165-166.

35

distribution of metrical accents in the line.8 Gascoigne, in other words, displays a rather manifest appreciation of that level of versifying which Roman Jakobson has termed verse design, by which Jakobson understands that underlying structure which ‘determines the invariant features of the verse instances [i.e. individual lines] and sets up the limits of variations’.9 In any useful description of versification the correlative concepts of verse design and verse instance are to be kept separate, not only from one another, but also from the level of delivery instance, which is Jakobson’s term to designate features pertaining to the performance in real time of verse instances.10 Rhythm, in its modern sense, is consequently a concept which is most relevant to the level of delivery design, whereas metre belongs primarily to the level of verse design. We shall have more to say about Jakobson’s influential contributions to metrical theory and poetics in general later; for now it suffices to say that Gascoigne, as early as 1575, adroitly identifies what has since been reified as the most basic principle of English metrics, namely that it is the consistent alternation of prominent and less prominent syllables in fixed numbers, and possibly feet, which makes a given verse instance metrical. The chief – though at the time largely unacknowledged – obstacle which Gascoigne and his contemporaries encountered when they sought to apply the principles of the ancient metrikoi to English versification rested, of course, with the fact that the opposition between long and short syllables in Greek and Latin was, by its very definition, binary: any given syllable was either long or short, with no intermediary level of duration being recognized. Only gradually, even hesitantly, did the Elizabethan metrists who codified the Greco-Latin system’s nomenclature in English realize that such a distinction was not viable in English, and that metrical length would have to be substituted by stress. Stress, however, is not a binary phenomenon in English, but rather a finely graded scale between two extremes, a fact which can easily be demonstrated by examining Gascoigne’s own discussion of the word understand in more detail. For the purpose of formalizing metrical patterns, Gascoigne devises a tripartite system of diacritics, by which marks a ‘grave’ (i.e. stressed) syllable position, marks a ‘light’ (i.e. unstressed) syllable position, whereas marks a syllable position which is ‘indifferent’ with regard to length/stress11 and therefore analogous to the classical concept of anceps. The two former diacritics are furthermore called upon 8

For further discussion of this, see John Thompson, The Founding of English Metre (London: Routledge & Kegan Paul, 1961): 73-74; for a dissenting view, see Alan Holder, Rethinking Meter: A New Approach to the Verse Line (Lewisburg, PA: Bucknell University Press/London: Associated University Presses: 1995): 38-39. 9 ‘Linguistics and Poetics’ [1960], in Roman Jakobson, Language in Literature, edited by Krystyna Pomorska & Stephen Rudy (Cambridge, MA & London: The Belknap Press of Harvard: 1987), 62-94: 78. 10 Jakobson [1960]: 78-79. 11 Vickers (ed.) 1999: 164.

36

to indicate the normative stress patterns of individual words, and Gascoigne clearly perceives that the syllable stand in understand receives the major lexical stress (for which reason he annotates it with ) and that the syllable der receives minimal stress (for which reason he annotates it with ). He refrains, however, from commenting on the phonetic status of the syllable un and tacitly annotates it with , as if it received the same amount of stress as der, which is clearly not the case. The pivotal question, then, which Gascoigne leaves unanswered is why an unstressed syllable such as ùn in ùndèrstánd may occupy a ‘grave’ syllable position in the metre whereas dèr may not. This predicament – to which we shall return more fully in Chapters 4, 5 and 7 – is one of several problems to arise as a consequence of mapping unto the finely graded scale of stress in English the binary metrical system of classical versification. Another such problem is that of the metrical foot and its ontological status in English verse. Gascoigne, as already noted, admits of one foot only in English poetry, namely that ‘of two syllables, whereof the first is depressed or made short, and the second is made long’. This, of course, is the familiar iamb, even though Gascoigne does not name it so in his treatise.12 Gascoigne’s contemporary, George Puttenham, in his Arte of English Poesie (1589), is even less inclined to accord metrical feet any constituent status in the construction of English verse. Having listed twelve permissible ‘auncient feet’ to facilitate the scansion of English verse, Puttenham experiences a sudden change of heart and declares that he thinks his own remarks ‘but vaine & superstitious obseruations nothing at all furthering the pleasant melody of our English meeter’; in fact he rather wishes ‘the continuance of our old maner of Poesie, scanning our verse by sillables rather than by feete’.13 The metrical foot, in other words, is a concept which is embraced rather lukewarmly, if not straight out discarded altogether, by the Renaissance metrists; instead they tend to attach significance to the fixed number of syllables arranged in lines with a consistently alternating stress profile as the most basic criterion of metricality. In 1702, this approach to English metre was taken one step further when Edward Bysshe opened his highly influential Art of English Poetry by brazenly declaring: ‘The Structure of our Verses, whether Blank, or in Rhyme, consists in a certain number of Syllables; and not in Feet

12

The earliest instance of the term iamb to denote the most common foot in English poetry occurs in William Webbe’s Discourse of English Poetrie (1586), reprinted in Joseph Haslewood, Ancient Critical Essays upon English Poets and Poësy, Vol. II (London: Robert Triphook, 1815): ‘A myxt foote of 2. [sic] sillables ... of one short and one long [is] called Iambus’ (67). Curiously, Webbe gives as an example of this foot in English the word dying, which seems to imply that Webbe’s notion of syllabic prominence has nothing to do with stress and everything to do with the classical rules of quantity based on the number of consonants that follow the vowels. 13 George Puttenham, The Arte of English Poesie [1589] (Menston: Scolar Press, 1968): 107.

37

compos’d of long and short Syllables’.14 The fact that Bysshe feels compelled to reiterate what Gascoigne had already worked out more than a century earlier – that syllable duration is not a viable principle for composing verse in English – bears witness to the obstinacy with which quantitative theories about English metrics continued to be formulated during the seventeenth century.15 Bysshe’s treatise is a rebuttal against such reactionary theories in much the same way as Samuel Daniel’s A Defence of Rhyme (1603) had been a century earlier. But whereas Daniel’s treatise was a response to Thomas Campion’s Observations in the Art of English Poesie (1602), in which a return to strictly quantitative principles of versification was advocated, Bysshe’s main opponent was Charles Gildon, one of the many writers to be mocked by Alexander Pope in The Dunciad. In the first volume of Gildon’s Complete Art of Poetry (1718) feet based on quantity are singled out as the basic structuring unit of English versification, while stress is explained away simply as a feature pertaining to the level of performance: in Gildon’s view, proponents of stressbased metrics are guilty of conflating verse design and delivery design. This is a common enough flaw in writings on metrics,16 although in the case of Gildon’s attacks on his contemporaries the accusation is hardly justifiable. Gildon’s use of notes on musical staffs for the purpose of scansion reveals his own conception of metre to be congenial with that of the ancient rhythmikoi, according to whom, it will be remembered, the notion of quantity was derived from the actual time-values of musical performance. In employing musical transcription as a means for poetic scansion, Gildon can be seen as a precursor to that upsurge of musical and temporal approaches to metre which found its earliest English codification in Joshua Steele’s Prosodia Rationalis (1779). However, our focus here is limited to the mainstream view of English metrics as it unfolded during the eighteenth and nineteenth centuries. To those who have taken at face value the contention of certain modern introductions to English poetry that scansion by foot is ultimately a remnant from the classicism of Renaissance metrists,17 it will probably come as a surprise to learn the extent to which isosyllabism, and not feet, was promulgated as the basic principle of English versification, at least up until the late eighteenth century. While it is true that we owe the adaptation of Greco-Roman nomenclature for the purpose of metrical description to the Renaissance metrists, there is simply no evidence to support the claim 14

Edward Bysshe, The Art of English Poetry [1702] (Menston: Scolar Press, 1968): 1. On this and the experimental poetry created in this tradition, see Derek Attridge, Well-Weighed Syllables: Elizabethan Verse in Classical Metres (Cambridge: Cambridge University Press, 1974). 16 Hollander goes so far as to suggest that ‘Performative systems of scansion, designed as descriptive ones, have composed all but a few of the metrical studies of the past’ (1956: 239). 17 E.g. Baron Wormser & David Capella, Teaching the Art of Poetry: The Moves (Mahwah, NJ: Lawrence Erlbaum Associates: 2000): 4. 15

38

that these theorists and their immediate successors thought of English verse as consisting first and foremost of feet based on the distinction between stressed and unstressed syllables. This is true especially of that century whose poetry we have nowadays come to think of as neoclassical: from a metrical point of view, English verse of the eighteenth century was far more influenced by Romance remouldings of classical models than by classical metrics itself. In no small part is this due to the unprecedented success which Bysshe’s Art of English Poetry came to enjoy in the decades after its publication: from 1702 to 1762 it passed through as many as nine consecutive editions, and the ‘Rules for Making Verses’, which is the section dealing with metrics, was reprinted as late as 1877 in Tom Hood’s Practical Guide to English Versification.18 The popular appeal of Bysshe’s treatise stands in stark contrast to the obscurity which shrouds its author’s identity: he may have been an ancestor of Percy Bysshe Shelley, but this has never been ascertained.19 It has, however, been established that Edward Bysshe must have been intimately familiar with the Jansenist grammarian Claude Lancelot’s Quatre Traitez de Poësies, Latine, Françoise, Italienne, et Espagnole (1663), as even Charles Gildon appears to have been well aware of: ‘[Bysshe’s indigested Notions are] mostly borrow’d from the Messrs. of the Port-Royal on the French Versification’.20 A. Dwight Culler lends credibility to Gildon’s imputation by comparing Bysshe’s already quoted contention that ‘The Structure of our Verses ... consists in a certain number of Syllables; and not in Feet compos’d of long and short Syllables’ to this statement out of Lancelot’s treatise: ‘La structure ne consiste qu’en vn certain nombre de syllabes, & non pas en pieds composez de syllabes longue & breves’.21 On this point and others, Bysshe’s treatise on English metrics appears to be little more than a verbatim translation of Lancelot’s work, specifically that part of it which can be found under the heading of ‘Breve Instruction sur les Regles de la Poësie Françoise’. This was neither the first nor the last time that the tenets of French versification were called upon to provide inspiration for the composition of English poetry: in Chapter 3 we shall examine the role played by French (and Italian) verse in Chaucer’s founding contribution to English verse, and in Chapter 6 we shall discuss the impact of the French symbolists and vers libristes on the development of free verse in the Anglo-American world. However, whereas both Chaucer and the 18

Bysshe’s treatise was reissued in 1705, 1708, 1710, 1714, 1718, 1724, 1737 and 1762; see A. Dwight Culler, ‘Edward Bysshe and the Poet’s Handbook’, in PMLA, Vol. 63, No. 3 (Sep. 1948), 858-885: 861, from whence also stems the reference to Thomas Hood’s reprinting of the ‘Rules’ in 1877. 19 Culler (1948): 860. 20 Charles Gildon, The Complete Art of Poetry, Vol. I (London: 1718): 93. 21 Culler (1948): 877.

39

proponents of free verse looked to France merely for inspiration and occasionally models, Bysshe takes over the most deeply embedded prosodic principle of French – namely that of syllabism – and applies it indiscriminately as a normative matrix for composing verse in English.22 The list of contemporary and later writers to have sneered at Bysshe for his lack of sophistication in performing this crude manoeuvre is almost as long as the list of eighteenth-century writers who, perhaps inadvertently, adopted and propagated Bysshe’s conclusions. A. Dwight Culler counts among the latter such prominent figures as Alexander Pope, Isaac Watts, Samuel Johnson and Thomas Gray as well as a host of lesser writers.23 He also provides the following apt characterization of the dubious regard in which Bysshe’s treatise ostensibly was held, even by its owners (among whom we find both Samuel Johnson and William Blake): ‘Bysshe [was] the sort of book one consults surreptitiously and keeps locked in a drawer when not in use’. 24 Over the course of the eighteenth century, the strict syllabism of Bysshe was occasionally paired with some rather feeble attempts to carve out a place for the foot in English metrics, but these endeavours were largely confined to technical treatments of the subject, characterized by a persistent tendency to confuse quantity and stress.25 Outside the narrow circle of theoretical metrics, however, Bysshe’s syllabist tenets reigned supreme, at least throughout the first three quarters of the century. ❦ It is against the backdrop of this prevailing view of English poetry as basically syllabic that we must read Samuel Taylor Coleridge’s remarks on his chosen metre for Christabel (1816) in the preface to the first edition of that poem: I have only to add that the metre of Christabel is not, properly speaking, irregular, though it may seem so from its being founded on a new principle: namely, that of counting in each

22

For a dissenting and closely argued view on the merits of Bysshe’s contribution to English metrics, see Peter L. Groves, ‘The Chomsky of Grub Street: Edward Bysshe and the Triumph of Classroom Metrics’, in Versification [an electronic journal of literary prosody], Vol. 3 (1999). 23 Culler (1948): 881. 24 Culler (1948): 864. 25 See, for example, the two essays appended to Samuel Say’s Poems on Several Occasions (1745), John Mason’s Two Essays on the Power of Numbers (1761) and Daniel Webb’s Observations on the Correspondence between Poetry and Music (1769).

40

line the accents, not the syllables. Though the latter may vary from seven to twelve, yet in each line the accents will be found to be only four.26 This claim – that regulating the number of accents rather than the number of syllables should constitute a ‘new principle’ – has been a source of consternation to more than one subsequent metrist. However, this consternation is largely the result of the ways in which earlier English poetry, in particular that of the seventeenth and eighteenth centuries, came to be reinterpreted during the course of the nineteenth century. The theoretical foundation of this process of reinterpretation may well be said to have been laid by Thomas Sheridan in his Lectures on the Art of Reading (1775). In the second volume of that work, Sheridan criticizes the prevailing tendency to employ French syllabism as a model for composing verse in English: Thus, because the French measured their verses by the number of syllables which composed them, on account of a defect in their tongue ... we did the same; and in consequence of this, our English heroic line was said to consist of ten syllables27 He also criticizes the way in which the concept of metrical feet had hitherto been approached by earlier English metrists. Sheridan thus perceives with a reasonable amount of clarity that the common fault with all previous attempts to introduce feet into English metrics was that they tended to treat English syllables as if they possessed classical quantity: The chief source of [earlier English metrists’] errors lay in ... considering the English poetic feet as exactly the same with the Roman, and treating them as such, when in reality there is a material difference between them; for the Latin poetic feet are formed by quantity, the English by accent.28 At the same time, Sheridan is unequivocal when it comes to his own view of English metrics: he blatantly declares that it is the fixed number of feet rather than syllables which structures English 26

Samuel Taylor Coleridge, The Collected Works, Vol. 16, Poetical Works, edited by J.C.C. Mays (Princeton, NJ: Princeton University Press, 2001): 482-483. 27 Thomas Sheridan, Lectures on the Art of Reading; Second Part: Containing The Art of Reading Verse (London: J. Dodsley et al.: 1775): 8-9. 28 Sheridan (1775): 5.

41

metrical verse, and he cites the following catalogue of feet as being relevant to any accomplished versifier working in English: trochee (‒ᴗ), iamb (ᴗ‒), spondee (‒ ‒), pyrrhic (ᴗᴗ), dactyl (‒ᴗᴗ), amphibrach (ᴗ‒ᴗ), anapaest (ᴗᴗ‒) and tribrach (ᴗᴗᴗ).29 Crucially – and in spite of the use of macrons and breves to indicate the metrical profile of each foot – these eight feet are defined by Sheridan exclusively in terms of stress vs. non-stress;30 in fact, Sheridan is probably the earliest writer to use the word stress consistently to designate the metrically salient phonological feature of English.31 Like Gascoigne exactly two hundred years earlier, Sheridan laments that English poets tend to allow of only the iamb in their verse, and calls out for a reanalysis of verse instances which challenge the isosyllabic structure of the English heroic line. As an example of such an aberrant verse line, Sheridan quotes the following, which had also been used by Samuel Say in one of his essays on metrics: And the shrill sounds ran echoing through the wood.32 Of this line, Sheridan asserts that proponents of syllabist metrics would have the word ‘echoing’ made into ‘ech’ing’ in order to keep the syllable count, to which he adds: ‘Can anything be more absurd than to omit a vowel in the writing, which cannot be omitted in the utterance?’33 Here, we must object that the answer to Sheridan’s rhetorical question is not quite as straightforward as he makes it out to be: in the score of Henry Purcell’s aria ‘Hark! The Ech’ing Air’ from The FairyQueen (1692), the word echoing is thus consistently rendered as ‘ech’ing’, and it furthermore occupies only two instead of three quavers in the score.34 The cocksure tone of Sheridan’s rhetorical question instantiates all the perils of conflating verse instance and delivery instance, but at the same time it must be admitted that this gaffe on the part of Sheridan is an understandable one since his quibble is with metrists who apotheosize the syllable as the metrical unit par excellence. 29

Sheridan (1775): 37. Sheridan does, in fact, perceive ‘quantity’ as a contributing factor in the recitation of English verse, but he wisely allocates this feature exclusively to the level of delivery and grants it no constituent value at the level of verse design; see Sheridan (1775): 29. 31 E.g. Sheridan (1775): 14; 35. The OED cites John Mason’s Essay on the Power and Harmony of Prosaic Numbers (London: 1749) as the earliest attestation of stress in the sense of ‘Relative loudness or force of vocal utterance’, but a closer look at Mason’s text reveals that the author employs the term in a manner which has little to do with the principles of English versification: ‘in Pronunciation they [the Greeks] laid the Stress or Force of their Voice on the long Syllables though they were not accented’ (Mason (1749): 24). 32 Sheridan (1775): 3. 33 Sheridan (1775): 3. 34 Henry Purcell, The Fairy Queen in Full Score (Toronto: Dover Publications, 2000): 179. 30

42

Instead of the typographical convention – so common in seventeenth and eighteenth century English verse – of employing apostrophes to indicate syncope, Sheridan argues in favour of the ‘full’, meticulous pronunciation of every syllable of every word in a line of verse. 35 Given his involvement with the elocution movement, which swept the Anglophone world in the late eighteenth century, this stance of Sheridan’s is hardly surprising, although it did open the gateway for a host of dubious scansions on his part. Take, for example, Sheridan’s scansion of Milton’s line O’er many a frozen, many a fiery Alp which is analyzed as consisting of thirteen syllables segmented into two iambs and three amphibrachs: │.

/ .│.

/│.

/ .│. /.│. / │

O’er many a frozen, many a fiery Alp36 Curiously, Sheridan does not object to treating ‘O’er’ as a monosyllable, though he does insist on counting ‘fiery’ as a trisyllable, just as he discards the possibility of synaeresis in the phrase ‘many a’. But even if we accept Sheridan’s syllable count and stress markings as valid, we are still left with the problem that Milton’s line could just as easily, and perhaps more elegantly, be scanned as consisting of two iambs and three anapaests: │.

/│ . .

/│.

/ │. . /│.. / │

O’er many a frozen, many a fiery Alp Neither of these scansions are, however, particularly useful beyond their ability to expose the fundamental arbitrariness of Sheridan’s system of scansion. They are both attempts at reinterpreting a predominantly syllabic verse design in terms of feet for the sole purpose of accommodating an idiosyncratic notion of orthoepy, which is at best relevant only at the level of delivery. And yet Sheridan’s observations on English metrics came to exert a considerable influence throughout the nineteenth century and even beyond. His chosen catalogue of eight feet soon became established to the point of orthodoxy among metrists and grammarians; we thus find the exact same 35 36

Sheridan (1775): 47-48. Sheridan (1775): 40.

43

inventory in John Carey’s Key to Practical English Prosody and Versification (1809) and in Alexandre Spiers’s Study of English Poetry (1835) as well as in the numerous editions of Lindley Murray’s English Grammar (1795) that were published in both Britain and America during the early decades of the nineteenth century. Disseminated via such channels, the stress-based metrical foot became reified as the most basic structural unit in English versification and then applied retrogressively to the works of poets who in all likelihood would have been perplexed to learn that amphibrachs and other such exotic creatures were lurking in their verses.37 Some later metrists saw fit to expand Sheridan’s list of permissible feet in English, culminating in Saintsbury’s inclusion of as many as 21 distinct species of foot, while others wisely rid themselves of the amphibrach and the tribrach and settled for an inventory of six feet: iamb, trochee, pyrrhic, spondee, anapaest and dactyl. This inventory is still presented as a helpful toolbox for scanning English verse in many modern introductions to that subject,38 but its potential usefulness and sheer familiarity should not, Derek Attridge admonishes us, be allowed to ‘confer on it any unwarranted authority’.39 Attridge, in fact, dates the rise of foot-based metrics in English significantly later than what has been proposed here, arguing that ‘the main tradition in prosodic theory up until the end of the nineteenth century was based on syllables and accents rather than feet, and only with the new interest in Greece and Rome in the nineteenth century did foot-scansion come into its own as a mode of analysis’.40 While it is certainly true that foot-scansion gained momentum only slowly and did not culminate until the second half of the nineteenth century,41 Attridge’s sweeping statement is in need of further qualification in a survey such as ours. We have already seen how Sheridan’s system of scansion by stress-based feet was propagated in major handbooks on metrics already from the beginning of the nineteenth century, and as for Attridge’s contention that it was the ‘new interest in Greece and Rome’ which paved the way for this approach to English metrics, it should be noted that what was decidedly new during the period with which we are concerned was the profound interest accorded to Greece; Rome, on the other hand, had held an unrivalled position as the centre of civilisation in the minds of English writers at least since the Renaissance. But in the course of the eighteenth century new methods of textual criticism began to emerge, primarily in Germany where 37

See Culler (1948): 883-885. E.g. John Lennard, The Poetry Handbook (Oxford: Oxford University Press, 1996), John Strachan & Richard Terry, Poetry (Edinburgh: Edinburgh University Press, 2000) and Rhian Williams, The Poetry Toolkit (New York: Continuum, 2009). The two latter also include the amphibrach, and Williams even lists the amphimacer, among those feet that can be helpful in the scansion of English verse. 39 Derek Attridge, The Rhythms of English Poetry (New York: Longman, 1982): 5. 40 Attridge (1982): 5. 41 On this, see also T.V.F. Brogan’s comments cited in Holder (1995): 246, n. 5. 38

44

scholars such as Johann Bengel, Johann Michaelis and Johann Griesbach approached the Bible, not as a revelatory monolith, but rather as a tapestry of disparate textual documents produced in specific socio-historical contexts. In Britain, however, it was first and foremost the work of Homer that was subjected to this type of historicizing scrutiny. Thomas Blackwell in his Enquiry into the Life and Writings of Homer (1735) presented Homer as a ‘Stroling Bard’42 whose work was crucially dependent upon the geographical and social environment in which it had been produced,43 and in 1775 – the same year that Sheridan published his Lectures on the Art of Reading – Robert Wood took Blackwell’s observations to their natural conclusion when he asserted that alphabetical writing had not yet been invented in Homer’s Greece,44 for which reason Wood surmised that the form in which Homer’s epics have been handed down to posterity must have been the result of later editors’ endeavours.45 This novel view of Homer as a decidedly elusive author figure is in glaring contrast to former generations’ conception of the Greek bard. Alexander Pope, in the preface to his translation of the Iliad (1715), compares Homer to Virgil and finds that ‘Homer was the greater Genius, Virgil the better Artist. In one we most admire the Man, in the other the Work’.46 What the subsequent advances in the discipline of textual criticism made apparent was that, in the case of Homer, there was in fact no man to admire, but only the work to which his name had been appended.47 We should also pay due attention to Pope’s use of the term genius to characterize Homer. Somewhat surprisingly, the OED gives as its earliest attestation of the use of genius in the sense of ‘Native intellectual power of an exalted type, such as is attributed to those who are esteemed greatest in any department of art, speculation, or practice; instinctive and extraordinary capacity for imaginative creation, original thought, invention, or discovery’ a quotation from Henry Fielding’s 42

Thomas Blackwell, Enquiry into the Life and Writings of Homer (London: 1735): 104. On this, see Donald M. Foerster, Homer in English Criticism. The Historical Approach in the Eighteenth Century (Hamden, CT: Archon Books, 1969): 26-40. 44 Robert Wood, An Essay upon the Original Genius of and Writings of Homer (London: 1775): 239. 45 A similar argument was, of course, put forward with considerably more scholarly rigour some twenty years later in Friedrich August Wolf’s Prolegomena ad Homerum (1795). 46 Alexander Pope, The Iliad of Homer (London: 1715): Preface (unpaginated). 47 Curiously, this demotion of Homer’s role in the actual production of those epics which are still attributed to his name led to a marked increase on a pan-European scale in the interest in matters Homeric. The increased interest in Homer can be seen in the sheer number of English translations of the Iliad and the Odyssey that were produced from the late eighteenth century to the turn of the twentieth. From 1790 to 1900, at least forty different translations of the Iliad were published, and the same period saw more than twenty new translations of the Odyssey. By comparison, less than twenty English translations of Homer’s two epics saw the light of day in the period between 1600 and 1789. This increase in the number of published translations of Homer cannot be attributed only to the advancements in the printing industry during the eighteenth century. Thus, if we look at the number of English translations of the Aeneid between 1600 and 1789, we find that about a dozen of such works were published in that period, whereas the number of new translations of the Aeneid between 1790 and 1900 did not exceed twenty. Please note, however, that only complete translations of the three relevant epics are included in this survey. 43

45

Tom Jones (1749) and furthermore notes that this sense is not recorded by Johnson in his dictionary.48 However, this is clearly the sense in which Pope employs the term to characterize Homer’s merits over Virgil’s, and it is even possible to find earlier instances of this sense being used by major English writers to describe the unique capacity of poets. Thus, in an issue of the Spectator from September 1711, Joseph Addison dedicates an entire essay to the concept of literary genius and begins by declaring: There is no Character more frequently given to a Writer, than that of being a Genius. I have heard many a little Sonneteer called a fine Genius. There is not a Heroick Scribbler in the nation, that has not his Admirers who think him a great Genius; and as for your Smatterers in tragedy, there is scarce a Man among them who is not cried up by one or other for a prodigious Genius.49 Addison’s diatribe reminds us that the so-called ‘cult of genius’ is by no means an exclusively, or even predominantly, Romantic trait. In fact, the cult was already in full flower in Britain at the beginning of the eighteenth century, from which point onwards it became increasingly commonplace to pronounce Homer as the prime exemplar of poetic genius, usually to the detriment of Virgil. The appropriateness of this epithet to Homer was, of course, reaffirmed in the title of Robert Wood’s Essay upon the Original Genius of and Writings of Homer (1775), which in turn echoes William Duff’s Essay on Original Genius (1767) and presumptively also Edward Young’s influential Conjectures on Original Composition (1759). It is in the latter tract’s discussion of Pope’s translation of the Iliad into heroic couplets that Young offers the following characterization of Homer’s inimitable metrical qualities: What a fall is it from Homer’s numbers, free as air, lofty, and harmonious as the spheres, into childish shackles and tinkling sounds!50 We should be remiss if, in a survey such as this, which seeks to trace the roots of free verse, we failed to point out Young’s use of the simile ‘free as air’ to characterize Homer’s metrics, but in our eagerness to do so, we should not overlook the severity of Young’s criticism of the prevailing mode 48

OED, s.v. ‘genius’, sense 5. The Spectator, No. 160 (September 3, 1711). 50 Edward Young, Conjectures on Original Composition (London: 1759): 58. 49

46

of English versification against which Homer is pitted. Young’s phrase ‘childish shackles and tinkling sounds’ echoes, of course, Milton’s denunciation of ‘the jingling sounds of like endings’ in the preface to the second edition of Paradise Lost, but whereas Milton retained ‘apt Numbers [and] fit quantity of Syllables’ as the cornerstone of any pleasing metrics, Young and his fellow commentators on poetic genius were willing to go further. Duff, for instance, declares that ‘Smooth versification and harmonious numbers will no more make genuine Poetry, than the atoms of a skeleton put together can make an animated and living figure’,51 and statements such as this bear witness to the slow but steady process by which the eighteenth century’s elevation of originality and invention over imitation and artistry was eventually extended to the province of metre. ❦ Sentiments such as those expressed in the quotations from Edward Young and William Duff above should obviously not be interpreted as arguments in favour of free verse. While it is true that the preoccupation with textual aspects of Homer’s work during the eighteenth century did produce at least one early specimen of what might be called proto-free verse in English – namely James Macpherson’s Ossian poems52 – the more significant outcome of these advancements was surely the gradual turn away from Latinate literature as the most apt model for English versification which followed in their wake. The turn away from Latinate models for English versification was twofold: just as Greece came to be favoured over Rome as the more original culture of which Latin was merely an imitation, if not straight out a decadent corruption, so the influence from Romance culture – especially from French critics such as Nicolas Boileau-Despréaux – came to be sidelined in favour of a renewed interest in the Germanic stratum of English culture, language and history. The former development finds its most complete expression in that movement which we nowadays refer to as Romantic Hellenism, but also makes itself felt in the gradual elevation of Homer over Virgil, which we have already commented on in some detail. As for the latter development – the turn away from France towards Germany – a few apt quotes may serve to illustrate the basic tenets of this process. We have already cited Sheridan’s somewhat crude contention that the syllabism of French poets’ versification was the result of ‘a defect in their tongue’, but a similar concern was voiced by 51

William Duff, Essay on Original Genius (London: 1767): 125. On which, see Chapter 6 and Kristine Louise Haugen, ‘Ossian and the Invention of Textual History’, in Journal of the History of Ideas, Vol. 59, No. 2 (Apr., 1998), 309-327. 52

47

Richard Hurd, the bishop of Worcester, who in 1767 had described the French language as ‘simple, clear, exact, that is, fit for business and conversation; but for that reason, beside it’s [sic] total want of numbers, absolutely unsuited to the genius of the greater poetry’.53 Ernst Robert Curtius points out that Hurd’s attack on French language and culture coincides almost exactly with Gotthold Ephraim Lessing’s critique of the influence of France on the German stage,54 over which Lessing championed the works of William Shakespeare. As a consequence of this, Lessing adopted for Nathan der Weise (1779) that verse design, whose development in English forms the topic of this dissertation’s Chapters 4 and 5, namely blank verse.55 Lessing’s reverence for the English language and its literature was not unrequited: among his earliest translators we find the Norwich-based scholar and polyglot William Taylor, who translated Nathan der Weise to English in the early 1790s. Hailed by his protégé George Borrow as ‘the father of Anglo-Germanism’,56 Taylor played a key role in promoting contemporary German literature to the British reading public at a time when – in Edward Dowden’s memorable phrasing – ‘German characters were as undecipherable to most Englishmen as Assyrian arrow-heads’.57 Thanks largely to Taylor’s efforts, a host of English translations of German writers, most prominent among them Friedrich Schiller and Johann Wolfgang Goethe, appeared in the final decade of the eighteenth century. These poets, and the Sturm und Drang movement with which their names are associated, in turn owed much to Young’s Conjectures on Original Composition, a detail which illustrates the reciprocal nature of the relation between German and English culture during these decades. It was in this Anglo-Germanic exchange of letters during the late eighteenth century that an old metrical acquaintance of ours made a surprising comeback. As we saw in the preceding chapter, the dactylic hexameter had suffered a particularly damaging setback when the distinction between long and short syllables had been lost in the classical languages, for which reason its subsequent usage was largely restricted to somewhat obscure venues. However, in 1748 the dactylic hexameter was resuscitated in German by Friedrich Gottlieb Klopstock when the first three Cantos of Der Messias were published anonymously. In Klopstock’s German recasting of the dactylic hexameter, however, the long syllable positions of the classical metre were consistently filled with stressed 53

Richard Hurd, Letters on Chivalry and Romance (London: 1767): 84. See Ernst Robert Curtius, European Literature in the Latin Middle Ages, translated by Willard R. Trask (London & Henley: Routledge & Kegan, 1953): 596. 55 Lessing was, however, not the first German writer to make use of blank verse based on English models. As early as 1758, Christoph Martin Wieland had employed blank verse for his tragedy Lady Johanna Gray, the title of which flaunts the play’s Anglophilia. 56 George Borrow, The Works of George Borrow, edited by Clement Shorter, Vol. VI (London & New York: Constable & Co. & Gabriel Wells, 1923-24): 220. 57 Cited in Robert Alfred John Walling, George Borrow, The Man And His Work (London: Cassell, 1908): 37. 54

48

syllables, in effect producing an accentual variant of the quantitative hexameter.58 This deceptively simple artifice encouraged the crafting of anisosyllabic lines at a time when isosyllabism reigned all but supreme, and it proved to be an unexpectedly versatile device, not only in Germany where the technique was copied by Goethe (Hermann und Dorothea), Schiller (Der Spaziergang) and Friedrich Hölderlin (‘Kanton Schweiz’) to name but a few, but also in Britain where it was William Taylor who first caught wind of the new trend. In the very first volume of the Monthly Magazine (1796), we thus come across a contribution by Taylor under the heading of ‘English Hexameter Exemplified’, in which nineteen lines from Macpherson’s Ossian are subjected to what Taylor calls ‘transversion’ into meticulously wrought accentual hexameters.59 A short extract of Taylor’s transversion may be in place: │/ . .│ / . .│ / . .│ / / │ / . .│/ . │ Oaks of the mountains decay, and the hard oak crumbles asunder; │/..│ / . .│ / / │ / . .│ / . .│ / . │ Ocean shrinks and again grows; lost is the moon from the heavens; │ / / │/ . .│ / . .│ / . .│ / . .│ / . │ Whilst thou ever remainest the same to rejoice in thy brightness.

As can be seen, Taylor adheres quite diligently to the basic compositional principles of his classical model: whereas spondees are allowed to roam freely in the first four feet of each line, the fifth foot is invariably a dactyl, just as the sixth foot is consistently a trochee, in keeping with the principle of anceps in that position.60 This is not the place to pass judgement on the artistic merits of Taylor’s experiment; for now, it suffices to say that Taylor’s dabbling with the accentual hexameter became the starting point for a parade of English poems in accentual versions of classical metres. Coleridge (‘Hymn to the Earth’, 1799) and Robert Southey (A Vision of Judgment, 1821) were at the forefront of this movement, but throughout the nineteenth century works based on similar principles continued to be 58

On Klopstock’s contribution to the development of the accentual hexameter and its ramifications in the Anglophone world, see Ernest Bernhardt-Kabisch, ‘“When Klopstock England Defied”: Coleridge, Southey, and the German/English Hexameter’, in Comparative Literature, Vol. 55, No. 2 (Spring, 2003), 130-163: 131-136. 59 Monthly Magazine, Vol. I (London: 1796), 404-405. 60 Taylor is, in fact, a good deal more faithful to the model of the classical hexameter than many other imitators of that form were: Klopstock frequently champions trochees to the exclusion of spondees, also outside of the final anceps position (see Bernhardt-Kabisch (2003): 135-136). In this Klopstock was followed by the majority of his English imitators, though not by Taylor in these lines.

49

composed in English, making accentual imitations of quantitative metres a staple, not only of Romanticism, but especially of post-Romantic verse on both sides of the Atlantic.61 Henry Wadsworth Longfellow’s Evangeline (1847), Arthur Hugh Clough’s Tober-na-Vuolich (1848) and Charles Kingsley’s Andromeda (1858) belong to the more prominent examples of this trend. In spite of the marginal position that these mongrel products of German and Greek influences on English verse hold in today’s literary canon, their historical significance should not be underestimated. Their very existence is indicative of an ideological process, initiated in the eighteenth century but brought to full flower during the course of the nineteenth, by which stress gradually came to usurp syllabic phrasalism as the most widely acknowledged constituent of English versification. After all, the very idea of consistently replacing long syllables with stressed ones within the framework of a sophisticated foot-based metre such as the Greek hexameter presupposes an unequivocal appreciation of both the foot and of linguistic stress as valid prosodic concepts in the Germanic languages. This metrical coup d’etát did not, however, assume any single form. In addition to the mushrooming of accentual imitations of quantitative verse during the nineteenth century, the same period saw, under the influence of Thomas Percy’s Reliques of Ancient English Poetry (1765), an increased interest in both folk song and purely accentual modes of versification. In all of these metrical modes, the criterion of observing a strict syllable-count in each line is loosened, if not altogether abandoned, in favour of keeping a fixed number of stressed syllables, and we find examples of all of these trends in, for instance, the works of Coleridge. His use of the ballad stanza in The Rime of the Ancyent Marinere frequently deploys syllabic license, which clearly sets apart the form of this poem from the tradition of the hymnal common measure,62 and in Christabel – as already mentioned – we find a mode of versification which, at least according to the author’s own account, is most accurately described as purely accentual.63 61

See Yopie Prins, ’Nineteenth-Century Homers and the Hexameter Mania’, in Sandra Bermann & Michael Wood (eds), Nation, Language, and the Ethics of Translation (Princeton, NJ: Princeton University Press, 2005), 229-256. 62 Consider, for instance, this stanza (ll. 263-266) – The móving Móon went úp the ský And nó where díd abíde: Sóftly shé was góing úp And a stár or twó besíde— – in which the two first lines adhere strictly to the expected pattern of common measure in terms of both syllable count and stress profile. The last two lines, on the other hand, both diverge from the expected syllable count of that measure while retaining the expected number of stressed syllables in each line. 63 It has, of course, been established by several commentators that Christabel contains numerous lines which do not contain exactly four strongly stressed syllables, but judging from the wording of Coleridge’s own preface to that poem, this compositional principle was clearly central to the poem’s metrical form.

50

Hailed by Eliot as ‘perhaps the greatest of English critics, and in a sense the last’,64 Coleridge epitomizes the full fruition of Greco-Germanic influences on nineteenth-century English verse, and there is ample evidence to be found for the validity of this claim in his idiosyncratic autobiography Biographia Literaria (1817). It is in that work that Coleridge’s memorable recollection of his meeting with Klopstock is recounted, but before that meeting took place, Coleridge had already immersed himself in Germanic languages and literature during a four-month stay at the University of Göttingen. Founded in 1734 by King George II, the University of Göttingen was at the time of Coleridge’s matriculation in February 1799 not only the youngest of the German universities, but also a beacon of the New Humanist approach to the classics,65 whose aim it was ‘to assimilate the substance, to form the mind and to cultivate the taste and lead up to the production of modern literature that was not to be a mere echo of a bygone age, but was to have a voice of its own’, according to J.E. Sandys.66 At the centre of this movement was the director of the university library, Christian Gottlob Heyne, who granted Coleridge unlimited access to the library upon his arrival. This allowed Coleridge to study a compendious selection of early German literature which, according to his own account in the Biographia Literaria, included ‘Ottfried’s metrical paraphrase of the gospel’, ‘the Minnesinger ... and the metrical romances’ as well as ‘sufficient specimens of the master singers’.67 His guide and instructor in these matters was one Thomas Tychsen, a professor of Theology, who in addition to teaching Gothic to Coleridge was also an expert in oriental literature.68 This last biographical detail of Coleridge’s activities in Göttingen is not only suggestive of the level of criss-crossing and amalgamation which characterized the New Humanist approach to the classic disciplines of learning as a whole; it also reminds us of the role which the East played in forming the new literary sensibility at that time. 64

T.S. Eliot, ‘The Perfect Critic’, in The Sacred Wood (London: Methuen, 1920): 1. ‘New Humanism’ is the term used by J.E. Sandys to translate the German pedagogue Friedrich Paulsen’s coinage Neuhumanismus and should not be conflated with Irving Babbit’s use of the same term at the beginning of the twentieth century. 66 J.E. Sandys, A History of Classical Scholarship, Vol. III [1908] (New York & London: Hafner Publishing Company, 1967): 7. Sandys further notes that the ‘interest in Homer is a note of the New Humanism. Thus far the Odyssey and the Iliad had only once been rendered in German, in 1537 and 1610 respectively. But the middle of the eighteenth century was marked by two translations of the early books of the Iliad, followed in 1754 by the illustrated translation that was Goethe’s first introduction to Homer. The text was edited by Ernesti in 1759-64. This was followed by five new translations, culminating in that completed by Voss in 1793, which was immediately succeeded by the edition of Wolf, with its memorable Prolegomena (1794-5), and by the edition of Heyne (1802 f )’ (8). 67 Samuel Taylor Coleridge, The Collected Works, Vol. 7, Biographia Literaria, edited by James Engell & W. Jackson Bate (Princeton, NJ: Princeton University Press, 1983): 208-209. For a more detailed account of Coleridge’s reading habits in Göttingen – and of his plans to write in detail about Lessing – see Maximiliaan van Woudenberg, ‘Coleridge’s Reading Notes in Göttingen: The Intertextual Research of the Projected Life of Lessing in 1799’ in Dirk Van Hulle & Wim Van Mierlo (eds), Variants 2/3: Reading Notes (Amsterdam & New York: Editions Rodopi, 2004), 245-270. 68 James Holly Hanford, ‘Coleridge as a Philologian’, in Modern Philology, Vol. 16, No. 12 (Apr., 1919), 615-636: 619. 65

51

This interest in the oriental world is also echoed in William Jones’s famous lecture to the Asiatick Society in February 1786: The Sanscrit language, whatever be its antiquity, is of a wonderful structure; more perfect than the Greek, more copious than the Latin, and more exquisitely refined than either, yet bearing to both of them a stronger affinity, both in the roots of verbs and the forms of grammar, than could possibly have been produced by accident; so strong indeed, that no philologer could examine them all three, without believing them to have sprung from some common source, which, perhaps, no longer exists; there is a similar reason, though not quite so forcible, for supposing that both the Gothic and the Celtic, though blended with a very different idiom, had the same origin with the Sanscrit; and the old Persian might be added to the same family.69 Published in 1788, Jones’s treatise was neither the first, nor the most cogent argument in favour of a common origin for languages as far apart as Celtic and Sanskrit. Similar notions had been put forward with more scholarly rigour by Joseph Scaliger in the sixteenth century and by Marcus Zuerius van Boxhorn in the seventeenth.70 Jones’s observations, however, found a far more receptive audience at the brink of the nineteenth century than had those of his predecessors in their time. Shaped as it was by the tenets of the New Humanism, the scholarly community in the late eighteenth century welcomed the implications of Jones’s theory and found in it confirmation of what was already suspected: that the world was at once bigger and smaller than had hitherto been assumed. A common origin apparently bound together the modern vernaculars, not only with the languages of the classical world, but also with languages even further removed. As a result of this realization the nineteenth century came to be, among other things, the century of philology. As far as scholarly disciplines go, comparative philology with its penchant for detailed diachronic studies may well lay claim to being the most schizophrenic of all. On the one hand, philology presupposes a cosmopolitan belief in the fundamental relatedness of the most disparate and apparently localized of phenomena, while on the other hand it tends to stipulate for these localized phenomena a privileged status as far as explanatory power is concerned. By this logic, 69

Cited in Benjamin W. Fortson IV, Indo-European Language and Culture: An Introduction, Second edition (London: Wiley-Blackwell, 2010): § 1.14. 70 In fact, Jones cannot even lay claim to being the first to hypothesize a connection between European languages and Sanskrit: in this he was preceded by Filippo Sassetti’s observations in the late sixteenth century on the similarities between a number of words in Sanskrit and Italian.

52

whatever is remote, exotic, local and isolated comes to serve as the paradigm by which the universal, the classical and the communal must be understood. This paradoxical nature of philology as a scholarly discipline is reflected in its own mixed ideological lineage: on the one hand, philology was a child of the New Humanism’s faith in the enduring value of the heritage from Greece and Rome and its continuous ability to forge links between the ancient and the modern world. On the other hand philology was equally a child of that intellectual current which Isaiah Berlin has fittingly termed the Counter-Enlightenment.71 Perhaps the single most important figure of the Counter-Enlightenment was Johann Gottfried Herder, whose ideas on language as rooted in the geographical and social conditions of individual Völker in many ways foreshadow the nineteenth century’s preoccupation with establishing genetic relationships between languages within the framework of the emerging sense of a distinct concept of nationhood. In the context of metrics, the rise of philology led to those early attempts to discover a common origin for Indo-European versification that were briefly touched upon in the previous chapter. Even more important as far as English versification is concerned, however, was the great attention which Anglicists bestowed upon Old and Middle English versification during the nineteenth century. Edwin Guest’s two-volume History of English Rhythms (1838) is notorious for its ambition to explain modern verse by the principles of Old English metre, for which purpose Guest devised an idiosyncratic system of twelve so-called ‘sections’ of stressbased syllable-groups. In spite of its quaintness, Guest’s work saw a second edition, edited and revised by W.W. Skeat, as late as 1882; this edition coincided almost exactly with the publication of the first volume of Jakob Schipper’s prodigious Englische Metrik (3 volumes, 1881-1888). Translated into English in 1910 as A History of English Versification, Schipper’s work was hailed by Saintsbury as ‘one of the foundation stones of a prosodic library’, and even today it remains a valuable work of reference. However, as T.V.F. Brogan points out, more than two thirds of Schipper’s History is dedicated to Old and Middle English versification, whereas modern versification is treated in considerably less detail.72 In terms of scope and ambition – if not in rigour and method – Guest and Schipper find their most obvious successor in George Saintsbury, whose History of English Prosody from the Twelfth Century to the Present Day (also 3 volumes, 1906-10) marks the last fully fledged attempt at encompassing the entirety of English verse practice in a single monumental work. Guest’s treatise 71

Isaiah Berlin, Three Critics of the Enlightenment: Vico, Hamann, Herder (Princeton: Princeton University Press, 2000), 168-242. 72 Brogan (1981): 11, from whence Saintsbury’s estimation of Schipper is also cited.

53

was published at the very brink of Queen Victoria’s reign, and the publication of Schipper’s English translation and of Saintsbury’s final volume coincided with the end of King Edward VII’s reign. This latter date marks the emergence of two new currents, both of which are crucial to this study: the introduction of free verse by the imagists and the development of linguistics as an academic discipline distinct from philology. While the emergence of free verse challenged the notion of metre as the basic prerequisite for versification, the rise of linguistics challenged the hegemony of diachronic language studies in general. These new currents heralded the era of modern metrics, which may be said to have begun with Otto Jespersen’s highly influential ‘Notes on Metre’, of which A. Walter Bernhard has commented: ‘as studies in modern linguistics tend ultimately to go back to Saussure, studies in modern metrics are generally based on Jespersen.’73 ❦ Otto Jespersen’s foundational text on the topic of metrics was originally delivered at the Royal Danish Academy of Sciences and Letters in 1900 and subsequently published as ‘Den psykologiske grund til nogle metriske fænomener’ [‘The Psychological Foundation of Some Metrical Phenomena’]. This essay was not translated into English until 1933, a fact lamented by Jespersen himself in a postscript to the translation: ‘I have often had occasion to regret that I wrote my paper in Danish and buried it in a place where fellow metrists in other countries were not likely to discover it.’74 In passing from Danish to English, Jespersen’s essay was renamed ‘Notes on Metre’ and emended in several respects; however, the overall purpose of the original essay as well as of its English translation remains the same: to provide ‘a fundamental revision of [the whole metrical science’s] principles, system of notation, and nomenclature.’75 In order to achieve this goal, Jespersen identifies three chief fallacies in earlier theorists’ dealings with metrics: the fallacy of longs and shorts, the fallacy of the foot and the fallacy of two grades. Having asserted that stress is the relevant feature on which metrical verse in the modern Germanic languages is based, Jespersen discards not only the concept of longs and shorts, but also the use of breves and macrons to scan this type of verse. Even when such symbols are redefined to designate weak and strong syllables respectively, their retention in critical writings, Jespersen insists, will make us prone to take over 73

A. Walter Bernhart, ‘Generative Metrics’, in Poetics 12 (April, 1974). Otto Jespersen, ‘Notes on Metre’ (1933), in Harvey Gross (ed.), The Structure of Verse. Modern Essays on Prosody, Revised Edition (New York: The Ecco Press, 1979), 105-128: 126. 75 Jespersen (1933): 106. 74

54

remnants of their former function. For the same reason, Jespersen argues in favour of discarding the whole nomenclature of foot-based prosody – terms such as iamb and trochee – because the use of it tends to reify the metrical foot, a construct for which Jespersen has no use in his metrical theory. Instead of thinking about verse lines as consisting of a certain number of feet based on two grades of stress (‘calling everything weak that is not strong’76), Jespersen proposes that we think of verse lines as consisting of a certain number of syllable positions, and of syllables as being stressed or unstressed only relatively to immediately adjacent syllables. Jespersen, in other words, is keenly aware that stress in English cannot be reduced to a simple binary opposition, for which reason he suggests a four-level hierarchy of stress, by which signifies minimal stress and signifies maximal stress, while intermediary levels of stress are signified by and . Remarking that iambic pentameters adhering to the form 1414141414 are quite scarce in English poetry, Jespersen explains: the only thing required by the ear is an upward and downward movement, a rise and a fall, an ascent and a descent, at fixed places, whereas it is of no importance how great is the ascent or the descent. It is therefore possible to arrange the scheme in this way, denoting the odd syllables a and the even ones by b: a / b \ a / b \ a / b \ a / b \ a / b ( \ a) ... It is the relative stress that counts.77 Jespersen’s four-level system for distinguishing relative stress was taken up by George Trager and Henry Lee Smith for their influential study of English phonology and morphology, An Outline of English Structure (1951). From this point onwards metrics, at least in America, came to be studied mainly within the framework of linguistic structuralism, as is evident from the proceedings of the Kenyon symposium of 1956, which were published in Kenyon Review 18 (1956) as English Verse and What It Sounds Like. Reintroducing the concept of the metrical foot, structuralist metrists such as Edmund L. Epstein and Terrence Hawkes employed the principles laid down by Trager & Smith to work out painstakingly extensive taxonomies for the English iamb,78 and the four-level stress system even found its way into decidedly popular treatments of metrics, 76

Jespersen (1933): 109-110. Jespersen (1933): 112. 78 See Edmund L. Epstein & Terence Hawkes, Linguistics and English Prosody (Buffalo: University of Buffalo Department of Anthropology and Linguistics, 1959), in which well over 6,000 types of possible iambs in English are mapped out. 77

55

such as G.S. Fraser’s student manual Metre, Rhyme, and Free Verse (1970).79 The most important outcome of American structuralist metrics remains, however, Seymour Chatman’s A Theory of Meter (1965), in which Jespersen’s four levels of linguistic stress are systematically mapped against the binary distinction between two levels of metrical accent, producing an admirably clear explication of tension between metre and speech rhythm in English. In much the same way as American structuralist metrics rose to prominence in response to a linguistic account of English phonology, namely Trager & Smith (1951), it also owed its eventual decline to the publication of yet another linguistic account of that same field. In 1968, Noam Chomsky and Morris Halle published their seminal The Sound Pattern of English, which work almost immediately spawned a string of new metrical theories – collectively referred to as generative metrics – that soon came to displace structuralism as the dominant theoretical framework for studying metrics in America. The first attempt at merging metrics with a generativist approach was made by Morris Halle and Samuel Jay Keyser in a 1966 article on ‘Chaucer and the study of prosody’, but it is their modified and expanded version of the original theory, put forward in the third chapter of English Stress: Its Form, Its Growth, and Its Role in Verse (1971), that is generally viewed as the foundational text of generative metrics. Taking the iambic pentameter as their object of inquiry, Halle & Keyser (1971) set out to provide a theory of metre that can account for our ability to distinguish between metrical and unmetrical lines as well as between different levels of metrical complexity. For this, the authors posit a tacit knowledge of certain principles of versification shared by all readers and poets of a given language. This knowledge, according to Halle & Keyser, is composed of two parts that are best studied separately: on the one hand, the recognition of abstract patterns underlying metrical utterances, and on the other hand, the correspondence rules that allow us to recognize certain strings of words as particularizations of a given pattern. The basic problem is illustrated by means of the following three verse instances: (a) Ode to the West Wind by Percy Bysshe Shelley (b) O wild West Wind, thou breath of Autumn’s being (c) The curfew tolls the knell of parting day80 79

More recently the four-fold classification of relative stress levels has been employed for the purpose of scansion in Timothy Steele, All the Fun’s in How You Say a Thing (Athens, OH: Ohio University Press, 1999). 80 Morris Halle & Samuel Jay Keyser, English Stress: Its Form, Its Growth, and Its Role in Verse (New York: Harper & Row, 1971): 139.

56

According to Halle & Keyser, only (b) and (c) fit the abstract pattern conventionally referred to as iambic pentameter, whereas (a) is unmetrical. Furthermore, English speakers familiar with the canon of English poetry will deem (b) more complex than (c), but certainly still acceptable. The task of the generative metrist, then, is to formulate a finite set of rules, the appliance of which will generate all permissible, but none of the impermissible realisations of a given abstract pattern, in this case the iambic pentameter. Halle & Keyser propose the following as a formalization of the abstract metrical pattern that underlies all utterances which can be recognized as iambic pentameters: (W)*S WS WS WS WS (X) (X)81 In this formalization each capital letter represents a position in the metre to be occupied by a syllable (see below, correspondence rule i, for clarification); the positions surrounded by parentheses are optional. W and S represent weak and strong positions respectively; X represents positions which can only be occupied by an unstressed syllable. An asterisk after a parenthesis indicates that the omission of the optional entity increases the line’s level of complexity; in this case an acephalous pentameter is thus deemed more complex than acatalectic variants. Two correspondence rules are then introduced to account for permissible realisations of this pattern: (i)

A position (S, W, or X) corresponds to a single syllable OR

to a sonorant sequence incorporating at most two vowels (immediately adjoining or separated by a sonorant consonant) ... (ii)

Fully stressed syllables occur in S positions only and in all S positions OR

Fully stressed syllables occur in S positions only but not in all S positions OR

Stress maxima occur in S positions only but not in all S positions82

81 82

Halle & Keyser (1971): 169. Halle & Keyser (1971): 169.

57

A ‘stress maximum’ is defined as ‘a fully stressed syllable [which] occurs between two unstressed syllables in the same syntactic constituent within a line of verse’,83 and only such syllables are barred from occupying a weak position in a line if it is to be metrical. Thus, if we take the three lines with which we began this section and map them onto Halle & Keyser’s metrical pattern, we get this result: 1

2

3

4

5

6

7

8

9

10

11

(W)*

S

W

S

W

S

W

S

W

S

(W) (W)

(a)

Ode

to

the

West

Wind

by

Per

cy

Bysshe Shell

ey

(b)

O

wild West Wind

thou

breath

of

Au

tumn’s

be

ing

(c)

The

few

the

knell

of

part

ing

day

cur

tolls

12

Syllables that receive lexical stress are given in bold, and we see that in (c) there is complete agreement between word stress and metrical pattern; this exemplifies the first formulation of correspondence rule ii above, making (c) highly metrical. In (b) there is a stressed syllable in the weak third position, which is a violation of the two first formulations of correspondence rule ii. However, since the syllable in question is enclosed on both sides by stressed syllables, it does not qualify as a stress maximum, and the line is therefore allowed by the third formulation of Halle & Keyser’s correspondence rule ii. The line is, of course, still metrical, but in a more complex manner than is (c). In (a), on the other hand, there are stressed syllables in weak positions 1, 5, 7 and 9, so clearly this line is in violation of the two first formulations of correspondence rule ii. However, since none of the syllables occupying positions 1, 5 and 9 is enclosed by unstressed syllables on both sides, they are not stress maxima and therefore acceptable under the third formulation of correspondence rule ii. The syllable occupying position 7, on the other hand, is a stress maximum in that it is enclosed on both sides by unstressed syllables belonging to the same prepositional phrase. The stressed syllable in position 7 is therefore a violation of all three formulations of correspondence rule ii, and consequently the whole line is rendered unmetrical, according to Halle & Keyser. The scope of this chapter does not permit a full review of generative metrics as it has branched out since Halle & Keyser. In accordance with Chomsky’s willingness to revise his own theories of grammar continuously (some might say compulsively), generative metrists have been 83

Halle & Keyser (1971): 169.

58

honing and reformulating their theories on metre for decades. Thus, as recently as 2008, Morris Halle (then in his mid-eighties) co-authored with Nigel Fabb Meter in Poetry: A New Theory. But whereas Chomsky’s modifications of his theories on grammar have generally been aimed at minimizing complexity, the field of generative metrics has tended to develop towards still higher degrees of complexity. While the early theory of Halle & Keyser can be fathomed by any reader with an interest in poetics, many later generative approaches to metrics presuppose a fairly advanced level of linguistic training. This, of course, is not a problem in and of itself, but it has problematic implications in that it divorces the area of metrics from the area of poetics. Whereas earlier theories of metre – say, Bysshe’s syllabism or Coleridge’s accentualism – were developed in tandem with or in response to the practice of versification and the actual reading of poetry, generative metrics is strictly the province of professionally trained linguists.84 For this reason, perhaps, generative approaches to metrics have attracted a good deal of criticism from metrists with more literary and discursive approaches. Some objections have been aimed at falsifying specific generative hypotheses, while other objections have been aimed at the very principles on which generative metrics rests. The daunting ambition of generative metrics – to produce a finite set of rules that correctly predict all metrical instances, while keeping out all unmetrical instances – makes individual hypotheses extremely vulnerable to falsification. The theory put forward by Halle & Keyser, for instance, discards a line such as ‘Ode to the West Wind by Percy Bysshe Shelley’ as being unmetrical, on account of the fact that it contains a stress maximum in the seventh syllable position. By that standard, the line ‘Burnt after them to the bottomless pit’ is not an iambic pentameter either, although the fact that it occurs in Paradise Lost would suggest otherwise. As Reuven Tsur, from whose work the Milton line has been borrowed, notes: ‘no criteria for metricality has yet been devised that were not violated by such masters of musicality as Milton and Shelley’.85 However, a far more serious criticism of Halle & Keyser is that by Derek Attridge, who upon concocting the line ‘Óde to the Wést Wínd by Jámes Élroy Flécker’ observes: such a line [i.e. Attridge’s own concoction, without stress maxima] would be regarded as an example of any metre which allows lines of eleven syllables, since the only way a line can be finally rejected as a realisation of an abstract metrical pattern is if stress maxima occur in 84

For discussion, see Beyers (2001): 233-236. Reuven Tsur, ‘Poetic Rhythm: Performance Patterns and Their Acoustic Correlates’, in Versification [an electronic journal of literary prosody], Vol. 1 (1997). 85

59

weak positions ... We could, therefore, exchange strong and weak positions and the line would still pass muster as a realisation, albeit a complex one, of the opposite metrical pattern.86 Attridge’s own theory of metre as set forward in The Rhythms of English Poetry (1982) and a string of supporting works continues to exert a considerable influence within contemporary studies of versification, and deservedly so. It combines the methodological rigour of the best generativist approaches to metre with a keen concern for the experience of poetic rhythm on the part of the reader. As a central part of his theory Attridge posits the concept of underlying rhythm, a substructure of which traditional metrical patterns are conventionalized particularizations. These underlying rhythms are simple patterns of energy pulses that are of both muscular and cognitive origin. Strong pulses (‘beats’) alternate with phases of relaxation (‘off-beats’), creating patterns that are reinforced by repetition and perceived periodicity. One of the most natural and versatile underlying rhythms is that of four beats repeated four times, i.e. ‘a single beat doubled, doubled again, doubled again, and doubled once more’.87 This basic pattern, Attridge argues, underlies a vast corpus of English verse that is inadequately accounted for by both traditional and generativelinguistic approaches to metre, namely demotic verse and folk song. Attridge’s own scansion of a section from A.A. Milne’s ‘Disobedience’88 may serve as an illustration of his theory’s merits: James James said to his Mother, B B B B “Mother,” he said, said he; B B B

[B]

“You must never go down to the end of the town, B B B B If you don’t go down with me.” B B B [B] Each B represents, of course, a single beat in the 4 X 4 beat structure that underlies these lines, and two things should be noted immediately. First, that the relaxation phase between each beat may be 86

Attridge (1982): 41-42. Derek Attridge, ‘Rhythm in English Poetry’, in New Literary History, Vol. 21, No. 4 (Autumn, 1990), 1015-1037: 1016-1017. 88 Attridge (1990): 1017. 87

60

occupied by two syllables (‘never go down’: double off-beat), one syllable (‘don’t go down’: single off-beat) or even no syllable (‘James James’: implied off-beat) at the poet’s discretion. Second, that a beat is not necessarily voiced in performance but can instead function as a mandatory pause; Attridge calls such beats ‘unrealized beats’ or ‘virtual beats’ and annotates them with a B in square brackets. The dependency on unrealized beats in our performance of certain types of verse has the important implication that rhythm is not reducible to patterns of syllables, as is commonly assumed in traditional approaches to metre. Rather, conventional forms and metres represent particular sets of rules by which their underlying rhythm may be realized. Whereas in Milne’s ‘Disobedience’ the utilization of the 4 X 4 pattern allows for a high degree of syllabic variation in each line (from six to twelve syllables) as well as for unrealized beats, the 4 X 4 pattern is used far more restrictedly in this example by Andrew Marvell: My Love is of a birth as rare B B B B As ’tis, for object, strange and high; B B B B It was begotten by Despair B B B B Upon impossibility. B B B B Marvell’s stanza is wrought in meticulous iambic tetrameter, a metre that does not allow for unrealized beats and in which there is a strong preference for single off-beats. Still, iambic tetrameter at its most basic level exploits the same underlying rhythm as does Milne’s stanza, i.e. 4 X 4. Combining the preference for single off-beats with the admission of unrealized beats, the hymn measures of English are also based on the underlying 4 X 4 rhythm. Common measure – and, by extension, the ballad stanza – has an unrealized beat in the eighth and sixteenth positions (as had indeed Milne’s stanza):

61

How eager are my thoughts to roam B B B B In quest of what they love! B B B

[B]

But ah! when duty calls them home, B B B B How heavily they move! B B B

[B]

In short measure, on the other hand, the fourth, eighth and sixteenth beats are unrealized, a pattern that is also found in limericks, in which genre double off-beats dominate: There was an old person of Skye, B B B

[B]

Who waltz’d with a Bluebottle fly: B B B

[B]

They buzz’d a sweet tune, to the light of the moon, B B B B And entranced all the people of Skye. B B B

[B]

The fact that the 4 X 4 rhythm also underlies the limerick neatly illustrates the important point that typographical conventions do not always disclose the underlying 4 X 4 structure. Limericks are conventionally printed in such a manner that the fully realized penultimate four-beat sequence is divided into two lines with two beats in each. With this in mind, it is quite easy to disregard typographical lineation and instead focus on spotting the underlying 4 X 4 structure of the fourteener – There’s not a joy the world can give like that it takes away B B B B B B B

[B]

When the glow of early thought declines in feeling’s dull decay; B B B B B B B [B] – as well as of poulter’s measure:

62

I could not though I would: good Lady say not so, B B B [B] B B B

[B]

Since one good word of your good wil might soone redresse my wo B B B B B B B

[B]

The ease with which the human brain perceives four-beat sequences and arranges them into hierarchies accounts for the great formal versatility of the 4 X 4 rhythm as a cognitive and muscular basis for verse and song. So powerful is the grasp of the 4 X 4 pattern that it can be used quite effortlessly as the basis for communal rhythmic activity even by children who are relatively young (counting-out games) and by adults who are relatively drunk (drinking songs). However, the same cognitive naturalness of the four-beat pattern also means that it has a tendency to assert itself even in forms and metres where a strong rhythmical pulse is not desirable. Verse forms based on the four-beat pattern thus have a tendency to fall into sing-song, for which reason such forms are generally deemed unsuitable as metrical frameworks for lengthy poems, as well as for serious poetry. The one common rhythm in English verse that resists the tyranny of the four-beat pattern is the five-beat pattern. As Attridge explains: A crucial feature of five-beat verse is the absence of any strong rhythmic hierarchy. There is no obvious way lines of five beats could divide into half-lines, and since no doubling movement is set up within the line, there is no encouragement to group the lines into pairs and larger units. Five-beat lines are consequently more self-sufficient than four-beat lines; unlike the latter they arouse no strong expectation for further lines to continue an unfolding pattern, but at the same time they do not provide the strong sense of finality brought about by the completion of one part of a larger pattern.89 The special character of the five-beat rhythm is also reflected in the fact that – unlike the four-beat rhythm, which underlies a wide array of different metrical forms in English – the five-beat rhythm only underlies a single, highly specialized, metrical form in English, namely the iambic pentameter.

89

Derek Attridge, Poetic Rhythm: An Introduction (Cambridge: Cambridge University Press, 1995): 162.

63

The next three chapters are devoted almost exclusively to the iambic pentameter: its history, its properties and its possibilities. However, before embarking on that venture, we must first make mention of a school of metrists whose body of work this dissertation is much indebted to. While generative metrics in the latter part of the twentieth century has been the dominant linguistic approach to metre in the west, especially in the USA, metrists trained in Russia have generally focused on statistical approaches to studying metre. Historically, the statistical approach to metre goes back to Andrei Bely’s 1910 study of stress omission in Russian verse, for which Bely developed a rather compelling method of representing metrical complexity graphically.90 Bely’s statistical approach to metrics was picked up by Boris Tomashevsky in the 1920s and also came to influence Roman Jakobson, as we shall see in Chapter 7. In more recent times, statistical metrics has found its most important theorists in M.L. Gasparov and Marina Tarlinskaja; their work has been of pivotal significance to this study, particularly Gasparov’s A History of European Versification (1996) and Tarlinskaja’s English Verse: Theory and History (1976). In the Anglophone world, the statistical approach to metrics has been explored most thoroughly by James Bailey during the 1970s and, more recently, by Martin J. Duffell. The latter’s A New History of English Metre (2008) has also been of great importance to this study.

90

Bely’s study – which, as far as I have been able to discover, has not been translated – was featured in his Simvolizm [‘Symbolism’] (Moscow: 1910). For an example of Bely’s graphical representation of metrical complexity, with an explanation in English, see B. Elan Dresher & Nila Friedberg (eds), Formal Approaches to Poetry: Recent Developments in Metrics (Berlin & New York: Mouton de Gruyter, 2006): 235. Bely’s treatise on metrics was held in particularly high regard by Vladimir Nabokov, who described it as ‘probably the greatest work on verse in any language’. Furthermore, Nabokov listed Bely’s 1913 novel Petersburg as one the four greatest novels of the twentieth century, the others being Joyce’s Ulysses, Kafka’s Metamorphosis and Proust’s À la recherche du temps perdu; on this, see Brian Boyd, Vladimir Nabokov: The Russian Years (Princeton: Princeton University Press, 1990): 149.

64

Chapter 3 Unseen Roots in Tongue-Tied Springs: The English Pentameter Still growing like the plants from unseen roots In tongue-tied Springs Elizabeth Barrett Browning, Aurora Leigh

It is tempting to think of the English pentameter as a rational historical compromise between the native Germanic accentual four-beat line and the invading syllabic hexameter line of the French alexandrine. Unfortunately it is also wrong. To say that the English pentameter is the happy result of mutual metrical hybridization between Germanic and Romance systems of versifications requires qualification and is true only in a very restricted sense. Provisionally, it should be remembered that the alexandrine, in spite of its subsequent air of quintessential Frenchness, is a relatively novel type of verse design in French poetry. Although it makes its first appearance in the twelfth century chanson de geste entitled Le Pèlerinage de Charlemagne à Jérusalem – i.e. almost a century after the Norman Conquest – the alexandrine does not assume its dominant position in French verse until the seventeenth century.1 For that reason alone the mock proposition with which we began this paragraph is a blatant anachronism. There are, however, significant similarities between the way in which free verse, through translations and adaptations, crept into English and the way in which the iambic pentameter made its entry in English literature. The honour of being the earliest English poem in what reasonably may be referred to as iambic pentameter has traditionally been awarded to Geoffrey Chaucer’s ‘An ABC’,2 which is a translation of a prayer out of Guillaume de Deguileville’s Pèlerinage de la vie humaine. The more technically accomplished chansons de geste, among which those by Guillaume de Deguileville must be counted, were composed in decasyllabic lines arranged into laisses of irregular length, which might lead us to assume that Chaucer for his earliest exercise in iambic pentameter simply took over the decasyllabic line from Guillaume’s original and adapted it to 1 2

Gasparov (1996): 130. Tarlinskaja (1976): 138.

65

English. The truth, however, is a good deal more complicated in that the French original which forms the source text of Chaucer’s translation is, in fact, an abecedarian poem within the poem, made up of octosyllabic lines arranged in twelve-line stanzas: Xristus, ton filz, qui descendi

Xristus, thi sone, that in this world alighte,

En terre et en la crois pendi,

Upon the cros to suffre his passioun,

Ot pour moy le costé fendu.

And eek that Longius his herte pighte

Sa grant rigour il destendi

And made his herte blood to renne adoun,

Quant pour moy l’esperit rendi,

And al was this for my salvacioun;

Son corps pendant et estendu;

And I to him am fals and eek unkynde,

Pour moy son sanc fu espandu.

And yit he wole not my dampnacioun —

Se ceci j’ai bien entendu

This thanke I yow, socour of al mankynde!

A mon salut bien entendi,

(‘An ABC’, ll. 161-168)3

Et pour ce, se l’ay offendu Et il ne le m’a pas rendu, Merci t’en rens, graces l’en di. (‘La Priere de Nostre Dame’, ll. 241-252)4 Both the metre and the stanza-form are changed by Chaucer in the process of translation, perhaps to escape the predicament of having to follow in English Guillaume’s insanely close-knit rhyme scheme (aabaabbbabba).5 Instead Chaucer makes use of that early rival to the rhyme royal which is sometimes referred to as the ballade stanza (i.e. ababbcbc)6 and which may best be described as a sort of curtailed precursor to the Spenserian stanza, which follows the same rhyme scheme but adds a concluding hexameter which rhymes with the line immediately preceding it. According to Thomas Speght’s printed edition of Chaucer’s works (1602), the ‘ABC’ was composed ‘at the request of Blanche, Duchesse of Lancaster, as a praier for her privat use’.7

3

All references to Chaucer are to The Riverside Chaucer, Third edition, edited by Larry D. Benson (Oxford: Oxford University Press: 1987). 4 Cited in The Complete Works of Geoffrey Chaucer, edited by Walter W. Skeat [1894], Second edition [1899] (Oxford: Oxford University Press, 1972). 5 After all, Chaucer did – in ‘The Complaint of Venus’ – complain about the ‘skarsete’ of rhymes in English. 6 E.g. Lewis Turco, The Book of Forms: A Handbook of Poetics (Hanover, NH: University Press of New England, 2000): 129-132. 7 Cited from Derek Pearsall, The Life of Geoffrey Chaucer: A Critical Biography (Oxford: Oxford University Press: 1992): 83-84). Here Pearsall also points out that Speght’s editorial comment is, in fact, all the evidence there is to

66

Blanche died in 1368 or 1369, probably as a result of the bubonic plague’s third influx in Britain, and tradition has it that it was in commemoration of her that Chaucer was commissioned by John of Gaunt to compose The Book of the Duchess, his first major work if the partial translation of Roman de la Rose is kept out of the chronology. The Book of the Duchess – like the later The House of Fame – is composed in octosyllabic couplets, a form for which there was already ample precedence in English versification at the time when it was taken up by Chaucer. It is, however, instructive to compare Chaucer’s octosyllables with those of his predecessors in that such a comparison sheds light on Chaucer’s unique contribution to the kind of metrical verse that would later come to dominate English poetry to such an extent that the modernists felt compelled to revolt against it. Before we do so, however, we must examine the history of the octosyllabic line in some detail. ❦ The English rhymed octosyllable is, not surprisingly, of French pedigree. Ultimately adapted from early Latin hymns,8 the octosyllable is the earliest metre to be testified in French versification, and it emerges by the late tenth century, in the poem La Passion du Christ.9 This poem is composed of 129 stanzas, each of which contains four octosyllables (aabb), but – like the later, decasyllabic Chanson de Roland and many other early works of French verse – it makes use of line-terminal assonance rather than end-rhyme to tie together its couplets. For some reason the principle of lineterminal assonance was never adopted by poets composing in English: from the earliest times at which octosyllabic and decasyllabic line forms were borrowed from the Romance languages, these forms invariably – at least until the introduction of blank verse10 – carried with them the requirement of end-rhyme in English. The reasons for this are sketchy, but worth exploring: after all, it does seem curious that a vernacular poetry, which up until the Norman Conquest did not employ end-rhyme at all,11 proved itself so receptive to that concept, not least in view of the fact that end-rhyme was not even a mandatory feature of the verse design that was being emulated. One might speculate that because the native Old English system of versification had relied primarily on patterned repetitions of stressed consonants and consonant clusters, such sounds were deemed too suggest that ‘An ABC’ was composed before the death of Blanche in 1369. Instead he proposes that the late 1370s as a more likely date of composition. 8 On which, see Gasparov (1996): 105-110, 128-129. 9 Gasparov (1996): 128. 10 On which, see Chapters 4 and 5. 11 With the notable exception of a text which is found in the third booklet of the Exeter Book and which is usually referred to as ‘The Rhyming Poem’.

67

important to be left out of the equation, especially in the crucial line-terminal position. An alternative explanation could be that the transmutation of terminal assonance into end-rhyme was simply the result of hypercorrection on the part of the English versifiers. Marina Tarlinskaja offers a modified synthesis of these explanations and suggests that ‘Rhyme was seemingly better suited to the strong stress which had emerged at the end of the phrase and helped to consolidate the uncertain new meter’.12 Finally, it might be prudent to take into consideration the very likely possibility of influence from Latin hymns. Such hymns did employ perfect end-rhyme and were widespread in Britain at the relevant point in time, for which reason it must be considered that their influence was so strong that end-rhyme was also adopted into secular verse. In order to trace the path by which end-rhymes eventually came to dominate over the use of assonance in the codification of the octosyllabic couplet in both French and English, we have to move forward in time a good century from the time of La Passion du Christ. At some point during the first quarter of the twelfth century, a monk who refers to himself as ‘li apostoiles danz Benedeiz’13 composes the earliest narrative work in rhymed octosyllabic couplets to have survived in Medieval French literature.14 Adapting the tradition of the Old Irish immram (a genre concerned with a hero’s sea journey to the Otherworld), Benedeit’s work details the legendary voyage of St. Brendan, the sixth century Irish abbot who allegedly journeyed to the ‘Isle of the Blest’. While this voyage served as the model for a large number of works in both prose and verse during the Middle Ages – in Latin as well as in several vernaculars – Benedeit’s Brendan was evidently the most successful version.15 The poem consists of close to a thousand couplets, the vast majority of which are every bit as regular as the two couplets which terminate the poem: Quant vint al tens que il finat, Ralat u Deus lui destinat. El regne Deu, u alat il, Par lui en vunt plusur que mil. (Benedeit, Brendan, ll. 1831-1834)16 12

Tarlinskaja (1976): 87. Cited from C.W. Aspland (ed.), A Medieval French Reader (Oxford: Oxford at the Clarendon Press, 1979): 19. 14 It should, however, be noted that the metre of St. Brendan allows couplets in which it is the seventh rather than the eighth syllable that receives stress. In such lines the stressed seventh syllable is always followed by an unstressed syllable. 15 Ian Short & Brian Merrilees (eds), Benedeit: The Anglo-Norman Voyage of St Brendan (Manchester: Manchester University Press, 1979): 3. 16 Cited in Short & Merrilees (eds) 1979: 79. 13

68

To the benefit of those of us whose knowledge of Old French is at best cursory, Benedeit’s most recent editors in English, Ian Short and Brian Merrilees, have appended to their introduction to this work a phonetic transcription of a short section from Brendan, which clearly demonstrates that Benedeit’s couplets are indeed remarkably regular in terms of both syllable count and end-rhyme.17 However, in spite of its metrical regularity and the popularity it enjoyed in its own age, Benedeit’s Brendan has rarely received recognition for its crucial contribution to the historical development of Anglo-French metrics. The same is true of Geffrei Gaimar’s Estoire des Engleis, which is an adaptation into French of the Anglo-Saxon Chronicle, probably composed in the late 1130s. With its surviving 3,300 octosyllabic couplets, the Estoire des Engleis is the earliest extant work of historiography in the French language,18 and like Benedeit’s Brendan it shows a high degree of regularity in its treatment of the octosyllable: En Denemarche le regnez aveit quatre riches contez, e en Bretaigne aveit conquis Cäir Cöel od le païs: (Estoire des Engleis, ll. 71-74)19 The fact that one of the earliest works in French to consist of rhymed octosyllabic couplets is based on The Anglo-Saxon Chronicle seems appropriate for a chapter like this which seeks to lay bare the shared Anglo-French roots of that particular verse design. However, the interest in the history of England, to which Gaimar’s work responded, is in fact a general trait of the period, as is evident also from the existence of a number of early twelfth century Latin texts on the subject.20 Two decades after its appearance, Gaimar’s Estoire was thus succeeded – and to a large extent replaced – by another ‘history’ of England in rhymed octosyllabic couplets, namely Wace’s Roman de Brut. 17

Short & Merrilees (eds) (1979): 15-16. Their transcription comes, of course, complete ‘with all the many reservations which this sort of exercise calls for’. 18 John Gillingham, ‘Kingship, Chivalry and Love: Political and Cultural Values in the Earliest History Written in French: Geoffrey Gaimar’s Estoire des Engleis’, in C. Warren Hollister (ed.), Anglo-Norman Political Culture and the 12th-Century Renaissance (Woodbridge, The Boydell Press: 1997), 33-58: 33. 19 Cited from Geffrei Gaimar, Estoire des Engleis, edited and translated by Ian Short (Oxford: Oxford University Press: 2009): 6. 20 Gaimar’s work was thus preceded by that of his more famous contemporary, Geoffrey of Monmouth, the Historia Regum Britanniae (1136). This work was in turn preceded by William of Malmesbury’s Gesta Regum Anglorum (1120) and Henry of Huntingdon’s Historia Anglorum (1129).

69

However, to say that three authors – Benedeit, Gaimar and Wace – were responsible for the early codification of the rhymed octosyllabic couplet in French during the first half of the twelfth century conceals a crucial point, namely that this form, strictly speaking, is not so much a French invention as it is an Anglo-Norman invention. The term Anglo-Norman I use to designate those dialects of French which developed on both sides of the Channel in the centuries following the Norman Conquest, as a result of prolonged contact between Norman and English. These dialects differed notably from the variety of French spoken in the area around Paris – occasionally referred to as Francien – in terms of both syntax and phonology.21 For that reason it is of great importance to notice that each of our three authors employed an Anglo-Norman variety of French for his octosyllables. Wace was born in Jersey and brought up in Normandy where he ended his days as canon of Bayeux: his writings are among the few extant records of Old Norman that we have. Less is known about Gaimar’s life, but scholars have identified Lincolnshire as his most likely place of residence during the composition of Estoire des Engleis.22 Of Benedeit’s identity nothing is known, but internal, linguistic evidence clearly positions him in the insular region of the Anglo-Norman area.23 Moving forward in time, while staying in the insular region of the Anglo-Norman area, to the end of the twelfth century, we encounter another equally obscure poet, who also employs the medium of the rhymed octosyllabic couplet, even for the purpose of introducing herself to her audience: Me numerai pur remembrance: Marie ai num, si sui de France (Les Fables, ‘Epilogue’, ll. 3-4)24 The couplet is from Marie de France’s Ysopet – i.e. her translation into French of a selection of Aesop’s Fables – and it is of great importance, not only as an exemplar of end-rhyme and metrical regularity, but especially in that it has provided us with an apt appellative for its otherwise elusive 21

See Johan Vising, Anglo-Norman Language and Literature (London: Oxford University Press, 1923): 27-33, and Martin J. Duffell, ‘Some Phonological Features of Insular French: A Reconstruction’, in Roger Wright & Peter Ricketts (eds), Studies on Ibero-Romance Linguistics Dedicated to Ralph Penny (Newark, DE: Juan de la Cuesta, 2005): 103125. 22 Short (ed.) (2009): x. 23 Short & Merrilees (eds) (1979): 6. 24 ‘I shall name myself for posterity: My name is Marie, and I am of France’. Cited from Logan E. Whalen, Marie de France & the Poetics of Memory (Washington, D.C.: The Catholic University of America Press, 2008): 124.

70

maker.25 A contemporary of Chrétien de Troyes, Marie de France can be said to personify in the most anonymous of manners the sort of multilingual cross-fertilization which went on between French and English versification in the centuries after the Norman Conquest. Her having translated Aesop’s Fables from Latin into Anglo-Norman26 indicates a surprisingly cosmopolitan horizon for a woman living in England at this point in history, but at the same time it should be remembered that according to her own presentation of her background – ‘si sui de France’ – she was, in fact, a native of the historical province of Île-de-France. This, at least, is how we have come to think of her nowadays, but in reality next to nothing is known of Marie de France’s true identity: even the scant biographical details provided above are speculative, at best.27 But Marie is an important figure nonetheless in that it is to her that the authorship of a series of lais preserved in a thirteenth century manuscript known as Harley 978 has been attributed.28 The lai is an exclusively Anglo-French genre of verse that is poised somewhere between the narrative and lyrical mode and usually borrows its themes from Breton folklore. It is a particularly interesting genre for our purpose in that lais provide us with some of the earliest extensive examples of relatively stable octosyllabic couplets in English literary verse. English couplets such as the following, from the opening of the Middle English lai Sir Orfeo, begin to emerge with increasing frequency by the late thirteenth century: Himself he lerned forto harp, And leyd theron his wittes scharp; He lerned so ther nothing was A better harpour in no plas. (Sir Orfeo, ll. 29-32)29

25

We owe the appellative ‘Marie de France’ to the scholar Claude Fauchet, who was the first to make use of this name in his Recueil de l’origine de la langue et poésie françoise (1581); see R. Howard Bloch, The Anonymous Marie de France (Chicago, IL: The University of Chicago Press, 2003): 2. 26 While Marie claims to have translated the Fables into ‘Romance’ from an English version which she attributes to King Alfred, scholars agree that at least the first forty of Marie’s Fables appear to be taken directly from Latin sources (Bloch (2003): 7; Whalen (2008): 105). 27 For a fascinating account of the uncertainty that surrounds the identity of Marie de France, see Bloch (2003). 28 The identification of Marie as the author of the lais contained in Harley 978 is a relatively novel occurrence. The first to have considered this idea is thus the eighteenth-century scholar and editor of Chaucer, Thomas Tyrwhitt (Bloch (2003): 3). 29 Cited from Anne Laskaya & Eve Salisbury (eds), The Middle English Breton Lays (Kalamazoo, MI: Medieval Institute Publications, 1995). Future references to Sir Orfeo are cited from the same source.

71

The syllable count, the exactitude of the end-rhymes and even the stress patterns of this short section are remarkably regular, especially when compared to slightly earlier metrical romances in Middle English such as, for instance, King Horn (mid-thirteenth century). The versification of King Horn – with its short, irregular lines30 and the occasional reliance on rather weak rhymes – certainly owes as much, if not more, to the quaint metre of Layamon’s Brut as it does to the octosyllabic couplet: Twelf feren he hadde That he alle with him ladde, Alle riche mannes sones, And alle hi were faire gomes, (King Horn, ll. 21-24)31 For an example of a work that can be said to represent a half-way stage between the rugged couplets of King Horn and the relatively smooth regularity of Sir Orfeo, one might turn to the romance Havelok the Dane, probably composed between 1280 and 1290: Here I schal biginnen a rym; Krist us yeve wel god fyn! The rym is maked of Havelok – A stalworthi man in a flok. (Havelok the Dane, ll. 21-24)32 In this excerpt, the syllable count is much stricter than what we saw in King Horn – three of the lines quoted thus contain exactly eight syllables (provided that ‘maked’ in line 23 is pronounced as a monosyllable) – and in terms of stress patterns, it should be observed that each line terminates in an unequivocally stressed syllable. But at the same time it must be conceded that the stress patterns leading up to those terminal stressed syllables are – especially in lines 21 and 24 – decidedly uncouth. Furthermore, it should be noticed that the partial rhyme between ‘rym’ and ‘fyn’, in which 30

More than 70 per cent of the lines in King Horn thus contain fewer than eight syllables; see Tarlinskaja (1976): 256. Cited from Ronald B. Herzman, Graham Drake & Eve Salisbury (eds), Four Romances of England (Kalamazoo, MI: Medieval Institute Publications, 1999). 32 Cited from Herzman, Drake & Salisbury (eds) (1999). 31

72

syllables with two different nasal consonants following the same vowel sound are made to rhyme with one another, closely resembles the phonological structure of the partial rhyme between ‘sones’ and ‘gomes’ in King Horn. It is thus possible to construe a genealogy of the octosyllabic couplet in English where – under the influence of French and presumably also Latin practices – each generation of English versifiers comes to observe more and more regularity in their handling of the form. From the Brut to King Horn, from King Horn to Havelok the Dane, from Havelok to Sir Orfeo: each step appears to validate such an interpretation, making it tempting to conclude that Chaucer’s use of the octosyllabic couplet was simply the next inevitable step towards regularity. Such an interpretation would, however, be valid only to a certain extent. While it is true that Chaucer’s couplets in The Book of the Duchess are far more regulated in terms of observing a fixed syllable count than those employed in Sir Orfeo,33 it would be unkind to the legacy of Chaucer to bestow on him the dubious honour of having merely regularized the octosyllable to meet the formal standards of its French counterpart. Thus, in addition to observing a stricter syllable count than anything we find in earlier English lais and romances, Chaucer’s octosyllables also exhibit a much smoother and more regular alternation between less stressed and more stressed syllables than do those in Sir Orfeo. This last assertion will, I think, be confirmed by most readers who venture even a cursory comparative reading of the two relevant texts, but it can also be backed up by statistical data. Again it is Marina Tarlinskaja who comes to our rescue. According to her count of metrical ictuses – i.e. metrical strong positions likely to be realized by lexically or syntactically stressed syllables – in early English octosyllables, only 79.3 per cent of the lines in Sir Orfeo contain four ictuses, whereas in The Book of the Duchess that percentage is 100.34 This feature of Chaucer’s octosyllabic verse is of great significance in that it represents a genuine structural addition to the French octosyllable, whose design is determined almost exclusively by isosyllabic considerations. A crucial structural difference between the octosyllable and the decasyllable in early French versification is that whereas the latter line-form must have a caesura, the former is characterized by not having one.35 We shall return to the intricacies of the caesura in the Romance decasyllable later in this chapter, but first we must, of course, consider the implications of its absence in the 33

According to Marina Tarlinskaja’s calculations, a staggering 84 per cent of the lines in The Book of the Duchess thus contain exactly eight syllables; the corresponding figure for Sir Orfeo is as low as 43.2 per cent. Furthermore, the number of syllables per line varies between five and eleven in Sir Orfeo, whereas in The Book of the Duchess, it varies between seven and nine (see Tarlinskaja (1976): 256; Table 18). 34 Tarlinskaja (1976): 257; Table 19. 35 Gasparov (1996): 125-127; Duffell (2008): 44.

73

octosyllable. However, before we do so, a distinction needs to be drawn between two different meanings of the term caesura if the claim just made about its different status in the octosyllable and the decasyllable respectively is to make sense. In its broadest, least restricted sense, the term caesura designates any syntactically motivated line-internal pause, regardless of whether that pause forms part of the line’s metrical structure or not. However, in its more restricted sense, the term caesura designates only a pause which also functions as a structural constituent of the metre: that is to say that such a caesura constitutes a non-optional part of the verse design. We shall, for the remainder of this chapter, refer to this latter type of caesura as a metrical caesura. To refer to the kind of syntactically motivated line-internal pause which does not form part of a poem’s verse design but is only a feature of the individual verse instance, we shall, whenever such a distinction is desirable, use the term non-metrical caesura.36 A metrical caesura, in other words, is a constituent at the level of the verse design, whereas a non-metrical caesura is a constituent only at the level of the verse instance. For an example of a type of verse which makes use of a metrical caesura, we might think of the French neo-classical alexandrine, in which the syntactical slicing of the verse in the middle is a compulsory feature of every line. For an example of a metre without a metrical caesura, we may, since it serves our present purposes so well, think of the French octosyllable. But the fact that a metrical caesura does not form part of the verse design of the French octosyllable does not mean that non-metrical caesurae are barred from occurring in individual verse instances of this particular verse design. We have already cited one line by Marie de France which clearly makes use of a non-metrical caesura, namely the one that reads ‘Marie ai num, ‖ si sui de France’. However, such lines are comparatively rare – both in the poems attributed to Marie de France and in other French works employing the octosyllable – but when they do occur, the caesura is almost inevitably placed in mid-line position as it indeed is in the line by Marie just cited. As we shall see, this position of the caesura after the fourth syllable foreshadows the most common structure of the French decasyllable, but it can also be observed in a few very early octosyllabic poems. Duffell thus gives as a rare example of an octosyllabic verse design which does employ a metrical caesura the metre of the eleventh century poem Vie de Saint Léger: in this metre, ‘the fourth as well as the final syllable in the line invariably has stress’, implying the presence of a mid-

36

Duffell (2008) draws the same distinction but insists that the term caesura should be used only to designate what is here called metrical caesura, whereas a line with a non-metrical caesura should be referred to as ‘containing two cola, thus emphasizing the grammatical as distinct from metrical nature of the subdivision’ (9). However, since the term colon is already a specialized rhetorical term with a more restricted sense than is required for the various workings of the non-metrical caesura, Duffell’s terminology is not adopted here.

74

line metrical caesura.37 The exact same verse design has been shown by Steven Guthrie to structure the metre of the early chanson de geste Gormond et Isembart (second half of the eleventh or first half of the twelfth century),38 and, finally, both La Passion du Christ and Benedeit’s Brendan, from which we have already quoted an excerpt above, make use of a similar verse design.39 However, these works should be viewed as early deviations from the subsequently established general rule of leaving out caesurae in the structural design of the octosyllable. Both the absence of a metrical caesura and the tendency to avoid the use of non-metrical caesurae in its place were taken over with a considerable degree of diligence by early imitators of the French octosyllable in English. In the more than 600 lines which make up Sir Orfeo, I have thus been able to find only a single couplet in which both lines with reasonable certitude can be said to contain a non-metrical caesura: Allas! thy rode, ‖ that was so red, Is al wan, ‖ as thou were ded; (Sir Orfeo, ll. 107-108) We get a much better impression of a typical line out of Sir Orfeo from quoting the line which follows immediately after the couplet just cited: And also thine fingres smale (l. 109) In this line, there is no obvious position in which to place a caesura, in spite of the fact that it contains as many as nine syllables. The line constitutes a single prosodic unit and therefore closely resembles the French octosyllable’s default intonational contour, in which lexical stress is largely subdued through the pivotal role accorded to phrasal end-stress.40 That early English imitators of 37

Duffell (2008): 43. Evelyn Birge Vitz, Orality and Performance In Early French Romance (Cambridge: D.S. Brewer, 1999): 7. 39 Gasparov (1996): 133; Aspland (ed.) (1979): 19. 40 That phrasal end-stress takes precedence over the lexical stress of individual words in French is, of course, a general feature of the French language; see, for instance, Bernard Tranel, The Sounds of French: An Introduction (Cambridge, New York: Cambridge University Press, 1987): 194-200. However, this trait is more pronounced in present-day French than it was in Old French: ‘In the earlier period [of Old French], when the tonic stress was intense, words remained, broadly speaking, the unit of the phrase but, in later Old and Middle French, words closely connected in thought ... were more and more run together and thus the phrase or locution became the sentence-unit instead of the word’ (Mildred K. Pope, From Latin to Modern French (Manchester: Manchester University Press [1932] 1973): 82). 38

75

the French octosyllable proved themselves so willing to resist the temptation of putting in caesurae in their handling of the form is, in a certain sense, curious. After all, it is difficult to think of a verse design which is more dependent on the constitutive role of a metrical caesura than the Old English alliterative metre, in which the metrical function of the caesura is so pronounced that it has become customary to speak of lines in this metre as consisting of two hemistichs. No discussion of caesurae, metrical or otherwise, in any type of verse is complete without some consideration of the role played by enjambment in that verse type. Whereas a caesura represents a line-internal pause caused by a syntactical juncture in the text, enjambment represents the lack of a syntactical juncture at line ends. Speaking in the most preliminary of terms, enjambment occurs whenever a line break does not coincide with a syntactical juncture in the text, and enjambment therefore constitutes one extreme of a scale whose opposite extreme is constituted by an end-stopped line.41 On first sight, it is something of an oddity how we have come to use for this very common poetic device in English a term which so openly flaunts its French origin, since a number of French verse designs are characterized by an all but complete lack of enjambment.42 Prime among these is the neo-classical alexandrine: we recall the anecdote of how an enjambed alexandrine allegedly caused a riot in the audience at the premiere of Victor Hugo’s verse play Hernani in Paris in 1830. However, enjambment is a much favoured device in other forms of French verse: Chrétien de Troyes seems to have been the first poet in the European vernaculars to have employed the device systematically, even though the term enjambment was not used until the sixteenth century when it was coined by Pierre de Ronsard.43 Chrétien de Troyes’s use of the octosyllabic couplet thus diverges significantly from that of his Anglo-Norman predecessors, Benedeit, Gaimar and Wace. Whereas the latter poets used closed couplets exclusively, with little or no enjambment even between lines belonging to the same couplet,44 Chrétien de Troyes frequently champions la brisure du couplet by which he achieves a hitherto unprecedented metrical elasticity in his narratives:

41

For a considerably more detailed treatment of enjambment, see Chapter 5. This relative scarcity of enjambment in French versification is, of course, directly related to the point made above about the crucial role played by phrase-terminal stress in French: since lexical stress in French is largely subdued by phrase-terminal stress, it tends to undermine the phonological integrity of a verse-line to have anything but a phraseterminal word in line-terminal position. 43 Preminger & Brogan (eds) (1993): 359-360, s.v. ‘Enjambment’. 44 Vising (1923): 85. 42

76

Lors ne lessa mie cheoir La reïne ses ialz vers terre; (Le Chevalier de la charrette, ll. 4478-4479)45 In this extract, which is from the scene in which Lancelot and Guinevere are reconciled, and Guinevere refrains from lowering her eyes to the ground at the sight of Lancelot, the verb (‘cheoir’) is separated from its grammatical subject (‘La reïne’) across a line-break which also functions as a break between couplets. This kind of metrical audacity would have been frowned upon by earlier Anglo-Norman practitioners of the octosyllabic couplet, and by the time that verse design was first borrowed into English, the tendency to use end-stopped lines to the exclusion of enjambed ones was borrowed alongside with it. This, of course, is not to say that each and every octosyllable is unequivocally end-stopped in either Anglo-Norman or Middle English verse. It is certainly possible to find examples of couplets in which some degree of enjambment is involved; here, for instance, is one from Sir Orfeo that also involves a queen: And wel sone this fair quene Fel on slepe opon the grene. (Sir Orfeo, ll. 71-72) In this couplet, the grammatical subject is cut off from the verb phrase in a manner that can only be described as enjambed. But again, it must be emphasized that this is a rare instance, an anomaly, and that the overwhelming majority of octosyllabic couplets in Sir Orfeo and other English lais and romances do not employ enjambment at all. Even more importantly: when some degree of enjambment does occur in these texts, it usually occurs only within couplets and virtually never between couplets. We are now in a position to summarize the state of the Middle English octosyllabic couplet at the time when it was picked up by Chaucer, probably around 1370. From the Anglo-Norman tradition had been adopted the combined principle of keeping both caesurae and enjambments to a minimum, to the effect that lines and couplets closely corresponded to the over-all syntactical segmentation of the text. However, Middle English practitioners of the octosyllable differed from their Anglo-Norman colleagues by failing to observe a strict syllable count: early Middle English 45

Cited from Chrétien de Troyes, Le Chevalier de la charrette (Lancelot), texte établi, annoté et présenté avec variantes par A. Foulet et K.D. Uitti (Paris: Bordas, 1989): 252.

77

couplet lines, such as those in King Horn, could contain as few as four or five syllables, and only gradually did lines of seven, eight and nine syllables come to dominate, but never to such an extent that the ratio of lines containing exactly eight syllables exceeded 60 per cent.46 To the extent that metrical versification always entails a compromise between phonological and syntactical features of a text, it may thus be said that while early English imitators were keen to imitate the syntactical features of the Anglo-Norman octosyllable, they were decidedly more indifferent when it came to observing its phonological features. Chaucer turned this way of handling the octosyllabic couplet upside down. We have already touched upon Chaucer’s successful combination of a stricter syllable count with a more consistently alternating stress profile in The Book of the Duchess, and we may add that this development was continued and refined in The House of Fame. However, Chaucer’s unique, and indeed founding, contribution to a distinct English system of versification lies not so much in his efforts to tighten up the phonological features of the octosyllabic couplet – that feat was accomplished with far more zeal by John Gower in Confessio Amantis – but rather in the syntactical freedom he gained by doing so. By tightening up the phonological features of the metre, Chaucer can be said to have stabilized the phonological integrity of each line’s sound-shape to such an extent that a whole new range of syntactical possibilities was laid open. These new possibilities are taken advantage of from the very beginning of the Proem to The Book of the Duchess: I have gret wonder, be this lyght, How that I lyve, for day ne nyght I may nat slepe wel nygh noght; I have so many an ydel thoght Purely for defaute of slep That, by my trouthe, I take no kep Of nothing, how hyt cometh or gooth, Ne me nys nothyng leef nor looth. (The Book of the Duchess, ll. 1-8) Within the first couplet and a half, Chaucer here manages to set up an expectation in the reader that these octosyllables are going to be structured around a medial metrical caesura, only to disappoint 46

Tarlinskaja (1976): 256; Table 17.

78

that expectation again at the earliest possible moment. Of course, no poet of sound mind would seriously consider making a metrical caesura part of an octosyllabic verse design for a lengthy poem in English: the octosyllable is simply too short a measure for this kind of metrical subsegmentation to work, and the metre would be prone to fall into sing-song if it were attempted. But in the opening couplet of The Book of the Duchess Chaucer nonetheless allows himself the freedom to suggest exactly such a verse design by placing a quite strong medial caesura in both lines of the initial couplet. This immediately creates a very compelling sense of rhythmical closure and selfcontainment in that couplet, and sets up the expectation in the reader of a fully realized ‘4 X 4 structure’.47 However, already in the transition from line 2 to line 3, it becomes apparent that the adverbial which makes up the second half of line 2, ‘for day ne nyght’, does not modify the verb phrase which precedes it, but rather the one which comes after, in the next line. This is made clear by the editor’s helpful punctuation of the text where no mark of punctuation has been inserted at the end of line 2. However, the fact that the line break between lines 2 and 3 is unpunctuated does not necessarily entail enjambment. While it is true that enjambment for obvious reasons precludes any occurrence of line-terminal punctuation, the absence of punctuation at line ends is not in and by itself a sufficient criterion for diagnosing enjambment. In this particular instance, I will argue that the strictest criterion for enjambment – namely that a mark of punctuation cannot under any circumstances be inserted with reference to syntax – is not met, and that we should fare better by viewing Chaucer’s transition from line 2 to 3 simply as a run-on line. Even so, this is no way to treat an octosyllabic couplet: Chaucer blatantly violates the syntactical unity of the couplet while seemingly affirming its underlying rhythmical structure through his use of cleverly positioned caesurae. This feat is hardly the endearing result of a metrical blunder that can be explained away with reference to the poet’s lack of experience: Chaucer thus performs an equally daring variant of the selfsame trick in the transition from line 6 to line 7, and similar examples can be found throughout the rest of the poem. The startling novelty of Chaucer’s innovative use of syntax in relation to caesurae and couplets at the beginning of The Book of the Duchess can best be appreciated if we compare it to the French poem on which Chaucer’s Proem is modelled, namely Jean Froissart’s Le Paradys d’Amour. Even a scanty comparison between the opening lines of these two poems demonstrates simultaneously Chaucer’s thematic debt to his model and his metrical independence from it:

47

See Attridge (1982): 76-122.

79

Je sui de moi en grant merveille

I have gret wonder, be this lyght,

Comment je vifs, quant tant je veille,

How that I lyve, for day ne nyght

Et on ne point en veillant

I may nat slepe wel nygh noght;

Trouver de moi plus travaillant:

I have so many an ydel thoght

(Le Paradys d’Amour, ll. 1-4)48

(The Book of the Duchess, ll. 1-4)

If we allow ourselves to consider Chaucer’s treatment of the octosyllabic couplet in The Book of the Duchess as more than the half-baked attempts of an inexperienced poet to calque onto his mother tongue the virtues of French versification, it becomes evident that Chaucer’s poem in fact represents a defining moment in the history of English versification. Through his careful manipulation of caesurae, Chaucer pries open both his lines and his couplets and leaves both of these formal constituents susceptible to all the corrupting and alluring effects of run-on lines and enjambments. In so doing, Chaucer departs significantly from all of his predecessors, in French and English alike: whereas they had sought to adapt their syntactical phrasing to meet the requirements of the metre, Chaucer deliberately coerces syntactical phrasing and metre into a highly charged equilibrium of mutual subversion, or at least he reserves the right to do so whenever it suits his purposes. Viewed in this light, The Book of the Duchess is the earliest poem in English by a major writer to exemplify the sort of highly charged interplay between metre and syntax which so many prosodists of the twentieth century have hailed as being at the heart of English versification. ❦ The oppositional interplay between metre and syntax which characterizes Chaucer’s use of the octosyllabic couplet is, of course, much more typical of – and much easier to achieve in – verse that is based on the slightly longer decasyllabic line.49 Credit for introducing this line type in English versification also falls to Chaucer, and it is primarily for his handling of the decasyllable that Chaucer has been hailed as the father of English poetry at least since the time of John Dryden. 50 48

Cited from E.-G. Sandras, Étude sur Chaucer (Paris: Auguste Durand, 1859): 295. On Chaucer’s indebtedness to Froissart, see James I. Wimsatt, Chaucer and His French Contemporaries: Natural Music in the Fourteenth Century (Toronto: University of Toronto Press: 1991): 174-209. 49 See Attridge (1982): 132-138. 50 Dryden thus bestows on Chaucer this epithet in the Preface to Fables Ancient and Modern; see John Dryden: The Poems and Fables, edited by James Kinsley (London: Oxford University Press, 1961): 528.

80

However, like its octosyllabic counterpart, the decasyllable has a history in the Romance vernaculars prior to its being modified to fit the exigencies of the English language. To unravel this history, we must once again turn our gaze towards France where the decasyllable made its entry in the first half of the eleventh century, almost a century after the advent of the octosyllable. Among the earliest instances of the decasyllable in French poetry are the 125 quintains of La Vie de St. Alexis, of which here is one, taken from the section in which Alexis’s wife is mourning the death of her husband: O kiers amis, de ta juvente bela! Ço peiset mai que purirat [en] terre. E! gentils hom, cum dolente puis estra! Jo atendie de te bones novelses, Mais or les vei si dures e si pesmes! (La Vie de St. Alexis, ll. 476-480)51 Even without any training in Old French, it is fairly easy to detect in this stanza the presence of a metrical caesura after the fourth syllable: in two of the lines, the caesura is marked by a comma, and in the three remaining lines there is a word boundary after the fourth syllable position followed by a preposition or a conjunction. This position of the caesura, by which the verse is sliced in a proportion of 4 + 6 syllables, is by far the most common in the French decasyllable. Lines with the reverse proportion of 6 + 4 syllables also occur, though much less commonly, and even the rare division of the decasyllable in the proportion of 5 + 5 syllables can be found if one knows where to look.52 However, it is important to understand that these different types of lines are never combined within the framework of a single poem: decasyllabic poems in French always employ the same caesural proportion in every single line. Keen observers will already have noticed that none of the five lines quoted above consists of ten syllables; they all contain eleven. Even so, they are perfectly regular instances of the French decasyllable insomuch as French metricists count the syllables only up to and including the final stress-carrying syllable of each hemistich, after which an additional unstressed syllable is allowed. At this point it may be prudent to comment on the exact status of lexical stress in French. Strictly speaking, lexical stress is only a feature of the very earliest period of Old French, and most scholars 51 52

Cited from Wendy Ayres-Bennett, A History of the French Language through Texts (London: Routledge 1996): 41. Preminger & Brogan (eds) (1993): 276, s.v. ‘Decasyllable’.

81

agree that by the time when the texts with which we are concerned here began to emerge, phrasal stress had already begun to replace lexical stress as the dominant prosodic feature of French. 53 This, however, does not mean that word stress is completely absent in French: it only means that word stress in French is not phonemic, for which reason it is so easily subdued by phrasal stress. This is in stark contrast to English where word stress is indeed phonemic, i.e. capable of distinguishing two otherwise identical words from one another, cp. /ˈdɪsʧaːʤ/ (discharge, noun) vs. /dɪsˈʧaːʤ/ (discharge, verb). Such minimal pairs, in which stress is the only differentiating feature, are not found in French; even so, French words of more than one syllable do in fact come with a fully predictable phonetic (as opposed to phonemic) stress pattern, which can easily be discerned if the word is pronounced in isolation. Phonetic stress in French – Old French and Modern French alike – thus always falls on the last syllable of a word, provided that the vowel of that syllable is not schwa, in which case stress falls on the preceding syllable.54 Words that have stress on the last syllable are called oxytones; words that are stressed on the syllable before the last syllable are called paroxytones. The distinction between oxytones and paroxytones is, of course, what makes possible the important distinction between masculine and feminine line-endings in French versification, by which lines terminating in an oxytone are called masculine, and lines which terminate in a paroxytone are called feminine. The same distinction is also observed in the French octosyllable where masculine couplets are frequently made to alternate with feminine couplets in regular patterns. However, in the case of longer measures, such as the decasyllable, there is cause to be doubly aware of the distinction between masculine and feminine endings in that this distinction also applies to the end of the first hemistich of a verse line. That is to say that the constitutive function of a metrical caesura is so pronounced in early French decasyllables that both hemistichs are accorded the same freedom of terminating in either a stressed syllable (oxytone) or a stressed syllable followed by an unstressed extra-metrical syllable (paroxytone).55 This, of course, means that a decasyllable may contain as many as twelve syllables, provided that both hemistichs have feminine endings: 53

E.g. Pope [1932] (1973): 82; Rebecca Posner, Linguistic Change in French (Oxford: Oxford University Press, 1997): 229. 54 Frederick B. Agard, A Course in Romance Linguistics, Volume 2: A Diachronic View (Georgetown University Press: 1984): 35. For a more detailed exposition of the history and status of word-final schwa in French, see Posner (1997): 266-273. 55 By ‘extra-metrical’ is meant that an unstressed syllable at the end of a hemistich is not felt to influence the rhythm of the verse, for which reason it is left out of the metrical count. On this, see Bruce Hayes, Metrical Stress Theory: Principles and Case Studies (Chicago: University of Chicago Press, 1995): 56-60.

82

Vint la pulcele ‖ que il out espusede. (La Vie de St. Alexis, l. 467)56 However, it is important to realize that in lines like this, the unstressed syllable immediately before the caesura is extra-metrical in the sense that it is not counted as occupying a metrical syllable position. The first hemistich is simply treated as if it had only four syllables instead of five; this means that the first syllable of the second hemistich is counted as occupying the fifth metrical syllable position, and it is on the basis of this count that the crucial tenth syllable position is computed. If we adopt the conventional French practice of using a capital F to denote a feminine ending and a capital M to denote a masculine ending, the structure of the line quoted above can be formalized as 4F + 6F, whereas the five lines from St. Alexis we quoted before can all be formalized as 4M + 6F. The most common template of the French decasyllable – i.e. the 4 + 6 template – thus comes in four distinct versions: 4M + 6M (10 syllables), 4M + 6F (11 syllables), 4F + 6M (11 syllables) and 4F + 6F (12 syllables). By now, it ought to be clear why we previously described the presence of a metrical caesura in the decasyllable as opposed to its absence in the octosyllable as a crucial structural difference in the verse design: whereas the octosyllable comes in only two different versions – i.e. 8M and 7F – the decasyllable, even in its most common template (4 + 6), comes in four different versions, to which can be added the four different versions which follow from the 6 + 4 template. Decasyllabic lines in which the first hemistich takes the form 4F are said to have an epic caesura in that such lines are particularly common in early narrative works,57 such as in this example, the very last line of La Chanson de Roland: Ci falt la geste ‖ que Turoldus declinet. However, as we have already seen, the decasyllable was eventually displaced by the octosyllable as the preferred medium for narrative, epic verse in French, a process which saw its fullest realization

56 57

Cited from Ayres-Bennett (1996). Gasparov (1996): 125.

83

in Chrétien de Troyes’s innovative use of the octosyllabic couplet for that purpose.58 Liberated from the shackles of its epic responsibilities, the decasyllable was soon picked up by Occitan poets in the mid-twelfth century and employed for verse in a distinctly lyric mode: this is where Ezra Pound’s much revered ‘Provençal troubadours’ enter the stage.59 Gasparov points out that this turn of events led to a rather unusual state of affairs as far as verse goes, in that the shorter line-form now, at least temporarily, became associated with narrative poetry while the longer line-form became associated with lyric poetry: from a typological point of view it is usually the other way around. 60 Gasparov furthermore points out that in order to accommodate this change of scene for the decasyllable, an important structural concession had to be made regarding its traditional 4 + 6 template. Whereas the caesura in epic works was determined primarily by accentual concerns (stress had to fall on the fourth syllable, after which an unstressed, non-metrical syllable was allowed), the caesura in its new lyric setting came to be determined solely by syllabic concerns in that an obligatory word boundary between the fourth and the fifth syllable was introduced.61 This, of course, meant that lines with an epic caesura were barred from occurring in Occitan lyric verse, but it also meant that lexical stress came to play an even more marginal role in Occitan versification than it had done in Old French versification. The most immediate consequence of the introduction of an obligatory word boundary after the fourth syllable in the Occitan lyric decasyllable was therefore that paroxytones could no longer be placed in such a manner that their stress-bearing penultimate syllable coincided with the fourth syllable position of the line: this would leave the paroxytone’s final syllable hanging at the fifth syllable position. Instead – whenever a paroxytone was needed immediately before the caesura – the paroxytone had to be moved back one position in the line so that its final, unstressed syllable came to occupy the fourth syllable position: all of this to avoid violating the rule prescribing a word boundary before the fifth syllable. This phenomenon can be observed in the following line by Peire Vidal, the poet whom Ezra Pound, not without envy, gave credit for having been ‘the fool par excellence of all Provence’:62 58

Indeed, Chrétien de Troyes’s romances are frequently cited as the earliest specimens of that type of verse which would later evolve into the novel. 59 In accordance with current scholarly practice, I use the word Occitan to refer to that language and culture which Pound referred to as Provençal. Provençal, Frede Jensen informs us, is a far too restricted term in both its linguistic and geographical connotations; see F.R.P. Akehurst & Judith M. Davis (eds), A Handbook of the Troubadours (Berkeley & Los Angeles, CA: University of California Press, 1995): 349. 60 Gasparov (1996): 128. 61 Gasparov (1996): 129. 62 Ezra Pound, Personæ. Collected Shorter Poems, edited by Lea Baechler & A. Walton Litz (London: Faber and Faber, 2001): 28.

84

a nos cortes ‖ es trebalhs e dolors; (‘Bels Amics cars, ven s’en ves vos estius’, l. 17)63 This type of caesura, which may be formalized as 3F + 6M/F, is aptly called a lyric caesura, and both Gasparov and Duffell explain its emergence in Occitan lyric poetry with reference to the complex melodies to which these lyrics were set by their makers.64 These melodies – many of which are preserved to this day in manuscripts – were so fixed in terms of syllable count that extrametrical syllables could not be tolerated immediately before the caesura, and as a result paroxytones in this position were moved back one position so that they came to be pronounced with a wrenched accent.65 That lexical stress occasionally must yield to accommodate the exigencies of melody is neither unique to nor particularly characteristic of the Occitan troubadour tradition. William Beare succinctly points out that ‘when clash [between the fixed rhythm of polysyllables and the regular beat of the music] occurs, the music wins’, and that ‘song is not the perfection of verse; it is a different and in some ways cruder, more mechanical thing’.66 Even in English where word stress is both prominent and phonemic, the natural word stress will habitually yield to the force of music: a salient example of this occurs in the song ‘My Favorite Things’ from Rodger and Hammerstein’s last musical The Sound of Music. In that song, the 3/4 time signature of the music is so pronounced that the lexical stress of the word eyelashes is wrenched from the first syllable, [ˈaɪlæʃɪz], to the second syllable, [aɪˈlæʃɪz], because it is the second syllable of that word which happens to coincide with the downbeat of the music’s time signature (here marked with an accent aigu): Snówflakes that stáy on my nóse and eyeláshes67 In this particular setting, the wrenched accent of eyelashes may be said to contribute to the over-all air of innocence which permeates the song in general, but wrenched accents in English verse, 63

Cited from Veronica M. Fraser, The Songs of Peire Vidal (New York: Peter Lang, 2006): 187. On the relationship between the poetic and musical structures of Occitan troubadour verse, see Elizabeth Aubrey, The Music of the Troubadours (Bloomington, IN: Indiana University Press: 1996): 194-197. 65 See Gasparov (1996): 129, and Martin J. Duffell, ‘“The Craft So Long to Lerne”: Chaucer’s Invention of the Iambic Pentameter’, in The Chaucer Review, Vol. 34, No. 3 (2000), 269-288: 270. 66 Beare (1957): 26. But see also M.L. West, Greek Metre (Oxford: Oxford at the Clarendon Press, 1982) for evidence that this may not have been the case in Greek (21). 67 Cited from Oscar Hammerstein II, Lyrics, edited by William Hammerstein (Milwaukee, WI: Hal Leonard Books, 1985). 64

85

whether sung or recited, are most commonly associated with comical genres (cp. Gilbert & Sullivan), if not simply viewed as signs of poor craftsmanship. That this is so may in part be due to the fact that word stress is so prominent in English that a wrenched accent will always stand out as unnatural,68 but even in French where the role of word stress is minimal, wrenched accents, such as those caused by the lyric caesura, seem to have been regarded as less than optimal. The lyric caesura was thus never employed in Occitan lyric verse to the same extent as the epic caesura was employed in Old French narrative verse, and by the sixteenth century the lyric caesura disappears from use completely.69 However, the fact that lyric caesurae did emerge among the Occitan troubadours at some point is important in that it is an indication of the continuous progression towards that purely syllabic type of versification which would later come to characterize French poetry in general. This sets off very clearly the Occitan decasyllable from its origins in Medieval Latin versification, in which stress played a far more important role than it subsequently came to do in French. 70 Even more significant, though rarely commented on, is the fact that the emergence of the lyric caesura is also the first indication of a development which in effect destabilized the structural, constitutive function of the caesura in Occitan. Whereas in the Old French decasyllable, caesurae and linebreaks were treated as equals inasmuch as both of these junctures allowed the addition of an extrametrical syllable, it was only the line-break of the Occitan decasyllable which was allowed this freedom. In formal notation, this difference between the Old French and the Occitan decasyllable can be formalized like this: Old French epic decasyllable:

4M/F

+ 6M/F

Occitan lyric decasyllable:

3F/4M + 6M/F

This demotion of the caesura in Occitan was the first step towards establishing a strict hierarchy between line-ends and caesurae in which the structural weight of the caesura was firmly subordinated to that of the line break. This hierarchy gave prominence to the line at the expense of the hemistich, and as early as the mid-twelfth century Occitan poets seem to have regarded the decasyllable as a unified line with a 10M/F structure rather than a bipartite conglomerate of

68

Duffell (2008): 46. Gasparov (1996): 129. 70 Gasparov (1996): 88-118. 69

86

equipollent hemistichs.71 This subtle shifting in the conceptualization of the decasyllable was to have important ramifications for European versification over the next centuries. The Occitan troubadours did not confine their minstrel activities to the region of Provence. Peire Vidal, for instance, travelled as far as Hungary, and in 1194-95 he is known to have been at the court of the Marquis of Montferrat in Italy. Vidal was only one of several troubadours to pay extended visits to Italy, and as a result of this cultural exchange a whole school of Italian troubadours, of whom the most famous is Sordello, arose in the North of Italy in the early thirteenth century.72 But Occitan verse had an even more important bearing further south where crucial developments of Italian versification were taking place among a group of poets we nowadays refer to as the Sicilian School. In the aftermath of the Albigensian Crusade, Occitan poets found themselves forced to seek patronage outside of the langue d’oc area, and they found it at the court of Frederick II in Palermo where they came to exert a considerable influence on the local poetic scene.73 Unlike the troubadours of Northern Italy, who adopted the Occitan language of their teachers, the poets of the Sicilian School used Italian exclusively. They also differed from their Northern colleagues in that the Sicilians did not compose any melodies, for which reason we must assume that their poems were not usually sung.74 That the poetry of the Sicilian School was not necessarily set to music may well have been the decisive factor that made possible the cultivation of those strikingly complex stanza forms which more than anything else distinguish the Sicilian School.75 It was here that two of the most important Italian stanza forms were first conceived during the first half of the thirteenth century: the sonnet and the canzone.76 Both of these verse forms take as their primary structural constituent a line-type that owes much to the Occitan decasyllable, namely the Italian hendecasyllable. We should pause in order to clarify an important issue pertaining to terminology. Whereas in French metrical tradition, the ten-syllable line is named after that version which terminates in an oxytone (i.e. 10M), the Italian metrical tradition names the ten-syllable line after

71

Duffell (2000): 270-271. Hence, the later term vers de dix to designate this verse design in French. On Vidal’s stay in Montferrat and other troubadours visiting the courts of Northern Italy, see Akehurst & Davis (eds) (1995): 295-306. 73 Occitan poets also found a safe haven at the Castilian court where they exerted a considerable influence on Galician verse. 74 Akehurst & Davis (ed.) (1995): 280. 75 According to Ronald Martinez, this proposition was first advanced in 1975 by the eminent Italian philologist Aurelio Roncaglia (see Akehurst & Davis (eds) (1995): 280, n. 6). 76 See Michael R.G. Spiller, The Development of the Sonnet: An Introduction (London: Routledge, 1992): 11-28. 72

87

that version which terminates in a paroxytone (i.e. 10F).77 This terminological gallimaufry should, however, not be allowed to obfuscate the plain fact that the Italian endecasillabo and the French decasyllabe are simply two different words for the same basic verse design, namely a line of ten syllable positions with an obligatory stress on the tenth position, after which one or two optional unstressed syllables are allowed. The difference in terminology rather reflects a difference in the phonological structure of the two language’s lexicon: whereas the French lexicon at this point in time was rich in both oxytones and paroxytones, the phonological structure of the Italian lexicon was, and still is, characterized by being predominantly paroxytonic. For this reason a regular Italian ‘decasyllable’ will almost always contain eleven syllables; hence the name hendecasyllable: a line of ten syllables plus one. We have so far been content with simply noting when and where the earliest instances of decasyllabic lines occur in the vernaculars, as if this verse design appeared out of nowhere. This, of course, is not the case. Gasparov convincingly argues that both the French decasyllable and the Italian hendecasyllable owe their success and viability to the fact that they are ultimately modelled on the Medieval Latin dodecasyllable,78 of which this is an example from an anonymous ninthcentury poem from Modena: O tu, qui servas ‖ armis ista moen[ï]a, Noli dormire, ‖ moneo, sed vigila.79 Gasparov’s helpful scansion, by which stressed syllables are marked in bold, immediately alerts us to the fact that both of these lines terminate in a word that is neither oxytonic nor paroxytonic, but rather in one that is proparoxytonic, i.e. stressed on the antepenultimate syllable. Due to the highly inflected nature of Latin, proparoxytones became quite common when word stress replaced vowel quantity in that language around the fourth century. However, when passing into the Romance vernaculars, words with this structure were usually shortened; Ernst Robert Curtius comments on this in one of his erudite footnotes:

77

To make the confusion complete, the Galician metrical tradition follows the French model, and the Spanish metrical tradition follows the Italian model. 78 Gasparov (1996): 122-126. 79 Cited from Gasparov (1996): 104.

88

On the way from Latin to French the penultimate syllable of the proparoxytone succumbed. Mallarmé was so touched by this, that he wrote a prose-poem on the ‘Death of the Penultimate’ (Le Démon de l’analogie in Divagations). It ends: ‘Je m’enfuis, bizarre, personne condamnée à porter probablement le deuil de l’inexplicable Pénultième.’ Grammar too has its tragedies.80 Grammar does have it tragedies, but unlike the humanitarian tragedy in whose shadow Curtius penned this observation in the 1940s, grammatical tragedies are not entirely devoid of redeeming qualities: metrical innovation often thrives in the ruins of grammatical tragedy. To readers who find the kind of subtle linguistic changes with which we are concerned here marginal and irrelevant to the study of poetry, Curtius’s footnote is also a reminder that the modernists viewed such matters differently: they were indeed mindful of the historical relationship between grammar and metrics and the ways in which changes in the former area were likely to affect changes in the latter. In addition to the fact that two unstressed syllables were required after the tenth syllable in the Latin dodecasyllable, there is another important feature which sets off this verse design from its later off-springs in the Romance vernaculars. Thus, in the Latin dodecasyllable, the caesura invariably falls after the fifth syllable, usually – but not exclusively81 – in such a manner that word stress coincides with the fourth syllable position, as in the two lines cited above. This position of the caesura is of course the same as that of the epic caesura in French (4F + 6M/F), but whereas the epic caesura was always optional in French, the word boundary after the fifth syllable in Latin was obligatory to the same extent as the word boundary after the fourth syllable came to be in Occitan. In the Italian hendecasyllable these two principles came into conflict. In order to appreciate the nature of this conflict and its importance to the type of verse which would later develop in English, it may be prudent to recapitulate what we have said so far about the caesura in the decasyllable. When the Old French epic decasyllable was adopted for lyric purposes by the Occitan troubadours, the troubadours introduced an obligatory word boundary after the fourth syllable, by which the caesura came to be determined exclusively by syllabic concerns. As a result of this, the structural force of the caesura was in time undermined to such an extent that the line came to be seen more as a single structural unit than as a conglomerate of two equipollent hemistichs. This demotion of the caesura in the prosodic hierarchy was most likely the outcome of the fixed nature of the melodies to 80

Curtius (1953): 156, n. 33. Cp. this line out of the previously cited anonymous poem from Modena: ‘Prima quiete ‖ dormiente Troïa,’ in which stress falls on the third syllable (also cited from Gasparov (1996): 104).

81

89

which the Occitan troubadours set their lyrics. However, when the troubadours eventually found a safe haven in Palermo, these melodies were given up alongside the Occitan language, enabling the local poetic scene to develop a distinct species of Italian poetry characterized by elaborate stanza designs, the majority of which relied heavily on the hendecasyllable. The Italian hendecasyllable, as it developed among the poets of the Sicilian School, was in some respects more indebted to the structural characteristics of the Medieval Latin dodecasyllable than was the decasyllabic line of the troubadours: of all the major Romance literatures, Gasparov observes, ‘Italian poetry incorporated the heritage of Latin verse in the most direct and straightforward way’.82 This is hardly surprising: geographically and linguistically there are greater overlaps between Latin and Italian than between Latin and any other major language. However, as far as Sicily is concerned, the impact of the dodecasyllable was double: in addition to being familiar with the Latin species of this verse design, the poets of Sicily are likely to have been influenced also by the Byzantine species of the dodecasyllable. Ever since Antiquity, ties between Greece and Sicily had been strong – Archimedes, Gorgias and Empedocles were all natives of Sicily – and during the Gothic War (535-545) Sicily was the first part of Italy to fall to the Byzantine Empire. This historical presence of Greek and Byzantine culture in Sicily may well have made the poets that we now associate with the Sicilian School both exposed and susceptible to the influence of Byzantine versification. The case of the Byzantine dodecasyllable provides us with a particularly salient example of a metrical innovation that blossomed amidst the ruins of a grammatical tragedy, namely the loss of distinction between long and short syllables in Greek.83 It is with this metre that M.L. West concludes his authoritative Greek Metre (1982), arguing that the emergence of the Byzantine dodecasyllable brought to a close two thousand years of quantitative verse and cleared the way ‘for new patterns to form, based on the contrast of accented and unaccented’.84 The quantitative pattern of which the Byzantine dodecasyllable was a development is clearly that metre which Aristotle describes as ‘the most speakable of all metres’,85 namely the Greek iambic trimeter.86 This metre also consisted of twelve syllables, but – unlike its immediate successors – the iambic trimeter was segmented into three clearly defined units called metra, hence the designation trimeter. Metra, in 82

Gasparov (1996): 122. See Paul Maas, ‘Der byzantinische Zwölfsilber’ [1903], in Kleine Schriften (München: C.H. Beck’sche Verlangsbuchhandlung: 1973): 242-288. 84 West (1982): 185. 85 Aristotle, Poetics, 1449a, in D.A. Russell & Michael Winterbottom (eds), Classical Literary Criticism (Oxford: Oxford University Press, 1989): 56. 86 Gasparov (1996): 104. 83

90

this context, are not to be confused with feet: an iambic metron in Greek prosody consists of four syllables alternating between short and long (but in such a manner that the first syllable of each metron may be either short or long),87 whereas an iambic foot, such as we know it from the vernacular prosodies, consists, of course, of only two syllables.88 This distinction is, however, of limited relevance to our immediate purposes: in fact, it is the discarding of the sub-segmentation into metra which more than anything else sets apart the Byzantine dodecasyllable from the iambic trimeter.89 The segmentation into metra of the iambic trimeter was determined exclusively by the need to alternate short and long syllables according to predetermined patterns and had nothing to do with word boundaries. That, however, is not to say that caesurae were not in play in the iambic trimeter: this metre did in fact observe a fixed caesural word boundary in each line, either after the fifth or after the seventh syllable. This caesural segmentation of the iambic trimeter was – in contradistinction to the segmentation of that metre into metra – retained in the Byzantine dodecasyllable, as West succinctly explains: ‘So far as the ear was concerned [after quantity had ceased to be significant], all that was left was the count of syllables—five + seven or seven + five ... —and a habitual cadence now marked by accent’.90 Even more importantly, lines of 5 + 7 syllables could be mixed with lines of 7 + 5 syllables within the same poem, as in these lines from a rendering of Aesop’s Fables: ὄνος δέ ποτε ‖ ἐμακάριζεν ἵππον διὰ τήν τροφὴν αὐτου ‖ καὶ θεραπείαν 91 The movable caesura of the Byzantine dodecasyllable completes the picture of the various prosodic models that came into play in the mélange that was Sicilian poetry during the Hohenstaufen dynasty. From the Occitan lyric decasyllable came a verse design with a fixed caesura after the fourth syllable, and from the Latin dodecasyllable came a verse design with an equally fixed caesura after the fifth syllable. In both of these metres the tenth syllable position invariably coincided with a lexical stress after which only unstressed syllables could occur. In the Byzantine 87

Such a syllable, whose length is optional, is aptly termed anceps and is conventionally annotated as x in formal notation. The iambic metron thus has the form x‒u‒. 88 On the distinction between the classical metron and the vernacular foot, see Preminger & Brogan (eds) (1993): 787788, s.v. ‘Metron’. 89 Maas [1903] (1973): 242-247. 90 West (1982): 185. 91 Cited from Gasparov (1996): 105 [my markings of the caesuras].

91

dodecasyllable it was usually the eleventh syllable position that enjoyed this degree of fixedness, but more importantly poems in this metre were allowed to mix lines with a caesura after the fifth syllable with lines in which the caesura came after the seventh syllable. This principle of operating with a movable caesura within the framework of a single poem was adopted by the Sicilian School in their codification of the hendecasyllable, but in such a manner that the caesura could occur either after the fourth or after the sixth syllable. In the late thirteenth century, the Sicilian hendecasyllable was picked up members by the Tuscan School, founded by Guittone d’Arezzo, and, even more significantly, by Guidi Guinizelli, the founder of the dolce stil nuovo. This style of courtly poetry would find its most famous practitioner in Dante Alighieri, and it is indeed in Dante’s Purgatoria that we find the earliest occurrence of the term dolce stil nuovo: it occurs when the Wayfarer – presumably Dante himself – encounters the poet Bonagiunta da Lucca, who, on being told that the Wayfarer writes verse dictated by the breath of love, responds: ‘O brother, now I see ... the knot Which held back the Chief Clerk, Guittone and me, From the sweet new style of writing that I hear of.92 The ‘sweet new style’ is, of course, the dolce stil nuovo; the ‘Chief Clerk’ is the Sicilian poet Jacopo da Lentini, who invented the sonnet; and ‘Guittone’ is the aforementioned founder of the Tuscan School. By letting Bonagiunta respond in this manner, Dante manages to profile himself against both his Sicilian and his Tuscan predecessors, though it is from them that he must have learned the principle of operating with a movable caesura. The movable caesura as it gradually came to be codified in Italian poetry differed from its Byzantine origin in one crucial respect. Whereas the position of the Byzantine movable caesura had been determined on a purely syllabic basis – i.e. either after the fifth or after the seventh syllable – the position of the Italian movable caesura came to be determined primarily by stress. In Italian it was thus word stress that was fixed to either the fourth or the sixth syllable, allowing for the addition of an extra syllable before the caesura whenever a paroxytone occurred in pre-caesural position. The first hemistich of an Italian hendecasyllable could thus take any of four different basic forms, exemplified here by four lines out of the first canto of Dante’s Inferno:

92

Dante Alighieri, The Divine Comedy, translated by C.H. Sisson (Oxford: Oxford University Press: 1980), Purgatoria, XXIV, ll. 55-57.

92

4M (four syllables):

mi ritrovai ‖ per una selva oscura (l. 2)

4F (five syllables):

ed una lupa, ‖ che di tutte brame (l. 49)

6M (six syllables):

‘Miserere di me,’ ‖ gridai a lui (l. 65)

6F (seven syllables): che la verace via ‖ abbandonai (l. 12)93 In the formal notation above, I have for the purpose of consistency retained the French practice of designating the hemistich structure by numbering the position of the final stressed syllable, followed by an indication of whether that syllable is succeeded by an unstressed syllable (F) or not (M). However, in Italian metrics, lines in which the first hemistich takes the forms 4M and 4F are, perhaps more eloquently, said to be a minore, whereas lines in which the first hemistich takes the forms 6M and 6F are said to be a maiore.94 When first hemistichs end in an oxytone, i.e. 4M or 6M, they are termed tronco (curtailed), whereas those ending in a paroxytone, i.e. 4F or 6F, are termed piano (soft). In the rare event that a first hemistich terminates in a proparoxytone, such a hemistich is termed sdrucciolo (gliding).95 The selfsame triad of terms – tronco, piano and sdrucciolo – also applies to the line in its entirety, depending on whether it ends in an oxytone, a paroxytone or a proparoxytone. But whereas the word-final unstressed syllables of paroxytones and proparoxytones are extra-metrical when they occur at line-ends in Italian, such syllables do form part of the metre when they occur in pre-caesural position. That is to say that every syllable of the first hemistich of an Italian hendecasyllable occupies a metrical syllable position and must be taken into account when computing the position of the pivotal tenth syllable.96 In this, the Italian piano/sdrucciolo caesura diverts significantly from the Old French epic caesura (4F + 6M/F), in which unstressed syllables in pre-caesural position are blatantly left out of the metrical count. The difference may best be illustrated by a comparison between two lines we have already quoted; the first is from La Chanson de Roland, the second is from Inferno:

93

Cited from Dante Alighieri, The Divine Comedy, Vol. 2, Inferno, Italian text with translation and comment by John D. Sinclair (New York: Oxford University Press, 1939). 94 Spiller (1992): 205, n. 15. 95 Walter Thomas, ‘Milton’s Heroic Line Viewed from an Historical Standpoint’, in The Modern Language Review, Vol. 2, No. 4 (Jul., 1907): 289-315 (293-294). 96 Duffell (2000): 279-280.

93

1 2 3 4 ‖ 5 6 7 8 9 10 x x x x (x) ‖ x x x x x x (x) (5 + 7) Ci fait la geste ‖ que Turoldus declinet (4F + 6F) 1 23 4 5 x xx x x ed una lupa,

‖ 6 7 8 9 10 ‖ x x x x x (x) ‖ che di tutte brame

(5 + 6) (4F + 5F)

In both of these lines, the first hemistich clearly takes the form 4F, but whereas the second hemistich in the Old French line takes the form 6F, it takes the truncated form 5F in the Italian line.97 Every pronounced syllable in the first hemistich of an Italian hendecasyllable is thus relevant in the larger metrical scheme and therefore influences the structure of the second hemistich to a larger extent than is the case in Old French where each hemistich is treated as an autonomous structure in its own right. This difference between the Italian hendecasyllable and the Old French decasyllable represents a further step in the consolidation of the decasyllabic line as a unified 10M/F structure as opposed to that bipartite conglomerate of equipollent hemistichs which is characteristic of the Old French decasyllabe. We have already identified the introduction of the Occitan lyric caesura (3F + 6M/F) as the first step towards establishing this new metrical hierarchy in which the hemistich was firmly subordinated to the line, but it was among the troubadours’ Sicilian and Tuscan heirs that the possibilities of this new order were first explored to their fullest extent. The structural demotion of the caesura in Italian meant that lines a minore came to be seen as equivalent to lines a maiore, for which reason these line-types could now be mixed freely within the context of a single poem. In the hands of the best Italian poets, this device proved to be a potent remedy against the rhythmical monotony which haunts even the greatest chansons de geste. Compare the opening three lines of La Chanson de Roland with those of Dante’s Inferno:

97

Carles li reis, ‖ nostre emper[er]e magnes

Nel mezzo del cammin ‖ di nostra vita

Set anz tuz pleins ‖ ad estet en Espaigne:

mi ritrovai ‖ per una selva oscura,

Tresqu’en la mer ‖ cunquist la tere altaigne.

ché la diritta via ‖ era smarrita.

In the very rare event that a first hemistich of the form 4F in French is followed by a truncated second hemistich, i.e. 5M or 5F, this is felt to be so alien that French metricists call it a coupe italienne (see Duffell (2000): 271).

94

By alternating the position of the caesura from line to line, Dante successfully counteracts the rhythmical see-saw effect which is immanent in the older French poem, even after three lines. To a modern sensibility accustomed to the rhythmical elasticity of free verse or of Milton’s blank verse, the regular line-fragmentation of La Chanson de Roland lends to this poem and others like it a distinctly medieval quality that is entirely absent from La Divina Commedia. But the caesural flexibility which characterizes the work of Dante and that of his two most prominent successors, Petrarch and Boccaccio, came at a price. In order to compensate for the loss of structural stability which the fixed caesura had secured, Italian poets of the fourteenth century began to regularize the stress pattern of each line in such a manner that stressed syllables became significantly less likely to occur in even syllable positions. In the Old French decasyllable no such restrictions were observed: stressed syllables were free to roam the line as long as the last metrical syllable position in each hemistich was occupied by a stressed syllable. Again, this probably reflects the comparatively marginal role played by word stress in French; in Italian, on the other hand, word stress is prominent, for which reason this feature had to be taken into account in every syllable position. At any rate, the ensuing result in Italian versification was unequivocal: over time, stressed syllables came to be associated exclusively with even syllable positions, paving the way for a consistent stress profile by which less stressed syllables alternated with more stressed syllables. It is thus in the work of Dante, Petrarch and Boccaccio that we begin to recognize clearly the sort of decasyllabic line which in English would be codified as the iambic pentameter: a five-beat structure with duple alteration and great caesural freedom. This metre, Duffell notes, is a striking innovation entirely on the part of Chaucer: ‘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’.98

98

Duffell (2000): 284.

95

Chapter 4 The Chinese Wall of Milton: Towards Blank Verse There without sign of boast, or sign of joy, Sollicitous and blank he thus began. Milton, Paradise Regain’d

‘Chaucer should be on every man’s shelf. Milton is the worst sort of poison.’1 Such is the verdict of Ezra Pound in the first of his essays on that period whose English branch one leaps over in moving from Chaucer to Milton: the Renaissance. It was then that the iambic pentameter had to be resuscitated – or, more accurately, reinvented – by poets such as Thomas Wyatt and Henry Howard, the Earl of Surrey, for it had fallen into a peculiar state. The Renaissance was also the period during which blank verse was introduced into English poetry by Surrey, for his translation of Books 2 and 4 of Virgil’s Aeneid, published in the 1550s. This chapter traces the historical process by which Chaucer’s verse design was lost and then found again, only to have it stripped of one of its most emblematic features, namely end-rhyme. Recognition of the formative importance of Wyatt’s and Surrey’s contribution to the history of English versification came very swiftly. As early as 1589, George Puttenham takes the first step towards their joint canonization: In the latter end of the same kings raigne [Henry VIII] sprõg vp a new company of courtly makers, of whom Sir Thomas Wyat th’elder & Henry Earle of Surrey were the two chieftaines, who hauing trauailed into Italie, and there tasted the sweete and stately measures and stile of the Italiã Poesie as nouices newly crept out of the schooles of Dante Arioste and Petrarch, they greatly pollished our rude & homely maner of vulgar Poesie, from that it had bene before, and for that cause may iustly be sayd the first reformers of our English meetre and stile.2

1 2

Ezra Pound, ‘The Renaissance: I – The Palette’, in Poetry, Vol. 5, No. 5 (Feb., 1915), 227-234: 231. Puttenham [1589]: 48.

96

Puttenham’s characterization of earlier English verse as ‘rude’ and in need of polishing is revealing of the degree to which Chaucer’s metrical accomplishments, as detailed in the previous chapter, soon fell into disrepair among poets and critics alike; it also foreshadows the persistence of this confusion. Even John Dryden, writing almost exactly a century after Puttenham, still fails to detect the basic principles of Chaucer’s metrical practice: The Verse of Chaucer, I confess, is not Harmonious to us; ... There is the rude Sweetness of a Scotch Tune in it, which is natural and pleasing, though not perfect. ... [C]ommon Sense (which is a Rule in every thing but Matters of Faith and Revelation) must convince the Reader, that Equality of Numbers in every Verse which we call Heroick, was either not known, or not always practised in Chaucer’s Age. It were an easy Matter to produce some thousands of his Verses, which are lame for want of half a Foot, and sometimes a whole one, and which no Pronunciation can make otherwise. We can only say, that he lived in the Infancy of our Poetry, and that nothing is brought to Perfection at the first.3 The explanations for Dryden’s failure to appreciate the metrical structure of Chaucer’s work are well-known. The one most commonly cited is that the gradual loss of word-final -e during the early fifteenth century4 appears to have deafened the ears of Chaucer’s immediate successors to the syllabic regularity of his metrics. Without knowledge of when to pronounce word-final -e, the syllable count of a line such as this becomes a murky affair: The tendre croppes, and the yonge sonne (CT, GP, l. 7) At the same time the French-inspired oxytonic stressing of a considerable number of loanwords was on its way out of the English language, to the effect that the regularity of Chaucer’s stress-patterns was, if not lost, then at least obscured in lines such as this: And bathed every veyne in swich licour, (CT, GP, l. 3) 3

Dryden (1961): 528-529. On which see Donka Minkova, The History of Final Vowels in English: The Sound of Muting (Berlin & New York: Mouton de Gruyter, 1991). 4

97

Even at the time of Dryden’s comments, philology as a scholarly discipline with a sound critical methodology of its own was still decades away, and in the absence of such methods it is hardly surprising that Dryden’s reliance on ‘common Sense’ proved insufficient. Dryden’s predicament becomes even more understandable when one takes into account the text on which his comments about Chaucer’s metre are based. As pointed out by Hoyt N. Duggan, the received text of Chaucer in Dryden’s day was Speght’s second edition of 1602, and even a cursory look at the opening of ‘The Prologue’ to the Canterbury Tales in that edition reveals the kind of textual obstacles that a seventeenth-century reader of Chaucer would have to overcome: ‘shoures’ is shortened to ‘shours’, ‘smale’ becomes ‘small’ and so forth.5 John Urry’s 1721 edition of Chaucer has become infamous for its Procrustean strategy of randomly lengthening and shortening words to make the text fit the metre, so it was not until Thomas Tyrwhitt’s 1775 edition of Canterbury Tales that philological acuity and metrical concerns were both given due consideration: Dryden’s failure to understand Chaucer’s principles of versification is, in other words, easily excused. The same courtesy should, of course, be extended to those poets who composed their verse in the decades following Chaucer’s death and whose metrical competence has been much questioned on account on their perceived ineptitude at reproducing consistently the smooth verselines of Chaucer and Gower. Thomas Hoccleve and John Lydgate are only the two most widely known representatives of this generation of English poets, a generation whose metrical accomplishments were summed up by Saintsbury with condescension: while the line [of Hoccleve, Lydgate and their English contemporaries] sometimes loses all rhythmical sufficiency, though it does yield ten syllables to the finger, it at any other time fails to respond even to this mechanical test, and simply sprawls—a frank and confessed nondescript or failure.6 The dismissive and cynical tone of Saintsbury’s comments is reflective of the fact that they were made at a time when the rediscovery of the regularity of Chaucer’s and Gower’s versification was still relatively novel: the decisive breakthrough in pinning down Chaucer’s metrical practice did not 5

For the more detailed discussion, on which my own treatment of this matter rests, see Hoyt N. Duggan, ‘Libertine Scribes and Maidenly Editors: Meditations on Textual Criticism and Metrics’, in C.B. McCully & J.J. Anderson (eds), English Historical Metrics (Cambridge: Cambridge University Press, 1996), 219-237. 6 George Saintsbury, A Short History of English Literature (London: Macmillan, 1898): 160.

98

occur until 1863 when Francis James Child published his short treatise ‘Observations on the Language of Chaucer’ in Memoirs of the American Academy of Arts and Sciences. This treatise was the first to put forward with methodological rigour the hypothesis that word-final -e in Chaucer’s verse regularly counts as a syllable when not occurring before a vowel or an , and it sparked a vitriolic debate among Chaucer scholars that would not be settled until more than a century later when Nicholas Barber & Charles Barber published their watershed computer-aided study of the metre of Canterbury Tales in 1990-91.7 However, scholars outside of the Anglophone world showed themselves more receptive to the metrical endeavours of Chaucer’s English heirs than did Saintsbury. Josef Schick, for instance, prefaced his edition of Lydgate’s Temple of Glas (1891) with a five-fold classification system of Lydgate’s metrics, according to which the regular template of the iambic pentameter that we know from Chaucer was reinterpreted by Lydgate in such a way that four alternative templates of this basic design were allowed to occur with some frequency. Two of these alternative templates have to do with what happens around the caesura (which in Lydgate’s verse is fixed almost exclusively after the fourth syllable position8): in the first alternative template, an extra-metrical unstressed syllable is admitted before the caesura; in the second, an unstressed syllable is omitted immediately after the caesura. The other two alternative versions have to do with what happens at the beginning of the line: one version is the acephalous line-type, in which an unstressed syllable is omitted at the beginning of a line; the other version is one in which an additional unstressed syllable is permitted to occur at the very beginning of an otherwise regular iambic pentameter.9 In formal notation, Schick’s classification of Lydgate’s ‘pentameter’ templates looks like this:

7

A. x / x / x / x / x /

(regular, 10 syllables),

B. x / x / x || x / x / x /

(epic caesura, 11 syllables)

C. x / x / || / x / x /

(broken-backed, 9 syllables)

D. / x / x / x / x /

(acephalous, 9 syllables)

E. x x / x / x / x / x /

(double anacrusis, 11 syllables)

See Martin J. Duffell & Dominique Billy, ‘From Decasyllable to Pentameter: Gower’s Contribution to English Metrics’, in The Chaucer Review, Vol. 38, No. 4 (2004), 383-400, and Nicholas Barber & Charles Barber, ‘The Versification of the Canterbury Tales: A Computer-Based Statistical Study’, in Leeds Studies in English, 21 (1990), 81103, & 22 (1991), 57-84. 8 Based on an extensive analysis of Lydgate’s verse, Duffel (2008) finds that this is true of 97.5 per cent of the lines analyzed (103). 9 See J. Schick (ed.), Lydgate’s Temple of Glas (London, 1891): lvi-lx.

99

While hardly exhaustive as a descriptive framework for all fifteenth-century verse in pentameter, Schick’s five templates based on Lydgate’s practice nonetheless give the reader a pretty good notion of the sort of metrical experiments that were in play more generally during that century. Special attention ought to be paid to templates B and C inasmuch as these variants on the basic trellis sound particularly uncouth to ears accustomed to more recent English verse. John Skelton’s line – Full subtyll persones || in nombre foure and thre (The Bowge of Courte, l. 7)10 – exemplifies the epic caesura of template B with, whereas Lydgate’s That his entent || can no man bewreye; (Troy Book, 1, l. 224)11 may be called on to exemplify the broken-backed template C. The effect of both line-types is to overemphasize the two hemistichs at the expense of the integrity of the line in its entirety, and in this sense – since these variants are so much more frequent in fifteenth-century English verse than they are in Chaucer – the metrical practice of the major fifteenth-century English poets may be seen as a retrogression from Chaucer’s accomplishment.12 It will be recalled that Dryden in his denunciation of Chaucer’s metrical aptitude quoted above claims to hear in Chaucer’s verse ‘the rude Sweetness of a Scotch Tune’. Perhaps Dryden ought to have trusted his ears over his common sense, for it is a curious fact of the history of English versification that Chaucer found his most faithful and accomplished imitators during the fifteenth century north of the Border. First among these, in rank as well as chronology, is James I of Scotland, whose eighteen years of imprisonment in London if nothing else provided him with an excellent education that allowed him to compose The Kingis Quair. The very last stanza of that 10

Cited from John Skelton, The Complete English Poems, edited by John Scattergood (Harmondsworth: Penguin, 1983). 11 Cited from John Lydgate, Troy Book: Selections, edited by Robert R. Edwards (Kalamazoo, MI: Western Michigan University Press, 1998). 12 One must, of course, exercise extreme caution when judging historical developments within metrics or any other area in terms of retrogression and progression. Both terms imply a somewhat naive ideology that obfuscates the messy way in which historical shifts actually occur: however, since this chapter is mainly concerned with demonstrating how the iambic pentameter became established as a normative and well-defined form, it does make sense to talk about the shifts that occurred during the fifteenth century as retrogressions from Chaucer’s use of that metre.

100

work makes it perfectly clear which poets the king took as his models: ‘Unto the impnis [hymns] of my maisteris dere, / Gowere and Chaucere, ... / I recommend my buk in lynis sevin’.13 The phrase ‘lynis seven’ refers, of course, to the heptastichic stanza-form rhyming ababbcc which James had borrowed from Chaucer’s Troilus and Criseyde: a fact which is sometimes, erroneously, assumed to be responsible for the existence of the term rhyme royal to designate that particular stanza.14 Appropriately, the same stanza was also called upon by Robert Henryson to provide the formal framework for his Testament of Cresseid, which picks up the thread from Troilus and Criseyde and provides an alternative ending to Chaucer’s tale. In Henryson’s handling of the rhyme royal stanza, it becomes an apt medium for all levels of linguistic register, most charmingly perhaps for the colloquial everyday idiom that Henryson draws on to frame his narrative in a stanza such as this: I mend the fyre and beikit me about, Than tuik ane drink, my spreitis to comfort, And armit me weill fra tha cauld thairout. To cut the winter nicht and mak it schort I tuik ane quair, and left all uther sport, Writtin be worthie Chaucer glorious Of fair Creisseid and worthie Troylus. (Testament of Cresseid, ll. 36-42)15 The penultimate line of this stanza stands out from the majority of lines in Henryson’s poem in that it does not have a caesura after the fourth or the sixth syllable, for which reason it might indeed have been ‘Writtin be worthie Chaucer glorious’ himself. James and Henryson are the earliest representatives of an important group of Scottish poets who in the course of the fifteenth and early sixteenth centuries used Chaucer as a model with such diligence and skill that until recently they were most commonly referred to as the Scottish Chaucerians. Today, the term Makars seems to be the preferred designation for this band of poets, which in addition to James and Henryson also includes William Dunbar, David Lindsay and Gavin 13

The Kingis Quair, ll. 1373-1378; cited from James I of Scotland, The Kingis Quair, with an Introduction, Notes and Glossary by John Norton-Smith (Oxford: Oxford University Press, 1971). 14 One John Quixley thus, as early as 1400, used the term ‘royal’ about his chosen stanza-form for his translation of Gower’s French ballades, entitled Traitie pour essampler les amantz marietz; see Martin Stevens, ‘The Royal Stanza in Early English Literature’, in PMLA, Vol. 94, No. 1 (Jan., 1979), 62-76. 15 Cited from Derek Pearsall (ed.), Chaucer to Spenser: An Anthology (Oxford: Blackwell, 1999).

101

Douglas, alongside several other lesser poets. For our purposes, Douglas is of particular interest on account of his translation of Virgil’s Aeneid, a translation whose metrical qualities Pound held in the highest regard: ‘Gavin Douglas [in his Eneados] ... attains a robuster versification than you are likely to find in Chaucer ... the texture of Gavin’s verse is stronger, the resilience greater than Chaucer’s’.16 As is not uncommon with Pound’s scattered comments on matters relating to metre, it is difficult to know whether or not one agrees with him because it is difficult to know exactly what he means. It is, however, generally accepted that in terms of metrical regularity, Douglas and his fellow Scottish Makars came closer to the example set by Chaucer than did their English contemporaries. That is to say that the aberrant line-types that were discussed earlier as characteristic of the decasyllabic verse of Lydgate, Hoccleve and Skelton tend to occur with a lower frequency in the Makars’ use of that measure, though still more frequently than in Chaucer.17 The purpose of this short overview of the metrical practices of Chaucer’s English and Scottish successors is not to pass judgment on the literary merits of each branch, nor is it to claim that a rigid distinction even needs to be upheld between the two branches. Rather, the overview is included here partly because it provides the necessary backdrop against which the contribution of Wyatt and Surrey to the history of English versification must be viewed, and partly to make the point that vernacular poetry continued to be produced in Britain throughout the century that began with Chaucer’s death. This latter point is not quite as trivial as it may seem. For while it is true that Chaucer’s immediate legacy was slight from a purely metrical point of view, the high regard in which his work was nonetheless held secured a copious production of British verse in the vernacular throughout the fifteenth century. In this respect the situation in Britain after Chaucer’s death was almost the opposite of the situation in that region which had most influenced his versification, the Tuscany of Dante, Petrarch and Boccaccio. Rather than heralding in a prolonged period of shaky metrical experiments in the vernacular, as was the case in Britain after Chaucer’s death, the death of Petrarch in 1374 instead heralded in what Benedetto Croce has referred to as ‘un secolo senza poesia’,18 by which should be understood a century without poetry in the vernacular. As Carol Kidwell notes: ‘[From the death of Petrarch and Boccaccio u]ntil the last quarter of the fifteenth century ... no more poets wrote in Tuscan. The

16

Ezra Pound, ABC of Reading (London: Faber and Faber, 1951): 115. See Duffell (2008): 99-115. 18 Benedetto Croce, Poesia popolare e poesia d'arte. Studi sulla poesia italiana dal Tre al Cinquecento (Bari: Laterza, 1967): 209-238. 17

102

outstanding literary figures in Florence were all latinists [sic]’.19 Thus, throughout most of the Quattrocento, the reputation of Petrarch in Italy rested almost exclusively on his writings in Latin, whereas his vernacular verse occupied at best a marginal position. And how could it be otherwise? After all, the vast majority of Petrarch’s writings – all of his works in prose, most of his verse and all of his letters – were written in Latin; the Canzoniere and the Trionfi are the only works composed in his vernacular Tuscan. Not until the late fifteenth century did commentaries on the Canzoniere begin to emerge, and only in 1501 did Aldus Manutius print Pietro Bembo’s influential edition of Petrarch’s Tuscan poems, Le cose volgari di messer Francesco Petrarcha. As pivotal as this edition was to the promulgation of Petrarch’s vernacular work in Italy, it was, however, the publication of Bembo’s treatise Le Prose della volgar lingua in 1525, which ultimately secured for Petrarch’s vernacular verse the central role that it eventually came to play on a European scale. Given the decidedly international influence of Petrarch during the sixteenth century, it is not without irony that the work responsible for his canonization as a poet worthy of imitation was a treatise dedicated to creating a unified national language for Italy on the basis of a single local dialect. Bembo’s effort to promote the Tuscan of Boccaccio and Petrarch as a model for modern Italian met with some resistance – among others from Gian Giorno Trissino, who advocated a more geographically diverse approach to the question of how a national Italian language should be constructed – but in the end Bembo’s view prevailed and ultimately led to the foundation of the Accademia della Crusca in Florence in 1582 and the subsequent codification of ‘Florentine pronunciation’. By that time Petrarch had already become a household-name all over Europe, and the sonnet form which he had cultivated in the Canzoniere had been transplanted to shores as distant as those of Britain. The slowness with which Petrarch’s vernacular verse became canonized as a linguistic and stylistic model for Italian literature meant that by the time of Wyatt’s first visit to Italy, in the late 1520s, the English courtier did not encounter a declining century-old tradition of composing poetry in Tuscan, but rather a vigorous revivalist movement aimed at reintroducing the virtues of Petrarch to contemporary Italian verse. Among the better known Italian poets who followed Bembo’s example of imitating the style of Il Canzionere were his contemporary Serafino dell’ Aquilo and Annibale Caro, who belonged to the next generation of Italian Petrarchans. The challenge facing these poets was, in a sense, even greater than the challenge facing Wyatt in his task of transplanting the Petrarchan sonnet to English. For whereas early Italian Petrarchans such as Bembo and Serafino 19

Carol Kidwell, Pietro Bembo: Lover, Linguist, Cardinal (Montreal, QC: McGill-Queen’s University Press, 2004): 221. Appropriately the volume is typeset in 10/12 Sabon with Bembo display.

103

were imitating their master in the shadow of the Latinate secolo senza poesia and therefore had to rediscover a vernacular poetic idiom to go with the formal requirements of the Canzionere, Wyatt had the distinct advantage of having at his disposal the composite idiom of an uninterrupted tradition of vernacular poetry going all the way back to Chaucer.20 In spite of the metrical deficiencies of much fifteenth century Anglic verse, Chaucer’s poetic idiom had been kept alive and had even been further developed and refined by his Scottish and English imitators. This literary heritage was habitually recognized by many of Wyatt’s Anglic predecessors, most commonly in the form of an elaborate excuse for the poet’s own failure to meet the standard set by those who came before him. The following stanza from the Prologue to Stephen Hawes’s The Example of Vertu (1504) is representative of this particular brand of modesty formula: O prudent Gower in langage pure Without corrupcyon moost facundyous O noble Chauser euer moost sure Of frutfull sentence ryght delycyous O vertuous Lydgat moche sentencyous Vnto you all I do me excuse Though I your connynge do now vse (The Example of Vertu, ll. 22-28)21 Given the strong awareness of a vernacular literary tradition that lines such as these demonstrate, it seems unwise to follow C.S. Lewis and dismiss en bloc the immense body of work that the major fifteenth-century poets produced simply as ‘the late medieval swamp’ out of which Wyatt and Surrey had ‘to build a firm metrical highway’.22 In fact, it seems highly probable that the efforts of the post-Chaucerians to keep alive an Anglic poetic idiom throughout the fifteenth century may have been a decisive factor in securing for the sonnet the central position that it came to occupy in English poetry during the latter half of the sixteenth century.

20

On Wyatt’s indebtedness to Chaucer, see John Watkins, ‘“Wrastling for this World”: Wyatt and the Tudor Canonization of Chaucer’, in Theresa M. Krier (ed.), Refiguring Chaucer in the Renaissance (Gainesville, FL: University Press of Florida, 1998), 21-39; and Dennis Kay, ‘Wyatt and Chaucer: They Fle From Me Revisited’, in Huntington Library Quarterly, 47 (1984), 211-225. 21 Cited in A.S.G. Edwards, Stephen Hawes (Boston, MA: Twayne Publishers, 1983): 10-11. 22 C.S. Lewis, English Literature in the Sixteenth Century (Oxford: Oxford University Press, 1965): 237.

104

❦ Wyatt’s dabblings with his Petrarchan model have attracted their share of scholarly attention, not only because they represent the earliest examples of Petrarchism in English, but especially because of their metrical structure, or rather their occasional lack thereof. The palpable roughness of Wyatt’s versification in his earliest production has gained for him a secure position in the modern canon of English Renaissance poetry, which has been formed under the influence of the emergence of free verse. The relatively high regard in which Wyatt’s verse is currently held usually carries with it a correspondingly low evaluation of the verse practice of his friend and immediate heir, the Earl of Surrey. The relative critical evaluation of Wyatt and Surrey in various periods of literary scholarship constitutes, in fact, a surprisingly reliable barometer with which to assess the aesthetic preferences of a given age: ages in which regularity has been viewed as a quality in and by itself tend to favour Surrey, whereas ages in which regularity has been viewed as stifling tend to favour Wyatt. Wyatt’s mother tongue may have provided him with a workable poetic idiom, but what it did not provide him with was a stable metrical matrix of the iambic pentameter. As Peter Groves points out in his critique of early twentieth-century attempts to force Wyatt’s most aberrant lines into such a pentameter matrix, the primary fault of such endeavours is that ‘they assume that the pentameter already existed as a model for Wyatt, and that his chief problem was to mould the language of his verse to fit it’.23 However, as we have just seen, the poor textual transmission of Chaucer up until the late eighteenth century effectively concealed and obfuscated even the most basic principles of Chaucer’s metrical design, for which reason alone it seems improbable that the modern understanding of Chaucer’s verse design should underlie Wyatt’s metrical experiments in the late 1520’s and early 1530’s. The principal snag with Wyatt’s verse instances is therefore not that so many of them fail to meet the metrical requirements of the iambic pentameter, but rather that a significant portion of them do. Consider, for instance, the opening quatrain of Wyatt’s translation of Petrarch’s Rime 140: The longe love that in my thought doeth harbar And in myn hert doeth kepe his residence 23

Peter L. Groves, ‘Finding his Feet: Wyatt and the Founding of English Pentameter’, in Versification [An Electronic Journal of Literary Prosody], 4 (2005).

105

Into my face preseth with bold pretence, And therein campeth, spreding his baner.24 Line 2 scans as a perfectly regular iambic pentameter, and line 3 could well be interpreted as an iambic pentameter with an inversion of the third foot, licensed by the phrasal break which precedes it. Lines 1 and 4, on the other hand, cannot be interpreted as iambic pentameters in that the unstressed final syllables of the paroxytones harbar and baner occur in the crucial tenth syllable position.25 Attempts to rescue the metricality of line 1 by making the final -e of longe syllabic (as a result of which it would be the first syllable of harbar that would occupy position 10) are not warranted by linguistic criteria outside of narrow metrical concerns, and furthermore no similar procedure can be performed to rescue line 4. Nor can the claim that each of the four lines quoted contains exactly ten syllables be sustained and used as an argument in favour of isosyllabism as the guiding metrical principle of Wyatt’s sonneteering. In other sonnets by Wyatt we thus encounter lines with as few as eight syllables – /

.

/

.

/

.

/ .

Envy theim beyonde all mesure 26 – and with as many as twelve: .

/ . /

.

/

.

/ .

/ .

/

I fley above the wynde yet can I not arrise 27 The clear iambic profile of this latter line – and of the second line of the quatrain quoted above – indicates that Wyatt was fully capable of producing such smooth stress-profiles,28 but their relative scarcity in his sonnets also tells us that such lines represent only one possible realization of the metrical assumptions that he was working under. George T. Wright has proposed that Wyatt made 24

Cited from Collected Poems of Sir Thomas Wyatt, edited by Kenneth Muir & Patricia Thomson (Liverpool: Liverpool University Press, 1969). 25 For a more detailed account of the metrical problems inherent in this sonnet in its entirety, see Groves (2005): Figures 2 & 19. 26 ‘Love and Fortune and my mind remember’, l. 4; cited from Wyatt (1969). 27 ‘I find no peace, and all my war is done’, l. 3; cited from Wyatt (1969). 28 As can also be witnessed in Wyatt’s poems in poulter’s measure, in which the alternation between less stressed and more stressed syllables is excruciatingly regular.

106

use of a metrics rooted in the practice of Lydgate, but replaces the five templates identified by Schick with as many as eight different templates. Peter Groves, on the other hand, sees Wyatt’s versification as a misconstruction of the endecasillabo which only gradually stabilized itself as a ‘proto-pentameter’, and even then made use of a number of features not subsequently codified in English versification, especially catalexis, both line-initially and line-internally.29 Considerably less scholarly sophistication informed the endeavours of Wyatt’s most influential editor, the one who prepared his poems for inclusion in Tottel’s Songes and Sonettes (1557). The eight-syllable trochaic line, which was cited above to illustrate the syllabic sparseness of some of Wyatt’s sonnet-lines, occurs in ‘Love and Fortune and my mind, rememb’rer’, of which this is the opening quatrain as it appears in the Egerton manuscript: Love and Fortune and my mynde, remembre Of that that is nowe with that that hath ben, Do torment me so that I very often Envy theim beyonde all mesure.30 However, by the time that the Tottel-editor was through with this passage, it had come to look like this: Love, Fortune, and my minde which do remember Eke that is now, and that that once hath bene: Torment my hart so sore that very often I hate and envy them beyonde all measure.31 To a modern reader accustomed to the expressive possibilities of counterpointing metre and meaning, almost all of the Tottel-editor’s emendations will seem abject failures. In line 1, ‘Love’ has been demoted from its original and highly prominent beat-carrying position at the opening of an acephalous line to a mere anacrusis to ‘Fortune’ in Tottel’s reworked version. In the same line, the personification of the speaker’s mind – originally conveyed in the form of an apposition by the interesting epithet ‘remembre’ – is changed into a decidedly uninteresting relative clause, which is 29

Wright (1985); Groves (2005). Cited from Wyatt (1969). 31 Cited from Edward Arber, Tottel’s Miscellany, English Reprints (Westminter: A. Constable, 1897). 30

107

made all the more uninteresting by the insertion of the periphrastic and entirely redundant particle ‘do’. These editorial choices in turn affect the structure of the quatrain’s second line, which in Tottel’s version becomes an inelegant pentameter with an unwarranted inversion of the first foot: /

.

.

/

.

/

.

/

.

/

Eke that is now, and that that once hath bene: This editorial decision spoils not only the spectacular double occurrence of the homographic echo ‘that that’ in Wyatt’s original line, but also the sophisticated manner in which the stress-profile of the original line subtly underpins its own referential content: .

/

.

.

/ ˰ .

/

.

.

/

Of that that is nowe with that that hath ben, In Wyatt’s original wording, this decasyllabic line splits up quite effortlessly into two completely equipollent pentasyllabic half-lines – each with the accentual profile . / . . / – and thus seems to force upon the reader rhythmically the unavoidable commensurability of that which is in the present and that which is in the past. The Tottel-editor’s treatment of the quatrain’s two final lines is equally disastrous: in order to furbish Wyatt’s original lines in such a manner that they meet the requirements of the iambic pentameter, the editor feels compelled to insert into them, not only the speaker’s heart, but also a redundant adverbial to go with it, as well as a new verb phrase. These editorial insertions leave us with two impeccably toothless iambic pentameters with feminine endings: .

/

.

/

.

/

.

/ . / .

Torment my hart so sore that very often . /

.

/ .

/

. /

.

/

.

I hate and envy them beyonde all measure. This emendation, admittedly, takes care of the metrical embarrassment caused by Wyatt’s original concluding trochaic tetrameter, but it also precludes the reader from pondering whether the notion of envy beyond all measure might, in fact, inspire a poet to venture beyond his poetic measure, too.

108

What is furthermore lost in this editorial process is the way in which the metrical structure of the original last line works in tandem with that of the one immediately preceding it: / .

/

.

/˰ / . / .

/ .

Do torment me so that I very often /

.

/

.

/

.

/

.

Envy them beyonde all mesure. In this scansion of the first line, six metrical beats are allowed to occur in order to make up for the missing beat of the concluding trochaic tetrameter; in this way the two lines – when conceived of as a single unit – do meet the requirement of having ten metrically accented syllables between them. And yet for all of his mind-numbing, automatized emendations of Wyatt’s original texts, the anonymous editor of Tottel’s Songes and Sonnettes is not the villain of the story that this chapter aims to relate. In regularizing Wyatt’s experimental verse lines to meet the expectations of the emerging book-buying market, the Tottel editor not only made Wyatt more accessible to the reading public but also made his poems more instantly commensurable with those of his younger friend and colleague, the Earl of Surrey, who was the real star of the Miscellany.32 Furthermore, it might very well be argued that without the metrical meddling of the Tottel editor, Wyatt would have been remembered today, not as the instigator of the Renaissance in English verse, but rather as the last of those obscure and rarely studied medieval post-Chaucerians, to whom a section of this chapter has been devoted. To put it more bluntly: by the mid-sixteenth century, English poetry was more in want of viable metrical models than of yet another experimental paragon of Lydgate, no matter how brilliant. Tottel’s editor made Wyatt’s verse into such a model, and in doing so he also laid the foundation for the current practice – already sanctioned by the time of Puttenham – of referring to Wyatt in tandem with Surrey as the great reformer of English versification. Yet it is quite obviously to Surrey that the honour of having reformed most influentially the very core of English versification ought to fall. Surrey’s contributions to Tottel’s Songes and Sonettes come down to us unblemished (so far) by the emergence of any Egerton manuscript to undermine this feat, for which reason we must assume that the poet who gave credit to Wyatt for having ‘taught what might be said in rhyme’33 was, in fact, himself the true reformer of English 32

As is apparent from the full title of the first edition of Tottel’s anthology: Songes and Sonettes Written By the Ryght Honorable Lord Henry Howard, late Earle of Surrey, and other. 33 Henry Howard, ‘Of the Death of Sir T. W. The Elder’, l. 13.

109

Renaissance versification. It was, we must assume, to a large extent Surrey’s poems which served as the model on which Wyatt’s poems were emended for the Tottel Miscellany, and it was that anthology’s presentation of our two poets which was to become formative for the important poetic developments which took place in the decades immediately following its publication. It is for that reason that the following considerations of the sonnets of Wyatt and Surrey are concerned exclusively with the texts as they appeared in Tottel’s anthology. If we adopt a somewhat purist definition of the sonnet as a poem consisting of 14 predominantly decasyllabic lines, Tottel’s Miscellany includes 27 sonnets ascribed to Wyatt and 14 ascribed to Surrey. The most conspicuous structural difference between the sonnets ascribed to Wyatt and those ascribed to Surrey is, of course, to be found in the rhyme schemes. With the exception of ‘Such is the course’ (which rhymes ababababababcc), all of Wyatt’s sonnets in Tottel open with a Petrarchan octave rhyming abbaabba followed by a sestet that usually rhymes cddcee.34 This rhyme scheme for the sestet differs from those preferred by Petrarch, in which the concluding couplet is avoided (typically by means of the rhyme schemes cdecde or cdcdcd), but Wyatt’s rhyme scheme has the interesting feature of balancing the sestet half-way between a quatrain followed by a couplet (cddc + ee) and two tercets (cdd + cee). Most commonly, though, Wyatt opts for letting his syntax emphasize the two tercets, but even then the couplet is always there, threatening to overturn the neat symmetry. Surrey, on the other hand, champions the closing couplet in his sonnets, and the effect of the couplet becomes all the more pronounced in that Surrey consistently avoids two consecutive rhymes in all other line positions. Only in ‘I never saw my Ladye’ (which rhymes abbacddceffegg) does Surrey make use of the enveloped quatrain rhymes that dominate Wyatt’s sonnets; in all other sonnets by Surrey the opening twelve lines are structured by means of alternating rhymes with no possibility of having two rhyming lines in immediate succession outside of the concluding couplet. The most common of Surrey’s rhyme schemes is the one for which he is also most famous, probably because it was picked up by Shakespeare for his sonnets: ababcdcdefefgg. This is a stanza design that lends structural stability to the quatrains and the couplet at the expense of the octave and the tercets, and we find this pattern in as many as nine out of the 14 sonnets attributed to Surrey in Tottel’s Miscellany. Two of Surrey’s sonnets (‘Brittle beautie’ and ‘Alas so all things nowe’) follow the ababababababcc pattern of Wyatt’s ‘Such is the course’; ‘The fansy, which that I’ has a pattern of ababababacaccc; and ‘The soote season’ has a pattern of ababababababaa. None of these rhyme 34

Two alternative rhyme schemes for the sestet occur in Wyatt’s sonnets: cdccdd (‘The longe love’) and cdcdee (‘If waker care’ and ‘The piller perisht’).

110

schemes became influential in the subsequent codification of the English sonnet tradition, and the reason for this is fairly obvious: the requirement of finding rhymes that can be extended for six or eight lines would impose too severe restrictions on the sonneteer’s lexical freedom in a comparatively rhyme-poor language such as English. The scarcity of English rhymes is also frequently called upon to explain why Surrey abandoned Wyatt’s preferred pattern of abbaabbacddcee in favour of the ababcdcdefefgg pattern. After all, the former makes use of only five rhymes, two of which must occur four times, whereas the latter makes use of as many as seven rhymes, none of which is under obligation to occur more than twice. While this explanation is not without merit, it fails to account for the fact that the Petrarchan octave – which is the section that requires a profusion of rhymes – remained by far the most popular way to open a sonnet among English poets in the centuries after Surrey’s invention. Based on a study of more than six thousand sonnets in English (a fairly comprehensive selection by any standard) L.T. Weeks as early as 1910 found that an astounding 58.5 per cent of these sonnets made use of the Petrarchan abbaabba octave, whereas only 22 per cent made use of Wyatt’s more flexible ababcdcd octave.35 English, in other words, is not incapable of yielding sufficiently varied rhymes to fill out the Petrarchan octave, and even if the scarcity of rhymes in English did play a part in Surrey’s decision to abandon Wyatt’s preferred rhyme scheme, it still does not explain why he opted for alternating rhymes to the exclusion of enveloped ones. ❦ In addition to the different types of rhyme schemes employed by Wyatt and Surrey in their sonnets, there is another distinguishing feature – albeit less conspicuous – that sets our two sonneteers apart from one another. This feature has to do with what qualifies as a rhyme, or, to put it in slightly more technical terms, what kinds of syllables are allowed to occur in the tenth syllable position. This position is, of course, the metrical crux of rhyming and from a structural point of view it is also the most severely restricted of any syllable position of the iambic pentameter. The most commonly mentioned structural distinction between line-endings in English is, of course, that between masculine and feminine endings, but there is another even more basic distinction to be drawn, and that is the distinction between lines in which the tenth syllable position is occupied by a word which 35

See L.T. Weeks, ‘The Order of Rimes of the English Sonnet’, Modern Language Notes, Vol. 25, No. 6 (Jun., 1910), 176-180). Weeks’s corpus contains 6,283 sonnets; of these 5,940 make use of an identifiable octave. The abbaabba pattern accounts for 3,477 of these; the ababcdcd pattern accounts for 1,306.

111

receives lexical stress and lines in which this is not the case. To illustrate this distinction – and to survey the various forms it may take in the English iambic pentameter – I have used lines from Paradise Lost to exemplify the different possibilities, and in each example the syllable which occupies the tenth position is underlined, whereas the lexical stress of each line-final word is indicated by putting the relevant syllable in bold. Thus, whenever the syllable occupying the tenth syllable position is both underlined and in bold, lexical stress and metrical prominence coincide, but whenever a syllable is underlined without being in bold it means that the tenth syllable position is occupied by a syllable that does not receive lexical stress. By far the most common way of securing the structural stability of an iambic pentameter in English is to fill the tenth syllable position with a lexical monosyllable that carries the fifth beat: Of man’s first disobedience, and the fruit (PL 1, 1)36 It can, of course, be argued that since stress is always a relational feature in English, it makes little sense to say that a monosyllable receives lexical stress in and by itself, but this is a discussion which is outside of our immediate concerns. What is important here is that – for quite obvious reasons – there is no syllable in a monosyllabic word which receives more stress than the one which occupies the tenth metrical position. Equally forceful, however, as a means to secure the structural stability of the pentameter is the choice of an oxytonic disyllable or polysyllable, positioned in such a manner that the syllable which carries the lexical stress occupies the tenth position: Of Oreb, or of Sinai, didst inspire (PL 1, 7) Had need from head to foot well understand; (PL 6, 625) Feminine endings typically occur when a line of iambic verse terminates in a paroxytonic disyllable or polysyllable in such a manner that the post-tonic syllable comes immediately after the tenth position: 36

All references to Paradise Lost are to John Milton, Paradise Lost, edited with an Introduction and Notes by John Leonard (London: Penguin, 2000).

112

And high disdain, from sense of injured merit, (PL 1, 98) Of rebel Angels, by whose aid aspiring (PL 1, 38) What all of the above options have in common is that the lexical stress of the word that terminates the line coincides with the tenth syllable position, thereby adding closure and structural stability to the line. But there are other, more problematic, ways of terminating a line, i.e. in which the tenth syllable position is occupied by a syllable that does not receive word stress. A paroxytonic disyllable is thus sometimes allowed to occur at line-ends, provided that the disyllable in question is a compound word: By ancient Tarsus held, or that sea-beast (PL 1, 200) Another version of the same phenomenon involves the line-final occurrence of a polysyllable with stress on the fourth to last syllable; this usually results in a feminine ending, e.g.: Well if thrown out, as supernumerary (PL 10, 887) The mere scarcity of polysyllables with this stress profile in English pretty much guarantees their infrequent presence at line-endings, and the option of concluding a line with a compound word such as sea-beast is equally rare. There is, however, one fairly common way by which the tenth syllable position may be occupied by a syllable that does not receive word stress, and that is when the linefinal word is a proparoxytonic polysyllable: No light, but rather darkness visible (PL 1, 63)

113

The familiarity of this line – and the relatively high frequency of lines that terminate in proparoxytonic words throughout large stretches of the history of English verse – is likely to make us insensitive to the fact that the crucial tenth syllable position in lines such as the one above is, in fact, occupied by a syllable that does not receive word stress. This issue is one which will be taken up in more detail in Chapters 5 and 7, but it is introduced here because it is also relevant to the discussion of why blank verse initially made its entry into English poetry, and why it was Surrey who cleared the channel. Rhymeless decasyllabic verse in English has been known as blank verse at least since the late sixteenth century, and the form made its first appearance in print during the 1550s, most prominently, perhaps, when Richard Tottel published Surrey’s English translation of Books 2 and 4 of the Aeneid under the title of Certain Bokes of Virgiles Aenæis turned into English Meter (1557). The unpretentious characterization of Surrey’s chosen verse form as simply ‘English meter’ hardly does justice to the striking novelty of Surrey’s decision to dispense with end-rhyme as early as 1540 when the translations were completed. In fact, the novelty of Surrey’s chosen form had been recognized much more adequately three years earlier on the title-page of John Day’s 1554 edition of Surrey’s translation of Book 4 where it says that Virgil’s poem has been ‘drawne into a straunge meter by Henry Earle of Surrey’. It is, of course, impossible to determine decisively what connotations we are to read into the word straunge, as several meanings of that word were in use during the sixteenth century. The majority of these, however, incorporate the quality of being foreign, alien and unfamiliar rather than merely peculiar or plain odd. At any rate there is an interesting semantic tension between the two adjectives ‘straunge’ and ‘English’ as modifiers of one and the same metre, a tension which is also present in blank verse itself: its lack of rhyme invokes the metres of the classical languages, while at the same time its asymmetrical five-beat structure, especially when not supported by rhyme, is so unobtrusive that it lends itself easily to the idiom of English speech. This tension between the familiar and the exotic that permeates blank verse as a poetic form is also mirrored in the peculiar fact that blank verse in English literature came to serve two very different functions. One function of blank verse – the one that Surrey, and later Milton, would use it for – was to provide the staple metre of epic, or heroic, verse for poets such as James Thomson, William Cowper and William Wordsworth; its other function was to provide the standard medium for dramatic verse from the late sixteenth century up until the closing of the London theatres in 1642. However, the history of blank verse cannot, and should not, be reduced to a tale of a single

114

metrical form that was originally developed for dramatic purposes and then taken over by poets with epic ambitions. Such an explanation would not only conceal crucial differences between dramatic and non-dramatic blank verse, but would also be at odds with the fact that the two uses of blank verse appear to have developed quite independently of one another, and that the non-dramatic use is, at least in English, the earlier. The problem is succinctly summarized in a 1937 essay by George K. Smart: Histories of literature and of prosody universally trace these developments [of dramatic and non-dramatic blank verse] very sketchily, jumping dextrously from Surrey’s non-dramatic blank verse to that in the drama Gorboduc, and passing on to its use in dramas by Marlowe, Shakespeare, and other dramatists, to return to the non-dramatic blank verse with Milton. ... In addition to being incomplete, this method is misleading, for dramatic blank verse is quite different from non-dramatic, and the two ought not to be considered one form.37 The quotation has been drawn from O.B. Hardison’s brilliant essay ‘Blank Verse before Milton’ (1984), which seeks to redress from a historical point of view the shortcomings in earlier metrical scholarship that Smart points out.38 We shall return to some of Hardison’s historicizing arguments later in this chapter, but before we do that we must first consider some structural and phonological arguments that may help to explain the unexpected emergence of blank verse in English. At the heart of these arguments lies the distinction drawn above between lines whose tenth syllable position is occupied by a syllable that receives lexical stress and lines in which this is not the case. Our main focus will be on lines of the latter type, and it will be recalled that such lines are most commonly the result of a proparoxytonic polysyllable in line-terminal position. Proparoxytones in English may conveniently be split into two different categories based on the relative prominence of the two syllables that come after the stressed syllable. By far the more common type is the sort of word in which the last syllable carries at least as much stress as the one preceding it; the second type is the sort of word in which the last syllable carries decidedly less stress than the one preceding it. This latter type of stress profile is quite rare in English and is found predominantly in compound words such as freshwater, handmaiden or warmonger: words like these 37

George K. Smart, ‘English Non-dramatic Blank Verse in the Sixteenth Century’, in Anglia, LXI (1937), 370-397. The enduring relevance of Smart’s challenge to contemporary scholarship is indicated by the fact that Robert B. Shaw makes no attempt to rise to it in his recent monograph Blank Verse: A Guide to Its History and Use (Athens, OH: Ohio University Press, 2007). 38

115

are generally barred from occurring at a position where their final syllable occupies the tenth syllable position of an iambic pentameter, for which reason they are of limited concern to the present discussion.39 The former type of stress profile is found in words such as wanderlust, nightingale or, indeed, visible. While line-final proparoxytones with this stress profile do not generally present a metrical problem in English, the practice of Surrey in his sonnets reminds us that they do present a potential problem pertaining to rhyme. This problem is especially salient with regard to proparoxytones in which the relative difference in stress between the two last syllables is minimal, the word visible being a case in point. A preliminary formalized classification of the different types of proparoxytones in English may thus take the following shape, in which 3 designates a syllable that receives primary lexical stress, 2 designates a syllable that receives secondary stress, and 1 designates a syllable that receives minimal stress:

Table 4.1 Structure Example

Transcription

Metrical status of final syllable Allowed at position 10

3.1.1

visible

/ˈvɪz.ᵻ.bl/

3.1.2

wanderlust

/ˈwɒn.də.lʌst/ Allowed at position 10

3.2.1

warmonger /ˈwɔː.mʌŋ.ɡə/ Barred from position 10

The principal problem with a 3.1.1-proparoxytone such as visible at the end of a line of rhyming iambic pentameter may be rephrased as a very simple question: what rhymes with visible? Milton, of course, did not have to provide an answer to this in Paradise Lost, but other poets have ventured suggestions. Here, for instance, are two suggestions by Shelley: Yielding not, wounded the invisible Palms of her tender feet where’er they fell: (‘Adonaïs’, ll. 211-212)

39

This is not to say that such compound words never occur in line-final position, only that they do so very rarely. When they do occur, it is usually in dramatic blank verse, particularly that of Shakespeare’s early plays, e.g. ‘A knot you are of damnéd blood-suckers’ (Richard III, 3, 3, 6). For more examples and further discussion, see Paul Kiparsky, ‘Stress, Syntax, and Meter’, in Language, Vol. 51, No. 3 (Sep., 1975), 576-616: 589-591.

116

Of day meanwhile had sunk behind the hill, And the black bell became invisible, (‘Julian and Maddalo’, ll. 133-134)40 Whether fell or hill constitutes the better rhyme with visible – or as in these examples invisible – is probably a matter of taste and is at any rate a question that pertains, not to the level of verse design, but rather to the level of verse delivery. However, the disconcerting fact that fell and hill do not rhyme with one another ought to make us aware that there is at least a potential problem with 3.1.1proparoxytones such as visible/invisible in rhyming positions. This problem has to do with the quality of the vowel which functions as the nucleus of the last syllable of visible. Unlike the vowels in fell or in hill, both of which are fully articulated in standard pronunciation, the vowel in -ble is reduced, for which reason the syllable of which it forms part is incapable of receiving stress in normal speech and therefore unable to support a rhyme on its own.41 Employing the final syllable of a 3.1.1-proparoxytone at the rhyming tenth position of an iambic pentameter therefore inevitably leads to one of two options, neither of which is entirely satisfactory to the exacting ear. Either the reduced vowel of the proparoxytone is promoted to rhyme with a full vowel – as in the two examples by Shelley quoted above – or it is rhymed with another reduced vowel such as in this example by Wyatt: Sins that disceit is ay returnable, Of verye force it is agreable, (‘Was never file yet half’, ll. 10-11) This particular rhyme-pair is an example of homoeoteleuton – i.e. the rhyming of similar or identical suffixes rather than word stems – and it is a device of which Wyatt was rather fond, as can be seen from the concluding sestet of ‘Ye that in love finde luck’: 40

Cited from P.B. Shelley, The Complete Poetical Works, edited by Thomas Hutchinson (Oxford: At the Clarendon Press, 1904). 41 In discussions such as these, the possibility of diachronic changes in pronunciation must, of course, be taken into account. While it is certain that the vast majority of 3.1.1.-proparoxytones would also have been pronounced with the lexical stress on their first syllable in the sixteenth and seventeenth centuries , it is decidedly more difficult to determine the exact vowel quality with which the unstressed syllables of such words would have been pronounced. On this, see Charles Barber, Early Modern English (Edinburgh: Edinburgh University Press, 1996): 132-134. However, the fact that one can easily find examples of inconsistent rhymes involving proparoxytones within the same poem from the period with which we are concerned implies that the pronunciation of the final syllable of proparoxytones, even at this early stage, must have been a murky affair: Christopher Marlowe, in Hero and Leander, thus rhymes she/chastitie (411-412), Virginitie/eie (269-270), harmony/by (105-106) and destroy/harmony (251-252).

117

Sephame saide true, that my natiuitie Mischaunced was with the ruler of the May. He gest I proue of that the veritie. In May my welth and eke my liff I say Haue stoude so oft in such perplexitie: Reioyse! let me dreme of your felicitie.42 Here, as many as four lines terminate in the suffixed syllable -tie, creating a monotonous sense of repetition without adding either closure or suspense. This solution to the problem of syllables with reduced vowels occurring in rhyming positions is unlikely to be perceived as successful by a modern reader, but at the same time it must be admitted that the alternative solution of rhyming reduced vowels with full vowels, such as in invisible/fell or invisible/hill, is not without problems of its own either. Both solutions will tend to create rhymes that challenge the natural exactitude of a full-throated rhyme between two lexically stressed syllables with identical unreduced vowels at their core, such as, say, May/say. ❦ For a chapter whose title asserts that it is about blank verse, we have so far spent an alarming amount of energy on discussing rhyme. But just as any informed discussion of free verse unavoidably entails a good deal of metrics, so the discussion of blank verse necessitates some consideration of the property by whose absence it is defined. The digression from which we have just returned about the problems created by proparoxytones at line-endings in rhyming pentameter has been necessary in order to attempt the following explanation as to why Surrey for his translation of the Aeneid chose to dispense with rhyme. The problem is summed up pithily by Hardison: When he decided to translate the Aeneid, Surrey was fully aware of the quality of his original. He would have learned about dactylic hexameter in grammar school and would have been exposed to more sophisticated comments on the music of Vergil’s line in

42

Cited from Wyatt (1969).

118

Renaissance editions of Vergil. Since his decision to use blank verse had no precedent in English, it must have been deliberate.43 The deliberateness with which Surrey discarded rhyme for his Aeneid may in part – as Hardison points out – be due to the lack of rhyme in the original, but perhaps there is also a clue to be found in Surrey’s own concept of rhyme, to the extent that it can be reconstructed from his practice in the sonnets. Apparently Surrey took a very different view from that of Wyatt on the appropriateness of proparoxytones at line-final positions: whereas Wyatt revels in them and the awkward rhymes they produce,44 Surrey seems to have taken pains to avoid them altogether. Thus, only three out of the 196 lines which make up Surrey’s fourteen sonnets in Tottel terminate in the highly problematical 3.1.1-form of a proparoxytone, and only one line terminates in the less problematical 3.1.2-form. This is in stark contrast to the distribution of such lines in the corpus of sonnets ascribed to Wyatt in Tottel’s Miscellany. Of the 378 lines that make up his 27 sonnets in Tottel, as many as 44 terminate in a 3.1.1.-proparoxytone and, again, only one terminates in a 3.1.2-proparoxytone. We should not read too much into the fact that 3.1.2-paroxytones at line-final position are only represented by a single occurrence in each of our poets’ sonnets; this is merely a consequence of the relative scarcity with which such words occur in English in general.45 But we must pay due attention to the relative frequency with which 3.1.1-proparoxytones at line-final position are allowed to occur by each of our sonneteers: whereas 11.6 per cent of Wyatt’s lines terminate in a 3.1.1-proparoxytone, the same phenomenon is found in only 1.5 per cent of Surrey’s lines. In much the same way as the two instigators of the sonnet in English diverge from one another in terms of their preferred rhyme scheme, they also seem to diverge from one another on the issue of what constitutes a rhyme powerful enough to bind together a poetic form as complex as the sonnet. In many respects the poems by Wyatt and Surrey that are under discussion here are, of course, protosonnets in the sense that they are experiments aimed at establishing the sonnet in English rather than expressions of an already well-established form. It therefore seems reasonable to take a look at the phenomenon we have just examined as it occurs in the sonnets of slightly later writers. For this purpose I have selected four sonnet sequences published between 1591 and 1609; they are Philip Sidney’s Astrophel and Stella (1591), Michael Drayton’s Idea’s Mirror (1594), 43

O.B. Hardison, Jr., ‘Blank Verse before Milton’, in Studies in Philology, Vol. LXXXVI, No. 3 (Summer, 1984), 253274: 264. 44 As noticed by William Harmon in ‘English Versification: Fifteen Hundred Years of Continuity and Change’, in Studies in Philology, Vol. XCIV, No.1 (Winter, 1997), 1-37: 18-19. 45 See Chapter 7.

119

Spenser’s Amoretti (1595) and Shakespeare’s Sonnets (1609). From the three latter sequences, all of which contain almost exclusively sonnets with fourteen lines of pentameter, I simply selected the first 36 sonnets for analysis. In the case of Astrophel and Stella, in which work the definition of what constitutes a sonnet is somewhat looser, I selected the first 36 sonnets that consisted of fourteen lines in pentameter while skipping over all other sonnet-variants. The frequency with which lines terminate in the last syllable of a proparoxytone can be seen in this table:

Table 4.2 Sonnets Lines

3.1.1

3.1.2

Surrey

Tottel’s

14

196

Wyatt

Tottel’s

27

378 44 11.6% 1