sign in sign up
<<
Home , Amphibrach, Blank verse, Charles Lock, Decasyllable, Fourteener (poetry), Free verse

FACULTY OF HUMANITIE S UNIVERSITY OF COPENH AGEN

PhD thesis

Jesper Kruse

Free and the Constraints of in

English Poetry

Name of department: Department of English, Germanic and Romance Studies

Author: Jesper Kruse

Title: and the Constraints of Metre in

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, 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 64 Chapter 4. The Chinese Wall of Milton: Towards 95 Chapter 5. On Golden Hinges Turning: in Blank Verse 131 Chapter 6. To that Sweet Yoke Where Lasting Freedoms Be: Free Verse in English 167 Chapter 7. To Enumerate the : The Territory of Neither--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 , the metrical tradition that free verse reacted against goes back to the late fourteenth century when abandoned and introduced the . Over the following centuries, this particular form of accentual- 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 . 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. , T.S. Eliot and , 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 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 , 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 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 ... 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. 2 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.

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, , line, word and , 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. 4 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). 6

be taken downwards to syntax and 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 line- structure 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 ‘ 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 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. 9 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. 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 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. Kirby- Smith’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. 18 Beyers (2001): 223. 19 Kirby-Smith (1996): 37. 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 , 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 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 ’s . 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 of , 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 put out his Les Palais nomades, the first collection of 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, 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 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 , 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 , Urdu and Persian poetry. If a Greek syllable contained a long , a , or a short vowel followed by two or more (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 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 . Metrical feet consisting of one long syllable followed by two short syllables (‒ᴗᴗ) are called dactyls; those consisting of two long syllables (‒ ‒) are called . 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 . But dactyls and spondees were not allowed to mix completely at random in any single line of Homeric verse: the sixth foot of a was invariably a ,11 and the fifth foot was almost always a ; 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 , 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 , 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 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. (ᴗ‒) and (ᴗᴗ‒), while others belong to the province of scholars specializing in Greek verse, e.g. amphimacer (‒ᴗ‒) and (‒ ‒ ‒). 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 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 . 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 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 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 <ː>, which indicates that the preceding vowel is long, we can see that our line contains seven such : 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 ‒ ᴗ ᴗ ‒ ᴗ ᴗ ‒ ‒ ‒ 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 Indo- European 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 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 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 , 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 .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 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 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 . 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 (‘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:

Oft Scyld Scēfing sceaþena þrēatum, monegum mǣgþum, meodo-setla oftēah; egsode Eorl[e]. Syððan ǣrest wearð fēasceaft funden; hē þæs frōfre gebād: wēox under wolcnum, weorð-myndum þāh, oðþæt him ǣghwylc þāra ymb-sittendra ofer hron-rāde hȳran scolde, gomban gyldan: 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 .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 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 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- rather than alliteration in his Evangelienbuch (a work which would later influence , 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 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 , 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 verse begins, somewhat pathetically, with the all but complete annihilation of that Italic verse form whose demise is brazenly condoned in ’s Epistle to Augustus in the following manner:

34 Gasparov (1996): 40-43. 35 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. 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’ (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 , 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 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 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 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 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. 39 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. 27

into closer alignment with the intonational patterns of everyday speech, for which reason it was especially favoured in . 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 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. 43 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). 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 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. 46 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. 30

Other classical verse forms were better suited for this kind of metrical recasting, prime among them the iambic and the iambic : the resyllabized forms of these metres would evolve into the dodecasyllable and the 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 , 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 ‘ 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 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 , ‘[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 . 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 ‘ 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, , 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 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 stress- based 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. 15 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. 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 , 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 ’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: (‒ᴗ), iamb (ᴗ‒), spondee (‒ ‒), (ᴗᴗ), dactyl (‒ᴗᴗ), (ᴗ‒ᴗ), anapaest (ᴗᴗ‒) and (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 Fairy- Queen (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 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. 30 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. 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 :

│ . /│ . . /│. / │. . /│.. / │ 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 Sheridan (1775): 47-48. 36 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. 38 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. 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. 43 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 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. 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 , 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 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. 49 The Spectator, No. 160 (September 3, 1711). 50 Edward Young, Conjectures on Original Composition (London: 1759): 58. 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 , 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. 52 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. 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 . 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 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. 54 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 ’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. 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 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 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 , 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 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. 65 ‘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. 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 stress- based 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). 74 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. 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 <1> signifies minimal stress and <4> signifies maximal stress, while intermediary levels of stress are signified by <2> and <3>. Remarking that iambic 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. 77 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. 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 Halle & Keyser (1971): 169. 82 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 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 12 (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 cur few tolls the knell of part ing day

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. 85 Reuven Tsur, ‘Poetic Rhythm: Performance Patterns and Their Acoustic Correlates’, in Versification [an electronic journal of literary prosody], Vol. 1 (1997). 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 generative- linguistic 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. 87 Derek Attridge, ‘Rhythm in English Poetry’, in New Literary History, Vol. 21, No. 4 (Autumn, 1990), 1015-1037: 1016-1017. 88 Attridge (1990): 1017. 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 , a metre that does not allow for unrealized beats and in which there is a strong preference for single off-beats. Still, 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 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 –

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 [‘’] (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 , 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 . 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 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 . 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 Gasparov (1996): 130. 2 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 :

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 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 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 line- terminal 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 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. 13 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. 68

To the benefit of those of us whose knowledge of 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 – 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): 103- 125. 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. 31 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). 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 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 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. 38 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). 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. 42 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 phrase- terminal word in line-terminal position. 43 Preminger & Brogan (eds) (1993): 359-360, s.v. ‘Enjambment’. 44 Vising (1923): 85. 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 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 sub- segmentation 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 self- containment 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 .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 Cited from Wendy Ayres-Bennett, A History of the French Language through Texts (London: Routledge 1996): 41. 52 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 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 Cited from Ayres-Bennett (1996). 57 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 while the longer line-form became associated with : 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 extra- metrical 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. 64 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). 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 line- breaks were treated as equals inasmuch as both of these junctures allowed the addition of an extra- metrical 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. 69 Gasparov (1996): 129. 70 Gasparov (1996): 88-118. 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 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 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 . 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. 72 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. 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 ninth- century 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 .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. 81 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). 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 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 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 .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. 83 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. 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): 787- 788, s.v. ‘Metron’. 89 Maas [1903] (1973): 242-247. 90 West (1982): 185. 91 Cited from Gasparov (1996): 105 [my markings of the ]. 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 , 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 lu pa, ‖ 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) fait la geste ‖ que Turoldus declinet (4F + 6F)

1 2 3 4 5 ‖ 6 7 8 9 10 x x x x x ‖ x x x x x (x) (5 + 6) ed una lupa, ‖ che di tutte brame (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:

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.

97 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, 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 Ezra Pound, ‘The Renaissance: I – The Palette’, in Poetry, Vol. 5, No. 5 (Feb., 1915), 227-234: 231. 2 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. 4 On which see Donka Minkova, The History of Final Vowels in English: The Sound of Muting (Berlin & New York: Mouton de Gruyter, 1991). 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 verse- lines 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:

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 , 11 syllables)

7 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), 81- 103, & 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. 17 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. 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 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 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). 30 Cited from Wyatt (1969). 31 Cited from Edward Arber, Tottel’s Miscellany, English Reprints (Westminter: A. Constable, 1897). 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 , 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 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 a spiring (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 line- final 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 ; its other function was to provide the standard medium for dramatic verse from the late sixteenth century up until the closing of the London 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 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. 38 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). 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 3.1.1 visible /ˈvɪz.ᵻ.bl / Allowed at position 10 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 3.1.1- proparoxytones 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: , 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, proto- sonnets 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), 253- 274: 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 3 1.5% 1 0.5% Wyatt Tottel’s 27 378 44 11.6% 1 <0.5%

Sidney Astrophel and Stella 36 504 25 5.0% 1 <0.5% Drayton Idea’s Mirror 1-36 36 504 19 3.8% 3 0.6% Spenser Amoretti 1-36 36 504 21 4.2% 1 <0.5% Shakespeare Sonnets 1-36 36 504 23 4.6% 0 0%

As can be seen, the relative frequency of lines terminating in the final syllable of a 3.1.1- proparoxytone in the four extracts from our sonnet sequences is relatively uniform, ranging from 3.8 per cent to 5.0 per cent. This would seem to indicate that the sonnet-form, as it became gradually codified in English, stabilized itself somewhere in between the two extremes represented by Surrey and Wyatt, neither shunning nor overusing line-terminal proparoxytones. Interestingly, the three instances of line-terminal 3.1.1-proparoxytones in Surrey all occur in the same sonnet, namely ‘The great Macedon’, whose epigraph reads ‘Praise of certain psalmes of Dauid, translated by sir. T[homas Wyatt] the elder’. Even without jumping to the premature conclusion that Surrey is simply paying tribute to Wyatt’s style by including a couple of his trademark line-terminal proparoxytones in this particular poem, the fact still remains that in all other sonnets by Surrey 3.1.1.-proparoxytones are completely absent in line-terminal position.46 Given the relatively stable frequency with which such words occur line-finally in the extracts from

46 3.1.1-proparoxytones are rare in Surrey’s sonnets in any position, but an extract such as this from ‘Thassirian king in peace’ (ll. 10-11) demonstrates that he had no difficulty integrating them into his pentameters: ‘Drenched in slouth, and womanish delight, / Feble of sprite, impacient of pain’.

120

Sidney, Drayton, Spenser and Shakespeare, it does not seem unreasonable to conclude that their all but total absence in Surrey’s sonnets must be the result of a deliberate stylistic choice: apparently Surrey simply disliked the types of rhymes that such line-ends tend to produce and therefore took pains to avoid them. If this indeed was the case, then his decision to forego rhyme in his translation of the Aeneid may in part have been motivated by the necessity of using proparoxytones more freely. An analysis of the 943 lines that make up Surrey’s translation of Book 4 of the Aeneid thus reveals that a total of 48 lines terminate in the final syllable of a proparoxytone, yielding a relative frequency of 5.1 per cent. This, of course, is very close to the frequency with which such lines occur in the four sonnet sequences that we looked at before (3.8 – 5.0 per cent), though still considerably lower than what is found in Wyatt (11.6 per cent). The mere suggestion that the introduction of blank verse in English might have been contingent upon a single poet’s idiosyncratic dislike of the rhymes produced by line-terminal proparoxytones has, admittedly, an outlandish ring to it. However, before it is dismissed as too fanciful to be taken seriously, it might be prudent to heed the following observation made by John Hollander in Vision and Resonance:

The crucial relation between the effects of word stress and the quality of rhyme in English has not to my knowledge been adequately considered. The rhetorical power of end rhyme in English ... depend in good measure on the necessity of the rhyming syllable’s being an accented one.47

Furthermore, it ought to be noted that Surrey is not the only English poet to have avoided the occurrence of 3.1.1-proparoxytones in line-final position. Samuel Johnson, for instance, seems to have been keenly aware of the problems entailed by line-final proparoxytones in rhyming pentameter and to have made use of a similar strategy of avoiding them. Thus, of the 263 lines of Johnson’s ‘London’ (1738) only three terminate in a proparoxytone (relative frequency 1.1%), and ten years later when Johnson published the 368 lines that make up ‘The Vanity of Human Wishes’ (1749) he avoided line-final proparoxytones altogether. In this he was followed by poets such as Oliver Goldsmith (zero occurrences in the 430 lines of ‘The Deserted Village’), George Crabbe (zero occurrences in the 553 lines of ‘The Village’) and William Cowper (zero occurrences in the 908 lines of ‘Conversation’). All of these poems are, of course, written in rhyming couplets, a form

47 John Hollander, Vision and Resonance: Two Senses of Poetic Form (New York: Oxford University Press, 1975): 118. 121

intricately associated with in general and with the works of Dryden and Pope in particular. These latter poets, however, did not refrain from spicing up their couplets with the occasional line-terminal proparoxytone, as can be seen from these examples:

Achitophel, grown weary to possess A lawful fame, and lazy happiness, (Dryden, Absalom and Achitophel, ll. 200-201)

And every shekel which he can receive Shall cost a limb of his prerogative. (Dryden, Absalom and Achitophel, ll. 391-392)48

Unfinished things, one knows not what to call, Their generation’s so equivocal: (Pope, An Essay on Criticism, ll. 42-43)

All this dread ORDER break—for whom? for thee? Vile worm!—oh madness, pride, impiety! (Pope, An Essay on Man, ll. 257-258)49

Thus, the same seemingly marginal feature that sets apart the rhyming practice of Surrey from that of Wyatt and the rest of the Renaissance sonneteers – i.e. the relative frequency of line-terminal proparoxytones – can also be called upon to provide a dividing line between the golden age of Augustan Poetry and the silver age which followed in its wake. Furthermore, as soon as Romanticism begins to rear its head, line-final proparoxytones begin to pop up again in the rhymed verse of all of the most celebrated poets of this period. Surrey’s apparent dislike of line-final proparoxytones in rhymed verse is, in other words, less idiosyncratic than it first seems: other major poets, even schools of poetry, have shared his lack of enthusiasm for the contorted kinds of rhyme this phenomenon tends to produce.50

48 Cited from John Dryden, Selected Poems, edited by Paul Hammond & David Hopkins (Harlow: Pearson, 2007). 49 Cited from Alexander Pope, The Major Works, edited by Pat Rogers (Oxford: Oxford University Press, 2006). 50 Other poets revelled in these rhymes for their couplets: in addition to Wyatt, one might think of Donne or Keats, both of whom, we note, occupy a far more secure position in today’s canon than do the poets who took pains to avoid such rhymes. 122

But even if Surrey’s decision to forego rhyme in his translation of Virgil was in part motivated by the necessity of including words of a more complex prosodic structure into his poetic vocabulary without offending his ear, such an explanation is hardly sufficient. It may shed light on why Surrey would have found the notion of rhymeless verse attractive, but at the same time it fails to account for the manner by which he may have been led to such a drastic solution to his problem. The contention voiced by Herbert Hartman in his 1933 edition of the Day version of Aeneid 4, that Surrey’s choice of blank verse was the ‘product of a strictly English humanism’51 which owed nothing to foreign influence, ignores the fact that the same Italy which had provided Wyatt and Surrey with a model for their sonnets may also have provided the latter poet with a model for blank verse. The central figure in the development of the Italian equivalent of blank verse, versi sciolti, is Pietro Bembo’s old adversary on the question of Italy’s national language, Gian Giorgio Trissino, whose Sofonisba (1515) is regarded not only as the first regular tragedy in any European vernacular but also as the first work to make extensive use of unrhymed .52 Trissino remoulded the same verse design – i.e. unrhymed hendecasyllables – to fit epic rather than dramatic purposes for his Italia Liberata dai Goti, which was published in 1547, but by then versi sciolti had already been used by other Italian writers as a medium for composing and translating epic verse.53 However, the significance of Trissino’s works in versi sciolti is not confined to their literary value, but also extends to the detailed theoretical considerations with which these works were prefaced.54 Anglophone readers unversed in the intricacies of Renaissance Italian remain indebted to O.B. Hardison, Jr. for having tracked down and translated Trissino’s prefatory remarks to his two most influential works in versi sciolti, and also for the conclusion that Hardison reaches based on his readings of these documents: both bear repetition in this context. Here, first, is Hardison’s

51 Herbert Hartman (ed.), Surrey’s Fourth Boke of Virgill (New York, 1933): xxvi. 52 Hardison (1984): 259-260. Trissino’s role as instigator of blank verse in European literature was already recognized in the Anglophone world by the mid-eighteenth century as is evident from Thomas Gray’s fiercely erudite ‘Observations on English Metre’, in which blank verse is explained as a decasyllabic measure ‘Without Rhyme. (Versi Sciolti of the Italians.) The invention is attributed to Trissino, about the year 1525.’ (Thomas Gray, Gray’s Poems, Letters and Essays (London: J.M. Dent & Sons, 1955): 337). Samuel Johnson in Lives of the Poets also mentions Trissino as the most likely source of influence on Milton’s use of blank verse in Paradise Lost. However, as early as 1735, Thomas Blackwell, explaining the principles of Italia Liberata, says that ‘[Trissino] abandoned the use of Rhyme [and] followed the natural Run of Speech in his Verse’ (Blackwell (1735): 32). 53 E.g. by Hippolito de Medici, Nicolo Liburnio and Luigi Alamanni. The latter has been suggested as an influence on Surrey’s blank verse by Edwin Casady; the two former by F.M. Padelford (for references, see Hardison 1984: 259, nn. 17-18). Both suggestions are, of course, prompted by the fact that since Surrey’s translation of Virgil was complete by 1540, it cannot have been influenced by Trissino’s epic use of versi sciolti in Italia Liberata (1547). 54 For general discussion of the theoretical framework in which versi sciolti appeared, see John M. Steadman, ‘Verse without Rime: Sixteenth-Century Italian Defences of Versi Sciolti’, in Italica, Vol. 41, No. 4 (Dec., 1964), 384-402. 123

translation of Trissino’s comments on the function of versi sciolti in tragedy, from the preface to Sofonisba, addressed to Pope Leo X:

I do not think that it can justly be considered a defect that [Sofonisba] ... does not use rhyme, as is commonly done, but is free in many places. ... [verso sciolto] is better and more excellent and perhaps less easy to write than it may be considered to be. And you will see that it is most useful not only in narrations and formal speeches but in the arousal of the piteous emotion that is necessary [in tragedy]. And this is because the speech that arouses such emotion is born from painful emotions, and painful emotion does not express itself in carefully thought-out words; and rhyme, which expresses careful thought, is, in fact, directly opposed to the expression of pity.55

Trissino’s argument here that the pensiveness required for crafting rhymes is not congenial with the expression and invocation of piteous emotions quite obviously rests on a theoretical notion of tragedy that has been borrowed directly from Aristotle. However, for his argument against rhyme in , Trissino relies on another Greek authority, namely Demetrius of Phalerum, whose concept of enargia forms the cornerstone of the preface to Italia Liberata. Hardison paraphrases Trissino’s argument in that preface thus:

Enargia (vividness) is essential to epic poetry. It is created by the particularizing of descriptions, which requires that the descriptive units be extended to the full length required by the subject. Poetic enargia therefore depends on an open verse form (i.e., versi sciolti) which does not, like couplets and stanzas, break the description into arbitrary units.56

It could, of course, be argued that the line is in itself an arbitrary unit – indeed the principal arbitrary unit that sets apart verse from prose – but Trissino’s point seems to be that it would be counterproductive to the purpose of epic verse to emphasize this fact by means of rhyme. The reason for this is, I believe, grounded in what may well be described as the fundamental paradox of rhyme. On the one hand, rhyme has the effect of undermining the line’s autonomous status by making it more prone to take part in the formation of larger prosodic hierarchies: ‘The bridging,

55 Hardison (1984): 261. 56 Hardison (1984): 260. 124

associating, linking function of rhyme is a dialectical turn upon its ability to handcuff. Rhyme links syllables, and thereby words, and thereby lines, and thereby larger versified structures’, John Hollander observes.57 On the other hand, rhyme also has the opposite effect, namely that of consolidating the metrical integrity of the line, thereby encouraging a kind of versification in which syntax becomes prone to subordinate itself to the measure of the line. In this latter respect Trissino’s admonition against rhyme in the preface to Italia Liberata prefigures Milton’s prefatory note on ‘The Verse’ of Paradise Lost, in which blank verse is described as championing ‘the sense variously drawn out from one Verse into another’.58 Hardison concludes:

To Trissino, in other words, verso sciolto is two quite different meters. As epic metre it permits the enjambment and verse paragraphing needed for heroic enargia. On the other hand, in drama it expresses the rapid movements and disordered sequences of intense emotion appropriate for dialogue. ... From the humanist point of view [heroic prosody and dramatic prosody] are separate, and although they may use the same form in vernacular languages because of the peculiarities of these languages, as imitations they express the qualities of different meters with different functions.59

By positioning the vexed question of blank verse in the humanist context in which the form(s) originally emerged rather than in the more linguistically oriented modern understanding of metrics, Hardison succeeds in elucidating, not only why the two main uses of blank verse developed independently from one another in both Italian and in English literature, but also why the two forms tend to be perceived so differently by most readers. The latter point is of no small importance for the larger purpose of our investigation, which is to outline the implications of blank verse for the Modernists’ metrical experiments at the beginning of the twentieth century. Their stance towards blank verse was, more than anything, one of fundamental ambivalence: on the one hand, we have Ezra Pound’s condemnation of Milton as ‘the worst sort of poison’, and on the other hand we have Eliot’s unveiled adulation for the verse practice of the Elizabethan dramatists: ‘the blank verse of their plays accomplished a subtlety and consciousness, even an intellectual power, that no blank verse since has developed or even

57 Hollander (1975): 119. 58 Milton, Note on ‘The Verse’ in Paradise Lost. 59 Hardison (1984): 262. 125

repeated’.60 In his essay on Christopher Marlowe’s blank verse, Eliot even has a paragraph whose opening sentences might have been cited as the epigraph of this chapter, instead of being merely alluded to in the chapter title:

The comparative study of English versification at various periods is a large tract of unwritten history. To make a study of blank verse alone, would be to elicit some curious conclusions. It would show, I believe, that blank verse within Shakespeare’s lifetime was more highly developed, that it became the vehicle of more varied and more intense art- emotions than it has ever conveyed since; and that after the erection of the Chinese Wall of Milton, blank verse has suffered not only arrest but retrogression. That the blank verse of Tennyson, for example, a consummate master of this form in certain applications, is cruder (not “rougher” or less perfect in technique) than that of half a dozen contemporaries of Shakespeare; cruder, because less capable of expressing complicated, subtle, and surprising emotions.61

One curious conclusion elicited by our own study of blank verse in this chapter is that Eliot’s conception of blank verse comes off as decidedly less sophisticated than that of Trissino. Eliot clearly judges the merits of dramatic and epic blank verse by the same yardstick – i.e. its ability to represent ‘intense art-feelings’ – and consequently and unsurprisingly finds Milton’s blank verse wanting. In fact, Eliot’s argument closely resembles the argument put forward by Trissino in his preface to Sofonisba, but whereas Trissino’s comments on dramatic blank verse are tempered by his erudite considerations of the same measure’s use in epic poetry, Eliot can only muster a sly comparison of Milton’s blank verse to the Chinese Wall. Still, that comparison is apt enough. It invokes not only the meandering sentence structure which characterizes Paradise Lost, but also the numinous enargia of that poem and the sense of impenetrable otherness it tends to instil in its readers: ‘None ever wished it longer than it is’, Samuel Johnson once remarked of Paradise Lost. Presumably the same goes for the Chinese Wall.

❦

60 T.S. Eliot, ‘The Possibility of a Poetic Drama’, in Eliot (1920), 60-70: 62. 61 T.S. Eliot, ‘Notes on the Blank Verse of Christopher Marlowe’, in Eliot (1920), 86-94: 87. Judging from the wording of the extract cited, Eliot appears unaware of the existence of John Addington Symonds’s Blank Verse (New York: Scribner, 1895), the first – and up until the publication of Shaw (2007) also the only – book-length study dedicated entirely to blank verse in English. 126

If the merit of a new literary form is measured solely by its immediate success and popular propagation, then Surrey’s invention of blank verse as a vehicle for translating and composing in English stands out as an abject failure. With the notable exception of Marlowe’s translation of the first book of Lucan’s Pharsalia published in 1600 and a couple of short extracts from the Aeneid and the Georgics translated by Thomas Norton and published in 1561, blank verse was not employed for lengthy translations of non-dramatic poetry until the early eighteenth century when Nicholas Brady and Joseph Trapp issued their blank verse translations of the Aeneid in 1716 and 1718, respectively.62 For a complete blank verse translation of Homer, English readers would have to wait until 1791 when William Cowper’s translations were published. Other uses of non- dramatic blank verse between Surrey and Milton are equally scant: Nicholas Grimald has two short pieces in blank verse in Tottel’s Songes and Sonnettes (‘The Death of Zoroas’ and ‘Marcus Tullius Ciceroes Death’), and then there is George Gascoigne’s idiosyncratic satire The Steele Glass, published in 1576, also in blank verse. Thus, when Milton concludes his note on ‘The Verse’ of Paradise Lost with the claim that his poem ‘is to be esteem’d an example set, the first in English, of ancient liberty recover’d to Heroic Poem from the troublesom and modern bondage of Rimeing’, he is not necessarily overstating his case: it is far more likely that Milton’s claim is an indication of the extent to which earlier instances of non-dramatic blank verse in English had fallen into oblivion by the late seventeenth century.63 Milton is thus keenly aware that ‘some both Italian and Spanish Poets of prime note have rejected Rime both in longer and shorter Works’ and adds that the same is true of ‘our best English Tragedies’.64 Though Norton and Sackville’s play Gorboduc (1561) hardly belongs in this category, it merits mention here because it is the earliest example of dramatic blank verse in English. Heavily inspired by the rediscovery of Senecan tragedy in the sixteenth century, Gorboduc combines

62 In the period between Surrey’s partial blank verse translation of the Aeneid and the complete ones of Brady and Trapp, Virgil’s epic was, in fact, subjected to an impressive variety of metrical forms in English translations: disregarding Caxton’s paraphrase in prose (1490), the Aeneid saw translations into English heroic couplets (by Gavin Douglas and Dryden), hexameters (by Richard Stanyhurst (1582)), fourteeners (by Thomas Phaer & Thomas Twyne (1573)), ottava rima (by John Harrington), and even Spenserian stanzas (by Richard Fanshawe). The first writer to pick up on Surrey’s aspirations to use blank verse as the appropriate medium for translating the Aeneid into English was Joseph Addison, whose translation in unrhymed decasyllables of that part of the Aeneid where Achaemenides is marooned on Sicily alongside a less than charitable Cyclops was published in 1704 under the title of ‘Milton’s Style Imitated in a Translation of a Story out of the Third Aeneid’. 63 Ignorance of pre-Miltonic epic blank verse in English in the eighteenth century is suggested by the uncertainty inherent in Samuel Johnson’s remark that ‘The earl of Surrey is said to have translated one of Virgil's books without rhyme’ (Lives of the Poets, Milton). 64 Milton, Note on ‘The Verse’ of Paradise Lost. 127

staunch rhetorical declamations with elaborate dumb-shows in a fashion that could indicate inspiration from Trissino’s Sofonisba,65 although it bears little resemblance to what is usually thought of today as Renaissance drama.66 The influence from Seneca is still very much present in Thomas Kyd’s Spanish Tragedy (1580s), which shares with Marlowe’s Tamburlaine the Great (1587) the honour of being the first English play to make effective use of dramatic blank verse. Marlowe in particular seems highly conscious of the expressive potential made possible by the sheer novelty of his chosen verse design, as may be inferred from the opening lines of the Prologue to Tamburlaine:

From jigging veins of rhyming mother-wits And such conceits as clownage keeps in pay, We’ll lead you to the stately tent of War, Where you shall hear the Scythian Tamburlaine Threat’ning the world with high astounding terms And scourging kingdoms with his conquering sword.67

In spite of this formal awareness on the part of Marlowe, the air of declamatory bluster that permeates Gorboduc is still looming over these early works of Marlowe and Kyd, as Eliot pointed out when he declared that the later Elizabethan and Jacobean drama had ‘grown away from the rhetorical expression, the bombast speeches, of Kyd and Marlowe to the subtle and dispersed utterance of Shakespeare and Webster’.68 John Webster is of singular importance to the discussion of the Modernists’ indebtedness to Renaissance blank verse. After all, it was he who provided Eliot with much of the material on which his charge against free verse in ‘Reflections on Vers Libre’ relies:

At the beginning of the seventeenth century, and especially in the verse of John Webster, who was in some ways a more cunning technician than Shakespeare, one finds the same constant evasion and recognition of regularity [as one finds in vers libre]. Webster is much

65 Hardison (1984): 259-260. 66 The direct influence of Gorboduc on English drama appears to have been limited to Gascoigne’s and Francis Kinwelmershe’s joint blank verse translation (1566) of Lodovico Dolce’s tragedy Giocasta (1549), itself in verso sciolto. 67 Cited from Christopher Marlowe, Tamburlaine the Great, edited by J.S. Cunningham (Manchester: Manchester University Press, 1981). 68T.S. Eliot, ‘“Rhetoric” and Poetic Drama’, in Eliot (1920), 78-85: 80. 128

freer than Shakespeare, and that his fault is not negligence is evidenced by the fact that it is often at moments of the highest intensity that his verse acquires this freedom.69

In addition to confirming Eliot’s view of blank verse as a monolithic form, whose ultimate merit is to be judged solely by its ability to express intense emotion, these scattered comments also remind us of the remarkable pace at which dramatic blank verse developed upon its introduction on the English stage. Only 25 years separate Webster’s best known plays from the earliest tragedies of Kyd and Marlowe, and during this short period dramatic blank verse underwent more radical changes than its non-dramatic counterpart has ever been subjected to since its invention by Surrey. The austere syllabic regularity of Paradise Lost should thus not be interpreted as a retrogression to a pre-Elizabethan formalism, as the citations from Eliot might be taken to imply, but rather as an attempt to codify in English a specifically epic brand of blank verse, in contrast to and distinct from its dramatic form. It will be recalled that when Surrey abandoned rhyme for his translation of the Aeneid, the ratio of lines terminating in proparoxytones in his verse went up from virtually nil to a level of about five per cent. This level was found to be comparable to the average ratio of line-terminal proparoxytones in a number of Elizabethan sonnet sequences (see Table 4.2), while still much lower than what is found in Wyatt’s verse (11.6 per cent). Thus, even when the requirements of rhyming are removed from the equation, Surrey’s employment of proparoxytones in line-terminal position remains comparatively frugal, especially – as we shall see – in comparison with the practice in dramatic blank verse. The paradigmatic figure to consider in this connection is, of course, Marlowe in that he alone among the Elizabethan and Jacobean poets tried his hand at both dramatic and epic blank verse. Marlowe’s translation of the first book of Lucan’s Pharsalia is the only lengthy piece of epic blank verse between Surrey and Milton and therefore of pivotal importance to our present concern. An analysis of its 751 lines reveals that some 5.6 per cent of these terminate in a proparoxytone; this ratio is, we note, only marginally higher than what we found in Surrey’s translation of Aeneid 4 (5.1 per cent). However, as soon as one moves from Marlowe’s Pharsalia to his dramatic works in blank verse, a remarkable shift occurs in the frequency of proparoxytones in line-final position. In the first 500 pentameters of Tamburlaine the Great, as many as 129 lines terminate in a proparoxytone, yielding a relative frequency of 25.8 per cent, and in the first 500 pentameters of The Jew of , the ratio is 22.0 per cent. The fact that lines ending in a

69 Eliot (1969): 186. 129

proparoxytone occur at least four times more frequently in Marlowe’s dramatic blank verse than they do in his epic blank verse makes it prudent to investigate the frequency with this feature occurs in other dramatic works from the same period. An analysis of the 500 opening iambic pentameters of a random selection of such works yields the following results:

Table 4.3 Marlowe Tamburlaine the Great 25.8% Marlowe The Jew of Malta 22.0% Thomas Kyd Spanish Tragedy 14.6% Mary Sidney The Tragedy of Antonie 12.8% Shakespeare Richard II 9.6% Shakespeare Julius Caesar 12.4%

As can be seen, Marlowe’s reliance on line-terminal proparoxytones in Tamburlaine the Great and The Jew of Malta really is exuberant, even by dramatic standards, but this just makes the comparatively low ratio of the same phenomenon in Pharsalia even more conspicuous. Evidently, Marlowe’s blank verse exhibits radically different characteristics depending on whether it is being used for dramatic or epic purposes. At the same time, however, we should note that the frequency with which proparoxytones occur in line-terminal position in the other dramas examined above is consistently much higher than what we found in the sonnet sequences detailed in Table 4.2. Shakespeare, for instance, more than doubles his usage of such lines in Richard II and Julius Caesar when compared to his Sonnets (4.6 per cent), without, however, going to the same extreme as Marlowe. So, whereas dramatic blank verse in English is characterized by exploiting how the absence of rhyme can admit proparoxytones in line-final position, epic blank verse – as explored by Surrey and Marlowe – seems to resist the same phenomenon. In spite of the fact that Milton must have been unaware of Surrey’s and Marlowe’s use of epic blank verse when he composed Paradise Lost, he appears to have inadvertently adopted his predecessors’ strategy of keeping the ratio of line- terminal proparoxytones low. Surprisingly low, in fact. An analysis of the first 500 lines in each book of Paradise Lost yields the following ratios of lines terminating in proparoxytones:

130

Table 4.4 Book I II III IV V VI VII VIII IX X XI XII Ratio 6.6% 2.6% 3.2% 2.6% 2.8% 2.2% 3.8% 4.8% 4.2% 6.0% 3.8% 4.0%

This amounts to an average ratio of a mere 3.9 per cent for the work in its entirety,70 i.e. a level on par with or even lower than the ratios found in the sonnet sequences detailed in Table 4.2. Given the much commented on Latinate and polysyllabic character of Milton’s diction in general, it is somewhat surprising to discover the restraint under which polysyllables of the proparoxytonic kind are allowed to occur in line-final position in Paradise Lost: on average, only one in twenty-five lines in Milton’s poem terminates in a proparoxytonic word. However, as we shall see in the next chapter, Milton’s predilection for ending his lines with a lexically stressed syllable has other and more far-reaching consequences for his style of versification, especially in the area of enjambment.

70 John Creaser reaches almost the same result concerning the frequency of line-terminal proparoxytones in Paradise Lost, although based on a much smaller sample: ‘of the 1,200 lines [of Paradise Lost] examined in especial detail for this study, only fifty-one [i.e. 4.25 per cent] have a line-end promotion (on words such as ‘argument’ and ‘providence’ in the opening invocation); in the other 95.75 per cent. of the lines, the fifth beat falls either on a stressed monosyllable or a disyllable with second-syllable stress (‘“Service is Perfect Freedom”: Paradox and Prosodic Style in Paradise Lost’, The Review of English Studies, New Series, Vol. 58, No. 235 (June, 2007), 268-315: 293-294). Creaser’s terminology is derived from Attridge’s system of beat scansion. 131

Chapter 5 On Golden Hinges Turning: Enjambment in Blank Verse

the gate self-opend wide On golden Hinges turning, as by work Divine the sov’ran Architect had fram’d.

John Milton, Paradise Lost

Before embarking on our investigation of how enjambment is utilized in dramatic and epic blank verse respectively, it may be wise to remind ourselves that – in spite of the prominent role that line- final proparoxytones were accorded in the previous chapter – such words are, of course, of limited interest in and by themselves. The principle reason for bestowing so much attention on them in this dissertation has been to discover what their different frequencies in different types of metrical verse might indicate. A low ratio – such as one finds in the epic blank verse of Surrey, Marlowe and, especially, Milton – indicates a verse design in which the structural integrity of the pentameter is reinforced and asserted by means of lexically stressed syllables in the crucial tenth position that carries the fifth beat. A high ratio, on the other hand – such as we find in Elizabethan dramatic blank verse – indicates a verse design in which the pentameter’s five-beat structure is loosened and made less palpable. In light of this, we cannot help but take issue with Samuel Johnson’s frequently cited criticism of Milton’s blank verse:

The musick of the English heroick line strikes the ear so faintly that it is easily lost, unless all the syllables of every line co-operate together; this co-operation can be only obtained by the preservation of every verse unmingled with another as a distinct system of sounds, and this distinctness is obtained and preserved by the artifice of rhyme. The variety of pauses, so much boasted by the lovers of blank verse, changes the measures of an English poet to the periods of a declaimer; and there are only a few skilful and happy readers of Milton who 132

enable their audience to perceive where the lines end or begin. ‘Blank verse,’ said an ingenious critick, ‘seems to be verse only to the eye.’1

Given the fact that the ratio of lines which do not terminate in a lexically stressed syllable is at least as low in Milton’s blank verse as it is in rhymed verse up until the Augustans, it must be admitted that Johnson’s criticism of blank verse would be far more pertinent to dramatic blank verse than it is to Milton’s heroic use of it. Milton, as we have just seen, goes to great lengths in order to secure the structural stability of his pentameters by getting all of his syllables, particularly those occupying the tenth position, to cooperate together in the formation of lines that resist prosodic intermingling. Johnson’s oddly structuralist stipulation that a line of verse should make up ‘a distinct system of sounds’ is therefore fully met by Milton’s verse design in Paradise Lost, even without the aid of ‘the artifice of rhyme’. In fact, what John Creaser observes about select passages from Paradise Lost may justly be extended to the poem in its entirety, namely that ‘the verse of Paradise Lost maintains a driving clarity, line by line ... and each line is felt as an independent rhythmical presence’.2 Another way of putting this – one more consonant with the terminology used in this dissertation – would be to say that the verse design of Paradise Lost is such that the vast majority of its verse instances hold at least the potential of being realized as individual and independent prosodic segments at the level of delivery. This contention, regardless of how it is formulated, would almost certainly have been rejected by Eliot, who not only considered Johnson ‘a dangerous person to disagree with’,3 but also made the following observation on Milton’s versification with direct reference to the passage from Lives of the Poets just cited:

It seems to me ... that Milton’s verse is especially refractory to yielding up its secrets to examination of the single line. For his verse is not formed in this way. It is the period, the sentence and still more the paragraph, that is the unit of Milton’s verse; and emphasis on the line structure is the minimum necessary to provide a counter-pattern to the period structure.4

1 Samuel Johnson, Lives of the English Poets, edited by George Birkbeck Hill (Oxford: At the Clarendon Press, 1905), Vol. I: 192-193. 2 Creaser (2007): 293. 3 T.S. Eliot, ‘The Metaphysical Poets’ [1921], in William R. Keast (ed.), Seventeenth-Century English Poetry: Modern Essays in Criticism (New York: Oxford University Press, 1962), 22-30: 30. 4 ‘Milton II’, in T.S. Eliot, On Poetry and Poets (London: Faber and Faber, 1957), 146-161: 157. 133

Eliot made this statement in what may be regarded as his single most important contribution to the field of inquiry with which this dissertation is concerned, his 1947 lecture on Milton to the British Academy, published ten years later in a slightly abridged form as ‘Milton II’.5 In that paper, Eliot recants some, and modifies others, of his earlier attacks on Milton’s verse and its harmful influence on the English language. It is a paper to which we shall find reason to return again later in this dissertation. First, though, we must consider Eliot’s denunciation of the line as a meaningful constituent in Milton’s verse, with particular reference to Paradise Lost. Eliot’s downplaying of the constitutive role of the individual line in Milton’s verse is at once right and wrong. He is right in claiming for the sentence and the paragraph the central role in establishing the particular character – as distinct from principle – of Milton’s versification in Paradise Lost, and also in observing that line structure functions as a counter-pattern to these period structures. But, even though Eliot is himself a dangerous person to disagree with, I believe he is wrong to claim that attention to the unit of the line and its internal structure is merely ‘the minimum’ required for counter-patterning the work’s syntactic structures. Rather, Milton’s disciplined verse design is a necessary precondition for his daring syntax. What Eliot seems determined to overlook is that Milton’s radical use of enjambment, which is responsible for creating that ‘physical sensation of a breathless leap’6 that Eliot senses in Milton’s long periods, depends entirely on a sufficiently strict verse design at the level of the individual verse instance. For Milton, syntactic freedom in metrical verse – the drawing out of the sense from one verse into another – can only be successfully accomplished when it is tempered by a firm metrical pattern, by apt numbers and fit quantity of syllables: ‘Milton’s prosody is at the heart of [Paradise Lost]’, observes Creaser, ‘The very movement of his verse enacts paradox in its fusion of the integrity of the individual line with the unpredictability of continuous enjambment’.7 In this sense, the syntactic freedom of Milton’s verse – his frequent enjambments and versatile caesuras – is but a more radical version of the freedom that Chaucer gained when he tightened up the verse design of his predecessors and thereby broke the syntactic straitjacket of the couplet. In fairness, it should be noted that the observations by Eliot that we have just discussed were put forward as a comment on ’s treatise on Milton’s Prosody (1889, revised edition 1921). That work – still considered a classic in its field – marked the beginning of an ongoing feud

5 The lecture was repeated at the Frick Museum in New York later that year, and the complete version of the lecture can be found in James Thorpe (ed.), Milton Criticism: Selections from Four Centuries (London: Routledge & Kegan Paul, 1951), 310-332. 6 Eliot (1957): 158. 7 Creaser (2007): 315. 134

between Bridges and Saintsbury through the first decades of the twentieth century concerning the metrical principles of Milton’s versification, and it may very well be from this pedantic feud that Eliot was seeking to distance his own comments. Whereas Saintsbury favoured foot-substitution as the key to analyzing Milton’s aberrant lines, Bridges posited ‘a decasyllabic line on a disyllabic basis and in rising rhythm’8 as the basic verse design of Paradise Lost and then proscribed a detailed system of to account for extra syllables in individual verse instances. The historical account of metrical theory given in this dissertation’s Chapter 2 strongly favours Bridges’s point of view, and Eliot readily admits that he ‘can find no fault with [Bridges’s] analysis’. Even so, Eliot insists that Milton’s verse is ‘not formed’ line by line, and furthermore adds that the detailed analysis of that level of verse structure contributes nothing to the ‘appreciation of the peculiar rhythm of a poet’.9 As Eliot’s comments stand, they oscillate uneasily between addressing the level of verse design and the level of delivery, so perhaps it would have been closer to the mark to say that Paradise Lost should not be performed line by line. While we can agree with Eliot that ‘the period, the sentence and still more the paragraph’ must be given their due in any effective delivery of Paradise Lost, we must object that these same structures cannot meaningfully be said to constitute ‘the unit’ of that poem’s verse design. The basic unit of Milton’s design in Paradise Lost – as of any other composition in verse – remains that of the line, and the extent to which a poet chooses to reify or subvert this unit syntactically has no bearing on its constituent status in the verse design. If anything, the relationship between syntax and line structure, at least in non-dramatic verse, appears to be such that syntactic freedom across line breaks tends to presuppose a strict and well-defined line structure, whereas a looser line structure tends to encourage syntactic reinforcement of line breaks.

❦

Probably the most notable exception to this general rule of thumb is the decasyllabic verse of , whose metrical ruggedness, especially in his Satires, is frequently accompanied by violent enjambments. Donne, of course, was singled out for apotheosization as the model poet of the seventeenth century in Eliot’s essay on ‘The Metaphysical Poets’, whereas Milton’s verse,

8 Robert Bridges, Milton’s Prosody with a Chapter on Accentual Verse & Notes, Revised, Final Edition (Oxford: Oxford at the Clarendon Press: 1921): 1. For Saintsbury’s view and his criticism of Bridges, see his History of English Prosody, Vol. 2: 257-270. 9 Eliot (1957): 157. 135

alongside Dryden’s, was held responsible for having introduced into English poetry a regrettable dissociation of sensibility. Here, however, we are primarily concerned with that dissociation of grammatical constituents which enjambment by its very nature introduces. Among the few lines by Donne that Eliot actually cites (albeit for other reasons than those for which they are cited here) in his influential essay are the following from ‘A Valediction: On Weeping’:

Till thy teares mixt with mine doe overflow This world, by waters sent from thee, my heaven dissolved so.10

The transition from the first of these lines to the second exemplifies that especially happy specimen of enjambment which instils in the reader a brief moment of syntactic anagnorisis:11 not until we commence our reading of the latter line do we realize that ‘overflow’ in this particular constellation is not an intransitive verb and that it takes as its direct object the absurdly hyperbolic noun phrase ‘This world’. The effect is made even more profound, or at least more iconic, by the fact that it is prompted by the ambitransitivity of the verb overflow; overflow being exactly what the syntactical structure of Donne’s sentence is allowed to do across this crucial line break.12 An equally happy specimen of enjambment, which relies on a related syntactic ambivalence for its effect, can be found at the beginning of Book 4 in Paradise Lost, when Satan falls prey, however briefly, to an unwelcome mixture of terror and regret atop Mount Niphates:

now conscience wakes despair That slumbered, wakes the bitter memory Of what he was, what is, and what must be Worse; of worse deeds worse sufferings must ensue. (PL 4, 23-26)

10 Cited from Eliot [1921]. 11 In using the term anagnorisis to designate that element of suspended grammatical understanding which certain types of enjambment can procure, I am indebted to Hollander (1975). 12 The aptness of the term overflow to characterize the effect of enjambment in verse is underscored by the fact that the OED recognizes the following sense of its nominal form: ‘Prosody. = ENJAMBMENT n.; a verse, couplet, or line displaying enjambement’, and furthermore cites this passage as its earliest recorded use: ‘1885 E. Gosse From Shakespere to Pope 6 Mr. Austin Dobson has proposed to me the term overflow for these verses in which the sense is not concluded at the end of one line or of one couplet, but straggles on, until it naturally closes; equivalent to the vers enjambé of the French’ (sense 5.a). The New Princeton Encyclopedia of Poetry and Poetics also uses that very word as the cornerstone for its definition of enjambment as ‘the overflow into the following poetic line of a syntactic phrase (with its intonational contour) begun in the preceding line without a major pause or juncture’ (359, s.v. ‘Enjambment). 136

Of these lines the New Princeton Encyclopedia of Poetry and Poetics, in its entry on ‘Enjambment’, pithily observes: ‘One had thought the series of parallel phrases complete at the end of line 25. One was wrong.’13 Less chary with his words, but ultimately making the same point, is John Hollander, who remarks that ‘the static pattern of line 25, framing the formula from the prayer-book (“As it was in the beginning, is now and ever shall be”), is jolted by the revelation that “be” was merely predicative (and of “worse,” at that), rather than existential.’14 I include both of these comments here, not merely because of the point they make, but also to demonstrate that this particular excerpt from Paradise Lost has taken on the status of a locus classicus in scholarly literature on enjambment and Milton’s use of it.15 The same passage had already been singled out for praise by Sheridan in his Art of Reading: ‘What an amazing force does this position give to the word worse! And in what strong colours does it paint to us the desperate state of reprobation into which Satan has fallen!’16 All of these appraisals are motivated by the fact that ‘be’, as we pass from line 25 into the next, transmutes before our eyes and within our ears from an intransitive verb to a copular verb by the realization that ‘Worse’ in line 26 must be its predicative. None of these comments, however, cares to make the point that our initial reading of ‘be’ as an intransitive verb is in no small part forced upon us by the proparoxytone ‘memory’ at the end of line 24. Due to the strictness of Milton’s verse design (the 23 lines leading up to the line in question all terminate in a lexically stressed syllable), the final syllable of ‘memory’ is likely to receive extra stress on account of its occupying the tenth syllable position, thereby becoming a perfect rhyme to the intransitive verb be. However, as soon as we realize that ‘be’ in this grammatical context is but a copular verb, we also realize that the lexical stress which we assigned to it in our initial reading was in fact grammatically unwarranted, in much the same way as the metrically induced stress on the last syllable of ‘memory’ was also unwarranted, albeit for different reasons. In point of fact, the reduced vowel of ‘be’ as a copular verb makes a perfect rhyme to the equally reduced vowel of the last syllable of ‘memory’. But this is a realization which only makes itself felt after the moment of syntactic anagnorisis has already passed, for which reason we must rely on our memory – the very word which led us astray in the first place – to correct our prosodic mistake. The suspicion that Milton carefully contrived this momentary lapse on the part of the reader by inserting end-rhyme into this

13 Preminger & Brogan (eds) (1993): 360, s.v. ‘Enjambment’. 14 Hollander (1975): 95. 15 E.g. James Longenbach, The Resistance to Poetry (Chicago & London: University of Chicago Press, 2004): 14 16 Sheridan (1775): 247-248. 137

particular patch of blank verse is confirmed by the fact that the two lines following immediately after make use of a perfect pair of end-rhymes, albeit obscured by orthography:

Worse; of worse deeds worse sufferings must ensue. Sometimes towards Eden which now in his view (PL 4, 26-27)

The ‘be / Worse’ transition from line 25 to 26 in Book 4 of Paradise Lost is a stellar example of enjambment, and in that sense its status as a locus classicus in discussions of this device is certainly merited. By its cunning interplay between lineation and syntax, between be as an intransitive verb and a as copular verb, the passage aptly illustrates Paul Valéry’s aphoristic definition of a poem as ‘a prolonged hesitation between sound and sense’.17 Giorgio Agamben, who cites Valéry’s aphorism, has called attention to the central function of enjambment in verse by stipulating for this device the honour of being the much sought after differentia specifica of poetry:

Awareness of the importance of the opposition between metrical segmentation and semantic segmentation has led some scholars to state the thesis (which I share) according to which the possibility of enjambment constitutes the only criterion for distinguishing poetry from prose. For what is enjambment, if not the opposition of a metrical limit to a syntactical limit, of a prosodic pause to a semantic pause? “Poetry” will then be the name given to the discourse in which this opposition is, at least virtually, possible; “prose” will be the name for the discourse in which this opposition cannot take place.18

If – as is most likely – Agamben means by ‘some scholars’ the Russian Formalists, we must first take issue with Agamben’s use of the term ‘poetry’ [‘poesia’] in this context: the proper antonym to prose is not poetry, but verse, and the Russian Formalists were well aware of this.19 Second, we must ask if Agamben in naming ‘the possibility of enjambment’ as the sine qua non of poetry is not guilty of conflating cause and effect. After all, the possibility of enjambment depends entirely on a

17 ‘[L]e poème, hésitation prolongée entre le son et le sens’, in Paul Valéry, Tel Quel II, ‘Rhumbs’, Pléiade II (Paris: Gallimard, 1960): 637. 18 Giorgio Agamben, The End of the Poem: Studies in Poetics, translated by Daniel Heller-Roazen (Stanford, CA: Stanford University Press, 1999): 109. 19 Agamben, judging from this excerpt, appears to be using the terms verse and poetry synonymously: ‘No definition of verse is perfectly satisfying unless it asserts an identity for poetry against prose through the possibility of enjambment’ (Agamben (1999): 39). 138

text’s being segmented into lines, and is, in other words, epiphenomenal to an even more basic formal criterion, namely that of lineation. We are thus on much safer ground by adopting M.L. Gasparov’s sober and concise criterion for drawing a meaningful distinction between our two most basic modes of verbal composition:

The main distinctive feature [between verse and prose] is … the fact that verse is segmented into equivalent and commensurable segments independent of syntax, and prose only in relation to syntax.20

It is exactly the fact that the segmentation of verse into lines is not principally governed by syntax which makes enjambment possible in verse, and not in prose. Enjambment is, in other words, to a greater extent than any other poetic device, characterized by its capacity to procure literary effects that are unattainable in prose. To see the truth of this claim, one need only compare Milton’s original lines from Book 4 with the corresponding passage in Dennis Danielson’s recent rendition of Paradise Lost into prose:

His conscience wakened long dormant despair, and with it woke bitter remembrance of what he had been, what now he was, and—even worse—what he must become (for worse crimes must bring worse punishments).21

This rendition effectively spoils that sublime moment of syntactic anagnorisis which Milton’s original passage so elegantly forces upon its reader. But it is, importantly, a stylistic shortcoming for which Danielson cannot be held responsible. Inasmuch as verse structure is a necessary precondition for enjambment, any translation of this particular passage from Paradise Lost into prose would be incapable of reproducing the extraordinary effect of Milton’s line-break between ‘be’ and ‘Worse’. But even as we celebrate the splendour of Milton’s use of enjambment at this particular juncture of Paradise Lost, we should not evade a serious issue raised by the same passage’s status as a scholarly exemplar of Miltonic enjambment. For while it is true that enjambment plays a crucial role in Milton’s versification, it must at the same time be admitted that the passage to which

20 Gasparov (1996): 97. 21 Dennis Danielson, Paradise Lost by John Milton, Parallel Prose Edition (Vancouver: Regent College Publishing, 2008): 147. 139

so much attention has been paid both here and elsewhere is hardly representative of Milton’s general use of enjambment in Paradise Lost. First of all, Milton rarely employs any form of the lemma be in line-end position: of the 10,565 lines that make up Paradise Lost only ten lines terminate in the form be; the form art terminates three lines; the forms am, is and been terminate two lines each; and was, are and were are only to be found at the end of a single line each. Furthermore, only a few of these instances employ enjambment, and those that do come nowhere near to achieving the spectacular syntactic artistry of the ‘must be / Worse’ line-break, as ought to be evident from these examples:

By Death at last, and miserable it is To be to others cause of misery, (PL 10, 981-982)

Gentle to me and affable hath been Thy condescension, (PL 8, 648-649)

These are both uncontroversial examples of enjambment, but they lack that anagnoristic quality which we experience in the break between lines 25 and 26 in Book 4. They also lack the idiomatic naturalness of word order which the sequence ‘what he was, what is, and what must be’ exploits; instead we find a warped, and decidedly Miltonic, syntax which all but drowns out the potential effects of enjambment. In this respect – and in spite of the general scarcity of lines terminating in forms of to be – the two excerpts quoted above are, in fact, much more typical of Milton’s general style of enjambment in Paradise Lost than is the celebrated passage from the opening of Book 4. This issue – that what is arguably the most widely celebrated instance of enjambment in Paradise Lost is not representative of the work in its entirety – is particularly problematic in a context such as ours, in which it is Milton’s general verse design that is under investigation. Given the pivotal role accorded to enjambment, or the possibility thereof, in modern treatments of poetics, it is somewhat curious that the critical nomenclature which we have at our disposal for the explication of this device is so limited. Whatever other problems poetics as a scholarly discipline may be struggling with, a dearth of technical terms is not usually one of them. But in the case of enjambment, no systematic attempt has yet been made to provide a critical 140

vocabulary that allows us to discriminate with accuracy between different types of this device. Hence the preliminary definition of enjambment that we gave in Chapter 3 and on which we have so far relied: enjambment occurs whenever a line break does not coincide with a syntactical juncture in the text, and therefore constitutes one extreme of a scale whose opposite extreme is constituted by an end-stopped line. Floating awkwardly somewhere in between these two extremes is the notion of the run-on line, but that’s about it, as far as critical terminology is concerned. This dearth of technical terms may well be the inevitable outcome of the fact that we have no commonly accepted Greek or Latin term for enjambment: as already stated elsewhere, enjambment is a French term whose existence we owe to Ronsard. The lack of a classical term to designate enjambment is all the more puzzling given that both Greek and Latin verse makes abundant use of the device; in fact, as Berkley Peabody points out, ‘In all the Indo-European compositional traditions that we know, enjambment of some kind is an established feature’.22 Enjambment, then, is not a novelty device that came into being in French sixteenth-century poetry and therefore needed a name; rather it is a feature so deeply embedded in Indo-European versification that no one before Ronsard felt the need to name it. A possible explanation for the lack of contemporary critical interest in enjambment as a poetic device in the classical languages may be sought in the typological profile of these languages as compared to that of most modern Indo-European languages. Both Greek and Latin are highly synthetic languages in which syntactic structures are established primarily by means of inflectional endings, for which reason word order plays only a marginal role. By contrast, the typological profiles of modern Indo-European languages such as English tend to be relatively analytical and to rely heavily on a fixed order of words to establish syntactic meaning. In synthetic languages an attributive adjective can thus be separated from the noun that it modifies by all sorts of grammatical inserts and still retain, through corresponding inflectional endings, its unequivocal binding to its head noun:

μῆνιν ἄειδε θεὰ Πηληϊάδεω Ἀχιλῆος οὐλομένην, ἣ μυρί’ Ἀχαιοῖς ἄλγε’ ἔθηκε,

In this familiar passage, the adjective ‘οὐλομένην’ (‘baneful’) modifies the noun ‘μῆνιν’ (‘wrath’) across several inserted phrases which do not form part of the noun phrase that the two words

22 Berkley Peabody, The Winged Word: A Study in the Technique of Oral Composition As Seen Principally through ’s Works and Days (New York: State University of New York Press, 1975): 142. 141

constitute in tandem. In analytical languages, on the other hand, an attributive adjective usually has to come immediately before or after the head noun it modifies:

Achilles sing, O Goddess! Peleus’ son; His wrath pernicious, who ten thousand woes

This is William Cowper’s translation of (parts of) the Iliad’s opening lines, in which the phrase ‘μῆνιν ... / οὐλομένην’ has been rendered as ‘wrath pernicious’, i.e. as a head noun followed immediately by its modifying adjective. In much the same way, synthetic languages can make do with inflectional endings to establish clause structure, whereas in analytical languages clause structure is established primarily by means of a fixed sequence of verb, subject and object. This latter principle, by which it is relative proximity in the syntagmatic concatenation of words that is responsible for establishing grammatical relationships, may well account for the fact that the dissociation of grammatically interlinked words by means of a line break is perceived as more intrusive, and therefore potentially more poetically significant, in analytical languages than in synthetic languages.23 The relationship between the inherent typological traits of a given language and the practice of versification is accorded a crucial significance in literary history by Ezra Pound in this central passage from the ABC of Reading:

The great break in European literary history is the change from inflected to uninflected language. ... The best way of using [an inflected] language is NOT the best way to use a language which has to be written in a certain order if it is to be clear.24

Having admonished his reader in this manner, Pound unsurprisingly picks out a passage from Paradise Lost to prove his point:

When Milton writes ‘Him who disobeys me disobeys’

23 A similar argument with regard to the effects of hyperbaton in synthetic and analytical languages is put forward by Stefan Daniel Keller in The Development of Shakespeare’s Rhetoric: A Study of Nine Plays (Tübingen: Francke Verlag, 2009): 52. 24 Pound (1951): 50. 142

he is, quite simply, doing wrong to his mother tongue. He meant Who disobeys him, disobeys me. It is perfectly easy to understand WHY he did it, but his reasons prove that Shakespeare and several dozen other men were better poets. Milton did it because he was chock a block with Latin.25

The charge against Milton is the usual: that his particular brand of what Matthew Arnold called the ‘grand style’ is achieved at the expense of both sense and syntax, by imitating the relatively lax word order of the classical languages for the sole purpose of euphony. Even Addison – perhaps Milton’s most influential apologist – confesses to finding Milton’s style ‘in some places too much stiffened and obscured by the frequent Use of those Methods, which Aristotle has prescribed for the raising of it’,26 and sentiments such as this were voiced with increasing fervour again at the beginning of the twentieth century, eventually giving rise to what has since been dubbed ‘The Milton Controversy’.27 In this controversy Pound and especially Eliot are rightfully counted among those who laid the ground for F.R. Leavis’s decidedly hostile attitude towards Milton, whereas critics such as William Empson, Cleanth Brooks and C.S. Lewis came forward as defenders of Milton’s style and work. Of these latter critics Lewis is of special significance to our present purpose in that he too has commented explicitly on the role of word order in Milton’s versification:

A fixed order of words is the price—an all but ruinous price—which English pays for being uninflected. The Miltonic constructions enable the poet to depart, in some degree, from this fixed order and thus to drop the ideas into his sentence in any order he chooses.28

Lewis’s point that Milton’s syntactic licenses frequently serve legitimate narratological and rhetorical functions is a welcome addendum to Eliot’s claim that Milton’s ‘syntax is determined by the musical significance, by the auditory imagination, rather than by the attempt to follow actual speech or thought’.29 But following actual speech or thought was, of course, never Milton’s ambition with Paradise Lost, as Lewis also points out with painstaking determination in his

25 Pound (1951): 51. 26 Joseph Addison, The Spectator, No. 285 (26 Jan. 1712). 27 See Christopher Ricks, Milton’s Grand Style (London: Oxford at the Clarendon Press, 1963). 28 C.S. Lewis, A Preface to Paradise Lost [1942] (London: Oxford University Press: 1960): 47. 29 T.S. Eliot, ‘Milton I’, in Eliot (1957), 138-145: 142. 143

Preface. In Lewis’s view, Eliot simply had a very hard time getting over the fact that Milton for his magnum opus chose to compose an epic poem instead of a dramatic poem. Regardless of what opinion one holds of the literary merits of Milton’s frequent departures from the normative word order of English, it remains the case that these selfsame departures form a crucial component in his distinct style of enjambment. But before we can venture an examination in detail of Milton’s style of enjambment, we must – in the absence of a commonly accepted taxonomy of this device’s various forms – first attempt to lay down some basic principles on which to proceed. So far, we have been able to get by with a negative definition of enjambment as a line break that does not coincide with a syntactical juncture in the text. However, as detractors of the term free verse have been keen to point out, negative definitions are difficult to work with in the long run, so perhaps we would fare better by stipulating that enjambment represents the breaking up of certain grammatical structures by means of a line break. One major type of grammatical structure that enjambment may affect is that of the phrase, by which a multi-word phrase – such as for instance ‘his wrath pernicious’ – is sliced by a line break in such a manner that the head of the phrase becomes separated from any of its modifiers. We shall in what follows refer to this type of enjambment as phrasal enjambment. The other major type of grammatical structure susceptible to enjambment is that of the clause, by which any of the basic syntactic constituents of the clause – verb, subject, object – becomes separated from any of the other syntactic constituents by means of a line break. This latter type of enjambment we shall refer to as clausal enjambment. To exemplify these two main types of enjambment, one need not look further than to the very opening lines of Paradise Lost:

Of mans first disobedience, and the fruit Of that forbidden tree, whose mortal taste Brought death into the world, (PL 1, 1-3)

The first line break is an instance of phrasal enjambment in that it separates the head of a noun phrase from its post-modifying prepositional phrase, whereas the second line break is an instance of clausal enjambment in that it separates a grammatical subject from its main verb and accompanying object. 144

The distinction between phrasal enjambment and clausal enjambment is not always quite as neat and clear-cut as it is implied above. Consider, for instance, these examples from the second Book of Paradise Lost:

with what eyes could we Stand in his presence humble, (PL 2, 239-240)

how in safety best we may Compose our present evils, (PL 2, 280-281)

In each of these examples it could be argued that the enjambment is simultaneously phrasal and clausal in that a verb phrase (‘could ... / Stand’; ‘may / Compose’) is being split across a line break, as a result of which process the grammatical subject becomes separated from the head of its verb phrase. However, it is important to note that in the second example, the line break is the only feature that severs the head of the verb phrase from its modifier, whereas in the first example the breaking up of the verb phrase is really effectuated by the insertion of ‘we’ between modifier and head verb, for which reason it is not the line break that does the disrupting of the phrase in this case. The first example, then, should be regarded as an instance of a predominantly clausal type of enjambment, whereas the enjambment in the second example is predominantly phrasal. The existence of fuzzy border cases between phrasal and clausal enjambment like the ones we have just discussed should not deter us from noting that the two types of enjambment operate on fundamentally different levels of grammatical structure. In phrasal enjambment it is, as the term suggests, the internal structure of a phrase that succumbs to the disruptive effect of a line break, whereas in clausal enjambment it is the syntactic hierarchy that such phrases establish between themselves which is being subverted. Both types of enjambment can, of course, be found in abundance, not only in Paradise Lost, but in virtually any English verse text of some length that does not preclude enjambment from its verse design. But the distinction is particularly useful in a survey such as this, in which we aim to pinpoint the characteristic style of Milton’s enjambment and the ways in which it differs from the kind of enjambment that can be found in dramatic blank verse. Using that very distinction as a point of departure, it may, in preliminary terms, be said that whereas 145

Milton’s use of clausal enjambment tends to be daring and innovative, his use of phrasal enjambment tends to be rather conservative. Such a sweeping statement is in obvious need of further qualification, and in what follows we shall attempt to provide this by comparing, from a formal point of view, the various ways in which Milton draws out his sense from one verse into another with the same phenomenon as it occurs in dramatic blank verse, especially that of Shakespeare. Clausal enjambment and phrasal enjambment will be examined separately, and we will begin with Milton’s use of clausal enjambment.

❦

I have already said that the celebrated anagnoristic clausal enjambment from Book 4 of Paradise Lost – ‘and what must be / Worse’ – strikes me as an atypical example of Milton’s general style of enjambment. There are two main reasons for this. The first reason has to do with the very anagnoristic quality which has made the passage famous: given the high frequency with which enjambment occurs in Paradise Lost as a whole, there are surprisingly few instances of line breaks that exploit syntactic ambiguities of the ‘be / Worse’ kind. Most of the time, we are in fact being made excruciatingly aware that Milton’s clauses are not complete when he employs clausal enjambment: throughout Paradise Lost clause-enjambed lines tend to flaunt their syntactic incompleteness. The other reason has to do with the kind of clausal pattern that is being disrupted by the line break in this particular instance of enjambment: the sequence ‘what must be / Worse’ represents the most common clause pattern of English declarative sentences: subject – verb – object (or, in this case, subject predicate). This is not a problem in and of itself, but the point is that clausal enjambments in which the normative clause pattern of English is disrupted by a line break can be found in almost any poem that allows for clausal enjambment in its verse design, for which reason there is nothing particularly Miltonic about it. The fixed position of syntactic constituents relative to one another – of the verb to its core arguments – within unmarked declarative sentences is not only what makes clausal enjambment such an effective poetic device in analytical languages like English; it is also an important parameter within the field of linguistic typology for classifying languages with a fixed word order . Using to denote subject, to denote object (or, for some verbs, depending on their valency, subject predicate), and to denote verb, linguistic typologists distinguish between six major groups of languages based on the different permutations that these constituents may be ordered in: 146

SVO, SOV, OSV, OVS, VSO and VOS. English, of course, strongly favours structures that embody the first category: in unmarked declarative sentences in English, the subject precedes the verb, which in turn is followed by the object. This normative word order – SVO – allows for two different kinds of clausal enjambment, depending on the exact point at which the structure is being disrupted by the line break: S / VO and SV / O. Both of these variants can, for reasons already stated, be found all over Paradise Lost; the following examples are but the earliest unequivocal occurrences of each variant:

S / VO whose mortal taste Brought death into the world, and all our woe, (PL 1, 2-3)

SV / O while it pursues Things unattempted yet in prose or rhyme. (PL, 1, 15-16)

Using the exact position of the line break in the SVO chain – i.e. whether the line break comes after the subject or after the verb – as a criterion for distinguishing between two discrete sub-types of this particular species of clausal enjambment may strike some readers as excessive, but the distinction is important nonetheless. In addition to allowing us to compare the relative frequencies with which each sub-type occurs across different works, genres or periods, it also serves as a reminder that the particular brand of anagnoristic quality which the ‘be / Worse’ enjambment embodies is only possible in the SV / O subtype. No comparable anagnoristic effect of syntax is made possible by the S / VO sub-type: such clausal enjambments are destined to flaunt their syntactic incompleteness. By keeping the initial position of the subject and reversing the relative position of verb and object, we arrive at another set of possible clausal enjambments, S / OV and SO / V:

S / OV Devil with devil damned Firm concord holds, (PL 2, 496-497)

147

SO / V Thus they their doubtful consultations dark Ended rejoicing in their matchless chief: (PL 2, 486-487)

The fact that I have been unable to procure in Paradise Lost an example of the S / OV subtype in which the break between subject and object is unmitigated by the insertion of an added phrase – in this case the prepositional phrase ‘with devil damned’ – is an indication of the scarcity with which this particular sub-type of clausal enjambment occurs, even in Milton.30 The SO / V subtype, on the other hand, rears its head several times throughout Paradise Lost, although obviously not nearly as frequently as do those sub-types of clausal enjambment which exploit the normative SVO pattern. Even so, it could well be argued that the two examples of SOV enjambment cited above are considerably more emblematical of Miltonic enjambment than any instance of SVO enjambment could ever be. The same holds true – perhaps to an even greater extent – for clausal enjambments that disrupt syntactic patterns in which the object is allowed to occur in clause-initial position. Clauses in which this is the case make possible four new subtypes of clausal enjambment that must be added to those already covered above: O / VS, OV / S, O / SV and OS / V. Paradise Lost provides us with unambiguous instances of each subtype:

O / VS such privilege Hath Omnipresence (PL 7, 589-590)

OV / S that proud honour claimed Azazel as his right, (PL 1, 533-534)

30 The example quoted by Pound to illustrate Milton’s Latinate syntax –

him who disobeys Me disobeys, (PL 5, 611-612)

– is, of course, another instance of S / OV enjambment, but one that is made considerably more complex by the fact that the S – ‘him who disobeys’ – is itself a clause with OSV structure. 148

O / SV So spake the fervent Angel, but his zeal None seconded, (PL 5, 849-850)

OS / V Him the Almighty Power Hurled headlong flaming from th’ ethereal sky (PL 1, 44-45)

Verb-initial clause patterns are rarely used for declarative purposes, even by Milton. The VSO pattern thus most commonly occurs either in imperative sentences such as

VS / O (imp) be thou in Adam’s room The head of all mankind, though Adam’s son. (PL 3, 285-286)

– or in interrogative sentences such as

VS / O (interr) remember’st thou Thy making, while the Maker gave thee being? (PL 5, 857-858)

But it is certainly possible to find examples of verb-initial clause patterns in declarative sentences, though usually only when an intransitive verb is preceded by an initial adverbial, e.g.

V / S Seasons return, but not to me returns Day, or the sweet approach of ev’n or morn, (PL 3, 41-42)

The VOS pattern occurs even less frequently in Paradise Lost than does the VSO pattern, although it does occur at least once across a line break:

149

VO / S About him exercised heroic games Th’ unarmèd youth of Heav’n; (PL 4, 551-552)

Milton’s obvious delight in playing around with the syntactic constituents of his sentences throughout Paradise Lost might in itself be described as a heroic game. His use of line breaks to split up clause patterns that differ radically from the normative syntactic pattern of English creates a type of clausal enjambment which is, if not exclusively, then at least predominantly epic. Against the sensible riposte that the above collection of Miltonic enjambments is merely an inevitable by-product of Milton’s licentious syntax, it must be insisted that freedom in word order does not necessarily entail the types of ‘epic’ enjambment that we find so prevalently throughout Paradise Lost. It is, of course, true that departures from the normative word order of English occur extremely frequently in Paradise Lost, but other poets have made copious use of inverted clause patterns without ever allowing themselves such liberties across line breaks as those we see in Milton. In an important study, John Porter Houston has catalogued the many ways in which Shakespeare departs from the normative word order of English,31 and based on a comparison of nine different plays by Shakespeare Stefan Daniel Keller has singled out as the play in which ‘Shakespeare used inverted or unusual syntax most often’.32 However, a closer examination of Hamlet reveals that only very rarely are the many instances of inverted clause patterns in that play allowed to extend themselves across a line break. Most of the time aberrant clause patterns in Hamlet are, in fact, confined within the space of a single line, as in this extract where Laertes has just learned of his sister’s death by drowning:

Too much of water hast thou, poor Ophelia, And therefore I forbid my tears. But yet It is our trick; nature her custom holds, (4, 7, 185-187)33

31 See John Porter Houston, Shakespearean Sentences: A Study in Style and Syntax (Baton Rouge: Louisiana State University Press: 1988). 32 Keller (2009): 53. 33 All references to Shakespeare are to The Arden Shakespeare: Complete Works, Revised Edition, edited by Richard Proudfoot, Ann Thompson & David Scott Kastan (London: Arden/Methuen, 1998). 150

The first part of the first line contains a clause with OVS structure, and the last part of the last line contains a clause with SOV structure. Shakespeare, in other words, was not exactly impervious to the allurement of inversions, although one could easily get that impression from reading Pound’s comments on Milton and Shakespeare in the ABC. The SOV pattern in particular occurs quite frequently throughout Hamlet,34 but only in a single case is this clause pattern allowed to overflow from one line into the next:

The violence of either grief or joy Their own enactures with themselves destroy. (3, 2, 198-199)

This couplet occurs – as indicated both by the use of end-rhyme and the use of italics – in that part of the play which is dedicated to the performance of Hamlet’s recasting of The Murder of Gonzago, for which reason it is ill suited as an example of Shakespeare’s general style of clausal enjambment. Outside of The Mousetrap section of Hamlet, the only type of aberrant clausal enjambment that occurs is the one in which an initial object is separated by a line break from its accompanying subject and verb, i.e. O / SV.35 This particular subtype of clausal enjambment can be found in a handful of instances, but when it does occur it is usually softened by the insertion of an adverbial between object and subject. However, the vast majority of inverted clause patterns in Hamlet are carefully positioned in such a manner that clausal enjambment is deliberately avoided, in stark contrast to the strategy employed by Milton in Paradise Lost. Before moving on to our discussion of phrasal enjambment, we must first consider that special subset of clausal enjambment by which a line break is allowed to come between a clause and the coordinator or subordinator which introduces it. This is a type of clausal enjambment which is all but absent in Paradise Lost, even though it occurs regularly in dramatic blank verse. By far the most common coordinators in English are and, but, and or; extremely frequent words in any linguistic register, and yet they are not allowed to occur in the tenth syllable position of a single line in the entirety of Paradise Lost. Milton, in other words, rejects this particular kind of enjambment,

34 E.g. Hamlet 1,1, 59; 3,2, 86; 3,3,3; 3, 4, 145. 35 There are also in Hamlet a few occurrences of the V / S pattern in combination with a preceding adverbial, e.g. ‘on his choice depends / The sanity and health of this whole state’ (1, 3, 20-21). However, this particular clause pattern is quite common in English and should therefore not be counted as a truly aberrant clause type. 151

in spite of the fact that it is pervasive in Shakespeare. Here is an example of each of the most common coordinators occurring before a line break, all collected from The Winter’s Tale:

Still neighbour mine. My ships are ready, and My people did expect my hence departure (1, 2, 449-450)

There is a sickness Which puts some of us in distemper, but I cannot name the disease, (1, 2, 384-386)

Be thereat glean’d: for all the sun sees, or The close earth wombs, or the profound seas hides (4, 4, 491-492)

In each of these examples, the coordinating conjunction (marked in bold) occupies the tenth syllable position, even though conjunctions generally receive a minimum of stress in delivery. In the two first examples, this does not constitute a metrical problem, inasmuch as the coordinator in both cases follows a paroxytone, but in the last example the coordinator or follows the lexical monosyllable ‘sees’, resulting in a severely weakened ending to that line. From a purely metrical point of view, such a line would, of course, be completely unacceptable to Milton, but the total absence of line-terminal coordinators in Paradise Lost indicates that the same holds true of any instance of enjambment that separates a coordinator from the clause it introduces. Whereas coordinators link clauses that are grammatically equipollent to one another, subordinators link clauses in which one clause is embedded as part of another clause. That is to say that a clause introduced by a subordinator is always a dependent clause, most frequently an adverbial clause that adds details of time, place, contingency or the like to the main clause. Only extremely rarely are line breaks allowed to sever a subordinator from the dependent clause it introduces in Paradise Lost. Shakespeare, on the other hand, makes frequent use of that kind of enjambment; here are but three randomly picked examples from Cymbeline:

152

At three or two years old, I stole these babes, Thinking to bar thee of succession as Thou refts me of my lands. (3, 3, 101-103)

The gods throw stones of sulphur on me, if That box I gave you was not thought by me A precious thing: (5, 5, 240-242)

I heard no letter from my master since I wrote him Imogen was slain. (4, 3, 36-37)

These three subordinators introduce adverbial clauses of comparison, condition and temporality respectively, and in each case the subordinator is left to dangle in the tenth syllable position, while the clause whose presence the coordinator signals does not commence until the following line begins. This type of enjambment is generally barred from occurring in Paradise Lost: there is not a single line in that work which terminates in as, if or since (or other common subordinators such as while/whilst, till/until or though for that matter). The only words capable of functioning as a subordinator that are allowed to occur in line-final position in Paradise Lost are when and where: three lines terminate in when, and two in where;36 in none of these instances, however, does the word in question function as a true subordinating conjunction. It is no coincidence that all of the examples given above of enjambment between co- /subordinator and clause have been taken from relatively late works by Shakespeare. In his early works, Shakespeare – like Milton – tended to avoid enjambments by which a weakly stressed function word would be left to carry the fifth beat of the pentameter without the aid of phrasal or clausal closure. This remarkable feature of Shakespeare’s mature style was first given due consideration in the latter part of the nineteenth century, especially after F.J Furnivall had founded the New Shakspere Society in 1873. Inspired by recent advances in Chaucer scholarship, the members of this loosely structured society set out to determine the chronology of Shakespeare’s

36 When occurs line-terminally in PL 5, 546; 6, 219 and 12, 384; where occurs line-terminally in PL 4, 451 and 5, 340. 153

plays by subjecting them to a variety of metrical tests and statistical investigations: an approach that met with some success and even more controversy. The Society’s positivist-tinted choice of methodology thus sparked a prolonged and undignified feud between Furnivall and Algernon Charles Swinburne, the latter of whom accused the Society’s members of being, among other and less flattering things, ‘scholiasts’ and ‘metre-mongers’.37 Of the numerous tests that these scholiasts and metre-mongers devised to determine the most likely chronology of Shakespeare’s plays, A.C. Bradley, with the benefit of hindsight, singled out three as being of permanent value:

The really useful tests, in regard to plays which admittedly are not widely separated, are three which concern the endings of speeches and lines. It is practically certain that Shakespeare made his verse progressively less formal, by making the speeches end more and more often within a line and not at the close of it; by making the sense overflow more and more often from one line into another; and, at last, by sometimes placing at the end of a line a word on which scarcely any stress can be laid. The corresponding tests may be called the Speech-ending test, the Overflow test, and the Light and Weak Ending test.38

The detailed study of lines ending in such words ‘on which scarcely any stress can be laid’ had been undertaken most meticulously by the polymath John Kells Ingram, who reported his findings in the New Shakspere Society’s first Transactions, published in 1874. Drawing on the earlier work of the Scottish critic George Lillie Craik, Ingram adroitly reasserted Craik’s distinction between two kinds of ‘weak endings’ in Shakespeare:

It is evident that amongst what have been called as a class weak endings, there are different degrees of weakness. Broadly, as Craik has already observed, there are two such degrees, which require to be discriminated, because on the words which belong to one of these groups the voice can to a certain small extent dwell, whilst the others are so essentially proclitic in their character ... that we are forced to run them, in pronunciation no less than in

37 See Oscar Maurer, ‘Swinburne vs. Furnivall: A Case Study in “Aesthetic” vs. “Scientific” Criticism’, The University of Texas Studies in English, Vol. 31 (1952), 86-96, in which the whole controversy between Swinburne and the New Shakspere Society is outlined. 38 A.C. Bradley, Shakespearean Tragedy, Second Edition (London: MacMillan and Co., 1919): 473. In 1930, however, E.K. Chambers asserted that ‘In view of all the uncertainties attaching to the metrical tests, I do not believe that any one of them or any combination of them can be taken as authoritative in determining the succession of plays which come near to each other in date’; in E.K. Chambers, William Shakespeare: A Study of Facts and Problems, Vol. I (London: Oxford University Press, 1930): 269. 154

sense, into the closest connection with the opening words of the succeeding line. The former may with convenience be called “light endings,” whilst to the latter may be appropriated the name (hitherto vaguely given to both groups jointly) of “weak endings.” 39

Based on this distinction, Ingram was able to show that the relative frequency of each category of weak endings in Shakespeare’s plays varied considerably, and he furthermore posited that a high frequency of such endings in any given play was an indication of that play’s lateness in Shakespeare’s production. The tentative chronology Ingram established based on this simple parameter comes very close to what is commonly accepted today, and it is therefore in Shakespeare’s later plays that we can find the most interesting contrast to Milton’s style of enjambment.40

❦

The fact that Shakespeare in his mature works allows for both light and weak endings, whereas Milton generally shuns both, also accounts for the majority of differences between the two poets’ use of phrasal enjambment. This is especially evident in that type of phrasal enjambment which might be termed prepositional enjambment, i.e. instances of line breaks that split up a prepositional phrase in such a manner that the preposition becomes severed from its complement.41 This type of enjambment occurs in both Milton and Shakespeare, but with an important difference. Whereas Shakespeare, at least in his later plays, permits enjambment after any given preposition, Milton is far more eclectic and uses only certain prepositions in line-terminal position. Milton’s criterion for distinguishing between prepositions that are permitted to partake in prepositional enjambment and prepositions that are not is, importantly, not semantic, but purely phonological. The only prepositions that Milton employs in his prepositional enjambments thus consist of more than one syllable and furthermore have an oxytonic stress profile. That is to say that while one can encounter instances of prepositional enjambment such as this in Paradise Lost –

39 John K. Ingram, ‘On the “Weak Endings” of Shakspere With Some Account of the History of the Verse-Tests in General’, in The New Shakspere Society’s Transactions (London: Trubner & Co., 1874), 442-464: 447. 40 When I refer to Shakespeare’s later plays, the ones I include are Timon of Athens, , Antony and Cleopatra, Pericles, Coriolanus, The Winter’s Tale, Cymbeline, and Henry VIII. I exclude, however, All’s Well that Ends Well on account of the fact that it shows relatively few of the metrical peculiarities which the other plays exhibit in abundance. 41 Instances in which a line break occurs within a phrase that functions as the complement of a preposition (e.g. ‘by work / Divine (PL, 5, 255-256)) are not counted as prepositional enjambments here, but rather as phrasal enjambments at the level of the phrase which is being severed by the line break. 155

But of the Fruit of this fair tree amidst The garden, God hath said, Ye shall not eat (PL 9, 661-662)

– one looks in vain for instances such as this, from Shakespeare, in which a monosyllabic preposition is allowed to take up the tenth syllable position:

Our Jovial star reign’d at his birth, and in Our temple he was married. (Cymbeline 5, 4, 105-106)

In total there are some twenty instances of prepositional enjambment in Paradise Lost, and the prepositions that I have noticed in those instances are – in addition to amidst – among, beneath, between and beyond.42 Given these instances of prepositional enjambments in Paradise Lost, it seems fair to conclude that while Shakespeare and Milton are in agreement about the appropriateness of this device in blank verse (both poets make use of it), they are in apparent disagreement about the metrical conditions of it. Whereas Milton’s verse design requires a syllable that receives lexical stress in the tenth syllable position before allowing prepositional enjambment, Shakespeare’s design can make do with less. This crucial difference in terms of verse design is also of pivotal importance when it comes to phrasal enjambments at the level of the verb phrase, i.e. enjambments in which the head of a verb phrase is separated from any of it modifiers by means of a line break. Unlike synthetic languages, such as Latin, in which a great many grammatical distinctions can be signalled by means of morphological changes within the main verb itself, English relies on a small, finite inventory of auxiliary verbs to signal features such as aspect, voice and modality. This makes possible in English a set of phrasal enjambments that would be close to impossible to procure in Latin, namely those in which a line break comes between an auxiliary verb and a head verb. By far the most common auxiliary verbs in English are be and have; their primary functions as auxiliaries are to partake in the formation of the perfect and the progressive aspect, and – in the case of be – also the passive voice.

42 For examples of each see, for instance, PL 10, 100-101; 6, 342-343; 10, 497-498; 1, 587-588. 156

The perfect aspect is formed in English by pre-modifying the past participle of a head verb with a form of have, and this distinctly Germanic construct supplies an English poet with the possibility of a rather poignant type of phrasal enjambment such as this one from The Tempest:

Work not so hard. I would the lightning had Burnt up those logs that you are enjoined to pile! (3, 1, 16-17)

By my count there are some forty instances of this type of enjambment in the entirety of Shakespeare. The majority of them occur, unsurprisingly, in the later plays, but solitary instances can also be found in several earlier plays, e.g. Richard III 4, 4, 395-396; The Merry Wives of Windsor 4, 6, 32-33 and Measure for Measure 2, 2, 11-12. In Paradise Lost, on the other hand, there is not a single instance to be found of this particular type of enjambment, by which a verb phrase marked for the perfect aspect is sliced by a line break immediately after the auxiliary verb. The same is true of enjambments by which a verb phrase marked for the progressive aspect is severed by a line break in such a way that the auxiliary becomes separated from the main verb. The progressive aspect in English is, of course, formed by pre-modifying the present participle of a head verb with a form of be: such verb phrases are never enjambed in Paradise Lost. But more interestingly, Shakespeare also avoids this particular subtype of phrasal enjambment, in spite of the fact that he allows of enjambments within verb phrases marked for the perfect aspect. The absence of enjambments within progressive verb phrases in Shakespeare is, to some extent, an inevitable consequence of the relative scarcity with which the progressive aspect is generally employed by Shakespeare. The progressive aspect is used far more commonly in Present Day English than it was in Early Modern English,43 but to the extent that it does occur in Shakespeare it does so most frequently in the later plays, with ‘a dozen or so examples’ occurring in Henry VIII.44 The other type of verb phrase in which auxiliary be may become severed from its main verb by means of enjambment is the passive verb phrase, in which construction the past participle of a main verb is pre-modified with a form of be. Given that Milton only rarely allows any form of be to occur in line-final position, it is hardly surprising to find that in the entirety of Paradise Lost there

43 See Marinel Gerritsen, ‘Divergent Word Order Developments in Germanic Languages’, in Jacek Fisiak (ed.), Historical Syntax, Trends in Linguistics: Studies and Monographs 23 (Berlin: Walter de Gruyter & Co., 1984), 107- 135: 126. 44 Elizabeth Traugott, The History of English Syntax (New York: Holt, Rinehart and Winston, 1972): 143. 157

is but a single occurrence of enjambment within a verb phrase marked for the passive voice; and that instance is even mitigated by the insertion of a prepositional phrase between auxiliary and main verb:

To cattle and each beast; which would not be To them made common and divulged, (PL 8, 582-583)

Unmitigated instances of enjambment between auxiliary and main verb in passive verb phrases can, however, be found throughout Shakespeare (e.g. The Merchant of Venice, 1, 3, 146-147; Coriolanus, 4, 5, 80-81; The Winter’s Tale, 1, 2, 239-240), although not very frequently. A similar picture emerges if one surveys the distribution of enjambments in which a line break severs a verb phrase marked for modality in such a manner that the modal auxiliary becomes separated from the main verb it modifies. This happens extremely rarely in Paradise Lost; three times to be exact. Two of these instances employ the modal verb might –

That with reiterated crimes he might Heap on himself damnation, (PL 1, 214-215)

The sooner for their arms; unarmed they might Have easily as Spirits evaded swift (PL 6, 595-596)

– and the third instance (PL 2, 280-281) employs may; that passage has already been quoted as an example of a fuzzy border case between clausal and phrasal enjambment. In Shakespeare, on the other hand, one can easily find instances of phrasal enjambments with all of the most common English modal verbs: shall, should, will, would, may, might and must.45

45 E.g. ‘Make good this ostentation, and you shall / Divide in all with us.’ (Coriolanus 1, 6, 86-87) ‘I pray thee, mark me,–that a brother should / Be so perfidious!’ (The Tempest 1, 2, 67-68) ‘I doubt not but this populous city will / Yield many scholars.’ (Pericles 4, 6, 184-185) ‘And, to be more than what you were, you would / Be so much more the man.’ (Macbeth,1, 7, 50-51) ‘That on the touching of her lips I may / Melt and no more be seen.’ (Pericles 5, 3, 42-43) 158

The frugality with which Milton employs line breaks to split up an auxiliary verb and the main verb it modifies is, I believe, to some extent the natural result of his being ‘chock a block with Latin’, to borrow Pound’s memorable phrase. As already hinted above, Latin requires but a single verb to express many of the grammatical distinctions in the verb phrase which in English can only be expressed by means of auxiliaries. St. Jerome’s laconic report quoted in Chapter 1 – Capitur Urbs quae totum cepit orbem – thus juxtaposes two different forms of the verb capio, both of which require auxiliary aid in English: capitur (present passive) would be ‘is captured’; cepit (perfect active) would be ‘has captured’. To a trained classicist like Milton, such translations are presumably barbaric enough in and by themselves, and are therefore in no immediate need of a line break to underscore their Germanic crudeness. But in addition to the lack of a classical precedent for splitting up such verb phrases across a line break, Milton’s general habit of avoiding this feature is also metrically motivated. Auxiliary verbs in English are almost exclusively monosyllabic, and since their function is to pre-modify a main verb, they generally receive a minimum of stress in delivery (in anticipation of the main verb), for which reason their vowel nucleus is likely to undergo weakening. The majority of those auxiliary verbs which we have examined here thus belong to that exclusive category of English words which come with two distinct pronunciations depending on their grammatical function and clausal position. Only in sentence-final position, or for the purpose of contrast or metalinguistic emphasis, are such words pronounced with a full vowel; in all other cases their vowel is reduced. There are just some forty of these ‘weak form words’ in English, but all of them are extremely common, especially in their weakened form.46 The only common English auxiliary verbs that do not belong to this class are may and might; on account of their vowel nucleus being a diphthong these words are relatively impervious to weakening. Incidentally – or not – this is the selfsame pair of auxiliaries that Milton allows to partake in enjambment of verb phrases marked for modality in Paradise Lost. This, again, seems to indicate that Milton’s use of enjambment is delimited primarily by the phonological properties of the syllable that occupies the tenth syllable position.

‘Not he: but yet heaven’s bounty towards him might / Be us’d more thankfully. ‘(Cymbeline 1, 7, 78-79) ‘Incite them to quick motion, for I must / Bestow upon the eyes of this young couple’ (The Tempest 4, 1, 39-40) 46 See Peter Roach, English Phonetics and Phonology: A Practical Course, 4th edition (Cambridge: Cambridge University Press, 2009): 89-96. 159

The final major category of phrasal enjambment that we will consider in this chapter is enjambment at the level of the noun phrase.47 We have already defined phrasal enjambment as a line break that severs the head of a phrase from any of its modifiers, and in the case of the noun phrase this means that we must distinguish between two discrete subcategories of noun phrase enjambment: those in which a pre-modifier is separated from its head noun, and those in which a head noun is separated from its post-modifier. Since adjectives can be used as both pre-modifiers and post-modifiers in English, we may use the following examples from Paradise Lost to illustrate the difference between the two subcategories of noun phrase enjambment:

God their Creator, and th’ invisible Glory of him that made them to transform (PL 1, 369-370)

War hath determined us, and foiled with loss Irreparable; (PL 2, 330-331)

There is a pronounced difference in the effect that these two examples produce. In the first example, in which a pre-modifier is left to dangle at the end of a line, we are intensely aware of the line’s incompleteness; whereas in the second example, in which it is the head of the noun phrase that dangles in line-terminal position, we are allowed some sense of closure. This sense of closure is, however, abruptly suspended as soon as we commence our reading of the following line, and the effect this has on the reading process is an integral part of Paradise Lost. We must therefore begin by outlining a few of the most common sub-types of enjambment between head noun and post- modifier in Paradise Lost.

47 By limiting our investigation to enjambment within prepositional phrases, verb phrases and noun phrases, we ignore enjambment within adverbial phrases and adjectival phrases. The most unfortunate consequence of this is that the following instance of adjectival enjambment in Paradise Lost is demoted to this footnote:

the serpent, whom they called Ophion with Eurynome, the wide- Encroaching Eve perhaps, had first the rule (PL 10, 580-582)

This type of enjambment, in which a word is split across a line break by means of a hyphen, occurs only this one time in Paradise Lost, and never in Shakespeare. More extreme examples of the same phenomenon can, however, be found in both Donne and Jonson, as well as in Hopkins. 160

By far the most common type of enjambment in Paradise Lost is that by which the head of a noun phrase is separated from a post-modifying of-phrase by means of a line break (‘the fruit / Of that forbidden tree’). So common is this feature that – based on my count for the first six books of Paradise Lost – more than 4 per cent of all lines in that work’s entirety are enjambed in this particular manner. It is, admittedly, a rather bland type of enjambment, unobtrusive to the point of not being experienced as true enjambment. A far more intrusive – but typologically different – variant of enjambment within a noun phrase post-modified by an of-phrase occurs in Henry VIII:

Our cause, that she should lie i’th’ bosom of Our hard-ruled King. (3, 2, 100-101)

This is as much an example of the late Shakespeare’s antics as of his metrics. Strictly speaking, it is of course an instance of prepositional enjambment, but one in which the enjambed prepositional phrase is embedded within a noun phrase (which in turn functions as the complement of yet another prepositional phrase). As we have already established that Milton never enjambs after a monosyllabic preposition, we do not need to look for comparable instances in Paradise Lost: we know that they are not there. But, more surprisingly, instances of the unobtrusive kind of enjambment between a head noun and an of-phrase (‘the fruit / Of that forbidden tree’), which are so prevalent in Milton, are quite scarce in Shakespeare. This is particularly salient in his earlier works: in The Two Gentlemen of Verona there are as few instances of this kind of enjambment as there are Gentlemen in the title, and in I have counted a mere seven instances, in spite of the fact that that play is composed entirely in verse. The ratio goes up in Shakespeare’s later plays, but in no play that I have examined for this feature are there more than 25 instances. In Paradise Lost, on the other hand, there are some 450 in total. In addition to adjectives and prepositional phrases serving as post-modifiers, head nouns may also be post-modified in English by various kinds of clauses. It is useful to distinguish between non-finite post-modifying clauses such as here –

Could merit more than that small infantry Warred on by Cranes: (PL, 1, 575-576) 161

– and finite post-modifying clauses, i.e. relative clauses, such as here:

Deserve the precious bane. And here let those Who boast in mortal things, and wond’ring tell (PL 1, 692-693)

In cases like these it hardly makes sense to speak of enjambment: as much as splitting up a complex noun phrase, the line break can also be said to signal the beginning of a new clause. However, in Shakespeare one can find instances in which complex noun phrases with post-modifying relative clauses are being subjected to a rather poignant type of enjambment, as in these instances from Cymbeline:

Fidele, sir: I have a kinsman who Is bound for Italy; (4, 7, 33-34)

To th’oath of loyalty; this object, which Takes prisoner the wild motion of mine eye, (1, 7, 102-103)

Should yield the world this ass! a woman that Bears all down with her brain, (2, 1, 52-53)

Strictly speaking, this kind of enjambment – by which a line break is allowed to separate a relative pronoun from the clause it not only introduces, but in which it also functions as the grammatical subject – is clausal rather than phrasal. It is in some respects reminiscent of the kinds of enjambment that occur between co-/subordinators and clauses, and like those it never occurs in Paradise Lost. Moving on to phrasal enjambment between pre-modifiers and head nouns, it is worth noticing that this is the one type of enjambment in which there are no systematic differences 162

between Milton’s and Shakespeare’s strategies. Both poets, on occasion, make use of enjambments between an attributive adjective and the head noun that it pre-modifies (such as in PL, 1, 369-370 (cited above)), although not very frequently. Even if we expand our definition of noun-phrasal enjambment to include enjambments between determiners and head nouns also, we still find more similarities than differences. Determiners are customarily divided into four different classes: articles, possessive determiners, demonstrative determiners and quantifying determiners. Milton and Shakespeare both allow enjambment after quantifying determiners, but reject it after the other three classes of determiners. The most common quantifying determiner to partake in enjambment, in both Milton and Shakespeare, is all:

Nor I on my part single, in mee all Posterity stands cursed: (PL 10, 817-818)

But other quantifying determiners and semi-determiners, like each and such, can also be found dangling at the end of a line. However, what cannot be found, either in Shakespeare’s dramatic works or in Paradise Lost, are the demonstrative determiners this/that and these/those in that same position. Apparently both poets must have felt that the structural unity between this kind of determiner and its head noun was too powerful to be challenged by a line break. This is particularly remarkable in the case of the plural demonstrative determiners these/those since these determiners have as their nucleus a long vowel and a diphthong respectively: for this reason they are unlikely to undergo weakening and are therefore, in principle, fit to occupy the tenth syllable position even before enjambment. It is thus hardly a coincidence that the earliest instance of enjambment between a demonstrative determiner and a head noun of which I am aware employs a plural, rather than a singular, demonstrative determiner. It occurs in Henry Vaughan’s ‘The Water-fall’ (1650):

My first consigner unto those Fountains of life where the Lamb goes?48

48 Cited from Henry Vaughan, Poetry and Selected Prose, edited by L.C. Martin (London: Oxford University Press, 1963). 163

The earliest instance of enjambment between any of the two singular demonstrative determiners and a head noun that I am familiar with occurs in John Philips’s Cyder (1708):

Tempting, not fatal, as the Birth of that Primæval interdicted Plant, that won Fond Eve in hapless hour to taste, and die.

Cyder is one of several mock imitations of Milton’s epic style to be composed during the early eighteenth century, as Philips himself boasts at the very opening of the poem: ‘Thy Gift, Pomona, in Miltonian Verse / Adventrous I presume to sing’.49 Too adventrous, perhaps, since Milton would never have presumed in his epic to this kind of enjambment. Possessive determiners are also barred from occurring in line-terminal position in both Shakespeare and Milton, although rare instances can be found in Jacobean dramatic blank verse, e.g.

Made only worthy in his love, and her Thankfull acceptance, (Thierry and Theodoret 4, 2, 127-128)50

Even so, there is a single occurrence in Paradise Lost of enjambment between a genitive noun used as a possessive determiner and the head noun it specifies:

When to enshrine his relics in the sun’s Bright temple, to Egyptian Thebes he flies. (PL 5, 273-274)

In this case it would seem that ‘sun’s’ to Milton’s exacting ear is sufficiently prominent, phonologically speaking, to partake in enjambment within a possessive noun phrase, but in the entirety of Shakespeare I have not been able to locate a single example of the same phenomenon.

49 Cited from John Philips, Poems Attempted in the Style of Milton (London: 1762). 50 Cited from Fredson Bowers (ed.), The Dramatic Works in the Beaumont and Fletcher Canon, Vol. III (Cambridge: Cambridge University Press, 1976). 164

Again, the structural binding between the possessive determiner and its head noun appears to have been deemed so strong that a line break was out of the question. When we come to the final major class of determiner – the articles the and a/an – we have moved into an area in which the unity of determiner and head noun is so pronounced, both phonologically and conceptually, that it all but resists enjambment in metrical verse. Neither Shakespeare nor Milton enjambs after an article, and to find an example of a poet who does we must look to William Blake, whose ‘To the Evening Star’ (1783) is a strong candidate for being the earliest poem in English to separate an article from its head noun by means of a line break:

Smile on our loves; and, while thou drawest the Blue curtains of the sky, scatter thy silver dew51

The effect of this enjambment is quite extraordinary, and it is made possible only by drawing out ‘drawest’ over two syllables, the second of which is in fact sufficiently weak (we must admit, to our own consternation) to go into iambic partnership with ‘the’. That it is Blake who first tries out this extreme type of enjambment in English is not surprising: in many respects Blake’s work is positioned firmly outside of the mainstream Tradition. Thus, in addition to providing us with the earliest instance of enjambment between an article and its noun, ‘To the Evening Star’ can also lay claim to being the first unrhymed sonnet in English; the collection of which it forms part furthermore includes the earliest poems in English to make use of unrhymed pentameter stanzas, e.g. ‘To Spring’ (quatrains) and ‘To Autumn’ (sestets). This disregard for formal conventions has also earned Blake a prominent position in the history of free verse; Alicia Ostriker, for instance, laconically characterizes ‘The Argument’ of The Marriage of Heaven and Hell as ‘The first piece of free verse in English’ in her edition of Blake.52 Without further qualification, however, this claim is dubious in that it disregards a number of earlier poems with an equal right to that distinction, the most important of which will be discussed briefly in the next chapter. Blake’s significance as a harbinger of free verse in English, in other words, does not rest principally with his versification in any single poem, but rather with his prose address ‘To the Public’ at the beginning of Jerusalem The Emanation of The Giant Albion:

51 Cited from William Blake, The Complete poetry and Prose, Newly revised edition, edited by David V. Erdman (New York: Doubleday, 1988). 52 William Blake, The Complete Poems, edited by Alicia Ostriker (London: Penguin Books, 1977): 898. 165

When this Verse was first dictated to me I consider’d a Monotonous Cadence like that used by Milton & Shakspeare & all writers of English Blank Verse, derived from the modern bondage of Rhyming; to be a necessary and indispensible part of Verse. But I soon found that in the mouth of a true Orator such monotony was not only awkward, but as much a bondage as rhyme itself. I therefore have produced a variety in every line, both of cadences & number of syllables. Every word and every letter is studied and put into its fit place: the terrific numbers are reserved for the terrific parts—the mild & gentle, for the mild & gentle parts, and the prosaic, for inferior parts: all are necessary to each other. Poetry Fetter’d, Fetters the Human Race!53

This remarkable manifesto quite unashamedly echoes Milton’s note on ‘The Verse’ in the 1674 edition of Paradise Lost, and the claim it makes for the significance of Jerusalem is therefore considerable. But what is particularly noteworthy in this context is that Blake – unlike Milton – cites neither ancient nor modern precedents to justify his principles of versification. In fact, it is made clear that Blake’s versification in Jerusalem marks a conscious rejection of blank verse, both epic and dramatic, in favour of a variety of cadences with specific connotations and functions within the poem. Blake’s statement, then, foreshadows a view of free verse which is fundamentally different from that of T.S. Eliot, according to whom freedom in metrics is always entangled in a parasitic relationship with some simple metre. These two positions on the issue of metrical freedom in verse – and the roles that each came to play in the early stages of Modernist poetry – will be considered in more detail in the next chapter.

❦

‘Blank verse, said an ingenious critic, seems to be verse only to the eye’. Thus concludes the paragraph by Samuel Johnson with which we began this chapter. The ingenious critic Johnson cites is the art collector William Locke,54 and the lasting ingenuity of his remark is reasserted by Eliot in ‘Milton II’:

53 Blake (1988): 145-146. 54 James Boswell, The Life of Samuel Johnson, edited with an introduction by David Womersley (London: Penguin, 2008): 786. 166

Some of my audience may recall that this last remark, in almost the same words, was often made, a literary generation ago, about the ‘free verse’ of the period: and even without this encouragement from Johnson it would have occurred to my mind to declare Milton to be the greatest master of free verse in our language.55

Eliot’s ingenuous remark is bordering on the absurd, of course. The pentameter matrix as it manifests itself in Paradise Lost is hardly reminiscent of a ghost lurking behind the poem’s arras; in fact it is the metrical strictness of Milton’s blank verse that allows him to construct those breathless leaps of epic enargia of which he is truly the greatest master in the English language. Milton’s syntactic freedom, and the metrical discipline it presupposes, does not come without a price. In the preceding chapter, it was shown how the strictness of Milton’s verse design influences and delimits his diction at line-ends, and in this chapter we have looked at a number of ways in which his verse design facilitates some and inhibits other types of enjambment. In both areas, Milton’s epic use of blank verse diverges systematically from the dramatic use of blank verse: this underscores the necessity of distinguishing sharply between epic and dramatic blank verse, in spite of the fact that they are habitually treated as a single metrical form. The most important feature of Milton’s verse design in Paradise Lost, i.e. the one which sets it apart most clearly from dramatic blank verse, is the scarcity with which syllables with a weakened vowel are allowed to occupy the tenth syllable position. This, importantly, is not a stylistic feature relevant only to the level of phonetics, or to the level of delivery. In terms of diction, it means that proparoxytones are comparatively rare in line-final position throughout Paradise Lost, whereas, in terms of syntax, it means that a variety of phrasal enjambments are avoided by Milton even though they occur regularly in Shakespeare. The fact that even small metrical differences at the level of the verse design have ramifications both at the level of diction and at the level of syntax is an important point in and of itself. With specific reference to blank verse, it helps explain the modernists’ appreciation of the Jacobean playwrights and their deprecation of Milton, and in a wider perspective it may also shed light on the modernist movement’s predilection for free verse.

55 Eliot (1957): 158. 167

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

Draw in thy beams and humble all thy might To that sweet yoke where lasting freedoms be; Which breaks the clouds and opens forth the light, That both doth shine and give us sight to see.

(Philip Sidney, ‘Leave me, O Love’)

Of all the lines in the cryselephantine rag-bag that is Ezra Pound’s Cantos, one is particularly likely to be graced with the odd citation now and then (though seldom cited in its parentheses), especially when the topic is modernist versification:

(To break the pentameter, that was the first heave)1

By the publication of The Pisan Cantos in 1948, the controversy that free verse had stirred in the early decades of the twentieth century was already a thing of the past: that same year saw T.S. Eliot being awarded the Nobel Prize in Literature ‘for his outstanding, pioneer contribution to present- day poetry’.2 In his acceptance speech, Eliot made the following observation:

Poetry is usually considered the most local of all the arts. Painting, sculpture, architecture, music, can be enjoyed by all who see or hear. But language, especially the language of poetry, is a different matter. Poetry, it might seem, separates peoples instead of uniting them. But on the other hand we must remember, that while language constitutes a barrier, poetry itself gives us a reason for trying to overcome the barrier. ... We may think also of the in Europe, and of the great influence that the poetry of one language can exert on another; we must remember the immense debt of every considerable poet to poets

1 Canto LXXXI, l. 55. Cited from Ezra Pound, The Cantos of Ezra Pound (New York: New Directions [1970] 1996). 2 Ottar G. Draugsvold (ed.), Nobel Writers on Writing (Jefferson, NC: McFarland & Co., 2000): 49. 168

of other languages than his own; we may reflect that the poetry of every country and every language would decline and perish, were it not nourished by poetry in foreign tongues.3

Eliot’s remarks on the paradoxical nature of poetry as simultaneously separating and unifying may well be extended (or contracted) to the province of metre. In the preceding chapters we have attempted to trace the rise of the iambic pentameter throughout the history of English versification, from its modest beginnings as an alternative to the octosyllable and through to its gradual canonization in the Renaissance and in Milton. The process by which the iambic pentameter came to be the dominant metrical framework for serious poetry in English was one of syncretism and hybridization rather than one of nationalist or aesthetic essentialism. The English pentameter was, after all, first used as a medium for translating an obscure octosyllabic poem in Old French. As for the development of blank verse, it took a translation of Virgil’s epic hexameters to get rid of the requirement of rhyme. The metre of Virgil’s Aeneid was itself an imitation of that used by his Greek predecessors, the most basic metrical principles of which were later reinterpreted and adapted to fit the phonological peculiarities of the modern vernaculars. Translation, imitation and adaptation are, in other words, not so much superadded functions that may or may not be imposed upon pre- existing metres: they are the very principles by which metre is kept alive and reinvented. These same principles were also at the centre of the modernist movement’s agenda and played, as we shall see, a pivotal role as catalysts for the manifold ways in which the pentameter was broken at the beginning of the twentieth century. Long before the modernists introduced vers libre to the Anglophone world, translations, imitations and adaptations of foreign verse designs had already brought forth a number of poems and genres that are frequently cited as examples of proto-free verse in English. The fascination with the Pindaric ode in seventeenth-century English verse is a case in point. The earliest unambiguous instance of Pindarism in English is by common consent ’s ‘Ode on the Death of Sir H. Morrison’ (1629),4 but it was Abraham Cowley’s collection of Pindarique Odes (1656) which more than any other work established the genre’s great popularity. Even so, Cowley appears to have been highly conscious of the difficulties that the formal properties of his odes might present to his readership:

3 Draugsvold (ed.) (2000): 49-50. 4 Preminger & Brogan (eds) (1993): 856. 169

For as for the Pindarick Odes...I am in great doubt whether they will be understood by most Readers.... The Numbers are various and irregular, and sometimes (especially some of the long ones) seem harsh and uncouth, if the just measures and cadencies be not observed in the Pronunciation. So that almost all of their Sweetness and Numerosity (which is to be found, if I mistake not, in the roughest, if rightly repeated) lies in a manner wholly at the Mercy of the Reader.5

This is a very early version of an argument that was also much favoured by certain modernist proponents of free verse, namely that failure to appreciate a new verse design has everything to do with a poor performance at the level of delivery. Cowley’s doubts were, however, unwarranted, and the Pindaric ode soon became extremely popular among English versifiers. This despite the fact the strict compositional principles underlying ’s odes6 were not understood by any of Cowley’s imitators until the publication of William Congreve’s pamphlet ‘A Discourse on the Pindarique Ode’ in 1706. In addition to outlining Pindar’s formal patterning of , antistrophe and epode, Congreve’s essay also offers valuable observations on the form’s use in English: ‘There is nothing more frequent among us,’ Congreve notes with disapproval, ‘than a sort of Poems intituled Pindarique Odes ... [whose character] is a Bundle of rambling Thoughts, express’d in a like parcel of irregular Stanza’s, which also consist of such another Complication of disproportion’d, uncertain, and perplex’d Verses and Rhimes’.7 And although Congreve speaks favourably of Cowley, he finds it necessary ‘to add, that I believe those irregular Odes of Mr. Cowley, may have been the principal, tho’ innocent Occasion of so many deformed poems since, which ... for the most part have been either Horrid or Ridiculous’.8 The general tone of Congreve’s comments is reminiscent of the tone in which modernist free verse has habitually been criticized by poets of more conservative persuasion in the course of the twentieth century. His comments also draw our attention to the fact that the Pindaric ode was far more widespread during the seventeenth century than might be assumed from glancing at any modern anthology of English verse.9

5 Abraham Cowley, Poems, edited by A.R. Waller (Cambridge: Cambridge University Press, 1905): 10-11. 6 ‘The true form of the Pindaric ode can be described in less than one sentence: it contained one or more triads, each of which consisted of two long stanzas, the strophe and the antistrophe, which were identical; and a concluding stanza, or epode, which was different from the strophe but the same form as all other in the same poem’ (Kirby-Smith (1996): 81-82). The identical structure of the strophe and the antistrophe reflects a dance pattern in which the dancers repeat the steps of the strophe in the antistrophe, but in reverse order. 7 William Congreve, ‘A Discourse on the Pindarique Ode’ (London: 1706): A1. 8 Congreve (1706): A4. 9 On which, see Kirby-Smith (1996): 81-102. 170

Another frequently cited example of proto-free verse in English which also owes its existence to the imitation of a distinctly non-vernacular form is Milton’s Samson Agonistes (1671), specifically the chorus parts. In the affixed note to that play, Milton explains his versification thus:

[The] Chorus is here introduc’d after the Greek manner, not ancient only but modern, and still in use among the Italians. In the modeling therefore of this Poem, with good reason, the Ancients and Italians are rather follow’d, as of much more authority and fame. The measure of Verse us’d in the Chorus is of all sorts, call’d by the Greeks Monostrophic, or rather Apolelymenon, without regard to Strophe, Antistrophe, or Epode.10

Gerard Manley Hopkins, to whom we shall return later in this chapter, contends that the choruses of Samson Agonistes are composed in what he calls counterpoint rhythm, ‘but with the disadvantage that [Milton] does not let the reader clearly know what the ground-rhythm is meant to be, and so they have struck most readers as merely irregular’.11 This perceived irregularity of Milton’s choruses is, of course, exactly what has earned them their prominent position in the canon of proto- free verse in English. However, it is worth investigating Milton’s use of the term apolelymenon to justify his principles of versification. Kirby-Smith speculates that Milton by employing this term may be alluding to the Greek dithyrambic poets in general and to Timotheus of Miletus in particular,12 but a plain translation of the word into English is also revealing: apolelymenon is a participle of the verb ἀπολύω and simply means ‘loosened’. Incidentally, the Italian word sciolto (from sciògliere) – as in versi sciolti – also means ‘loosened’, a fact that may go some way in explaining why some commentators on versification have seen fit to pronounce blank verse a precursor to free verse. From Milton’s comments cited above on his own versification in Samson Agonistes, however, it is quite clear that while his choruses were composed in imitation of ancient models, the Pindaric ode was not one of those models. Even more pertinent to our discussion of the ways in which translation, adaptation and imitation of non-vernacular verse-forms have contributed to the emergence of free verse in English is the use of the Hebrew versicle as a model. A century after Cowley’s popularization of the Pindaric ode, Hebrew prosodic models were taken up by two English writers simultaneously but

10 John Milton, Complete Poems and Major Prose, edited by Merritt Y. Hughes (New York: Odyssey Press, 1957): 550. 11 , The Major Works, edited with an Introduction and Notes by Catherine Phillips (Oxford: Oxford University Press, 2002): 107. 12 Kirby-Smith (1996): 78. 171

independently; they were Christopher Smart and James Macpherson. It is hardly a coincidence that the compositional principles of Hebrew poetry begin to assert themselves in English versification at this time, around 1760. In 1754 Robert Lowth had been appointed Doctor of Divinity by Oxford University in response to his treatise on Hebrew poetry entitled Praelectiones Academicae de Sacra Poesi Hebraeorum (translated into English in 1787 by George Gregory as Lectures on the Sacred Poetry of the Hebrews). The extent to which Robert Lowth’s theories have influenced the area of metrics may perhaps best be illustrated by pointing out the fact that it is to Lowth that we owe the introduction of the word parallelism as a prosodic term in English:

The correspondence of one Verse, or Line, with another, I call Parallelism. When a Proposition is delivered, and a second is subjoined to it, or drawn under it, equivalent, or contrasted with it, in Sense; or similar to it in the form of Grammatical Construction.13

In his treatise on Hebrew prosody, Lowth distinguishes between three species of parallelism: ‘The first species is the synonymous parallelism, when the same sentiment is repeated in different, but equivalent terms’;14 the second species is ‘antithetic parallelism ..., when a thing is illustrated by its contrary being opposed to it ... sentiments are opposed to sentiments, words to words, singulars to singulars, plurals to plurals, &c.’; the third species, ‘which may be called the Synthetic or Constructive Parallelism, ... [is when] the sentences answer to each other, not by the iteration of the same image or sentiment, or the opposition of their contraries, but merely by the form of construction’.15 Whereas the two first species are determined semantically and morphologically, the third species is defined in syntactic terms exclusively and thus includes the figure of anaphora by which the same word or phrase is employed at the beginning of successive lines, sentences or phrases.16 Christopher Smart should be credited with having introduced anaphora as a consistent structural device in English verse. He was intimately familiar with Lowth’s treatise and held it to be

13 Since the Praelectiones Academicae de Sacra Poesi Hebraeorum was written in Latin (for which reason Lowth used parallelismus in that work), this earliest instance of the word parallelism as a prosodic term in English (OED s.v. ‘parallelism’, 3) is from Lowth’s Isaiah; a new translation; with a preliminary dissertation, and notes critical, philological, and explanatory (1778); the term was subsequently taken up by George Gregory for his 1787 translation of Praelectiones Academicae de Sacra Poesi Hebraeorum. 14 Robert Lowth, Lectures on the Sacred Poetry of the Hebrews, translated by George Gregory [1787], New Edition with Notes by Calvin E. Stowe, Vol. II (Andover: Codman Press, 1829): 157. 15 Lowth [1787]: 161-162. 16 On the significance attached to anaphora by Pound, see Line Henriksen, Ambition and Anxiety: Ezra Pound’s Cantos and ’s Omeros as Twentieth-Century Epics (Amsterdam & New York: Rodopi, 2006), 163-223. 172

‘one of the best performances that have been published for a century’.17 Smart made this remark in the February edition of the Universal Visiter or Monthly Memorialist in 1756. Three years later he found himself in a private madhouse in Bethnal Green where he would spend the following four years completing his baroque, heavily anaphoric poem Jubilate Agno. In spite of his radical use of biblical cadences in Jubilate Agno, Smart cannot be said to have had any influence on the development of English free verse in the modernist vein. The manuscript of Smart’s now famous (or, at least, frequently anthologized) poem was not recovered and published until 1939, at which time the free verse movement was more in need of fresh blood than of old models. Macpherson, on the other hand, became the centre of one of the greatest eighteenth-century literary sensations as the self-proclaimed translator of the so-called Ossian poems.18 Though scarcely read today, Macpherson’s pseudepigraphic Fragments of Ancient Poetry, Collected in the Highlands of Scotland, and Translated from the Gaelic or Erse Language (1760) and its sister volumes created a stir all over Europe, winning the admiration of such prominent mainland literati as Klopstock, Herder and Goethe. Samuel Johnson, on the other hand, appears to have seen right through Macpherson’s Ossianic scam: on being asked ‘whether he thought any man of a modern age could have written such poems’, Johnson reputedly replied, ‘“Yes, Sir, many men, many women, and many children.”’19 The prose-like style of Ossian, whose origins Johnson contested, rapidly spawned a string of ‘versifications’, i.e. new versions of Macpherson’s text rewritten in metrical arrangements. We have already looked at William Taylor’s ‘transversions’ of Ossian into accentual hexameters in Chapter 2; earlier examples in other metres include versifications by John Wodrow (1771), Richard Hole (1772) and Ewan Cameron (1776). Dafydd More, whose work is responsible for bringing to my attention the Ossianic versifications of Wodrow, Hole and Cameron, finds ‘[t]he idea of responding to Ossian by versifying it, when perhaps its single most significant characteristic is its prose-poem form, ... worth noting’:20 an assessment so adamantly modest that one can only concur. Furthermore, the margins of the first printing of Fragments of Ancient Poems contained parallel passages from, among other sources, Homer, the Bible and Milton (all of which were removed from subsequent printings),21 thus

17 Cited in Arthur Sherbo, Christopher Smart: Scholar of the University (East Lansing: Michigan State University Press, 1967): 106. 18 For a more thorough outlook on the exact status and reception of Macpherson’s Ossianic poems than space permits here, see Haugen (1998) and Howard Gaskill (ed.), The Reception of Ossian in Europe (London & New York: Thoemmes Continuum, 2004). 19 Boswell (2008): 210. 20 Dafydd Moore, ‘The Reception of The Poems of Ossian in England and Scotland’, in Gaskill (ed.) (2004), 21-39: 33. 21 Haugen (1998): 314. 173

encouraging comparison with the most canonical of texts. While the early reception of Ossian tended to favour comparison with Homer,22 Macpherson’s consistent use of hemistichs and parallelism as structural constituents leads Kirby-Smith to the sensible conclusion that ‘[d]espite Macpherson’s efforts to make his work sound as much like the Iliad as possible, the King James Bible remained the true source of his style’.23 An even more pronounced influence from the King James Bible can be detected in Martin Farquhar Tupper’s Proverbial Philosophy, the first series of which was published in 1837. In spite of their immense popularity during the second half of the nineteenth century,24 Tupper’s philosophical proverbs are so little known today that it may be prudent to give an example; this short excerpt from the section ‘Of Prayer’ may be taken as indicative for the general, somewhat bombastic, style:

Thus, O worshipper of reason, thou hast heard the sum of the matter; And woe to his hairy scalp that restraineth prayer before God. Prayer is a creature’s strength, his very breath and being; Prayer is the golden key that can open the wicket of Mercy; Prayer is the magic sound that saith to Fate, So be it; Prayer is the slender nerve that moveth the muscles of Omnipotence.25

More than one contemporary reviewer of Walt Whitman saw the superficial resemblance between this type of versification and that employed by the author of Leaves of Grass. An unsigned review in The Examiner of Leaves of Grass thus dubbed Whitman ‘a wild Tupper of the West’, and The London Leader commented on Whitman’s ‘wild, irregular, unrhymed, almost unmetrical “lengths,” like the measured prose of Mr. Martin Farquhar Tupper’s Proverbial Philosophy’.26 Henry James,

22 Hugh Blair, who had penned the lengthy essay that accompanied Fragments of Ancient Poetry (and who was on the receiving end of Johnson’s scathing remark on the authorship of Ossian), thus devoted forty pages of his ‘Critical Dissertation upon the Poems of Ossian’ (1763) to comparing the ‘manners’ of Ossian and Homer. The topic was subsequently taken up by Herder in his Homer und Ossian (1795). However, in Blair’s 1783 Lectures the focus is on comparisons with the Old Testament and the Koran (Haugen 1993: 314, nn. 28-29). Like Christopher Smart, Blair was intimately familiar with Lowth’s treatise on Hebrew versification and praises it in the Lectures; on this, see Josef Bysveen, Epic Tradition and Innovation in James Macpherson’s Fingal (Uppsala, 1982): 26-30. 23 Kirby-Smith (1996): 149. 24 During the nineteenth century, Tupper’s Proverbial Philosophy thus passed through as many as forty editions in England and sold close to a million copies in the United States. 25 Cited from Martin Farquhar Tupper, Tupper’s Proverbial Philosophy: A Book of Thoughts and Arguments, Originally Treated, First and Second Series (Milwaukie [sic]: I.A. Hopkins, 1849): 58. 26 The complete texts of both reviews can be found in Milton Hindus (ed.), Walt Whitman: The Critical Heritage (London: Routledge and K. Paul, 1971): 90-96. 174

in his unsigned review of Drum-Taps in The Nation on 16 November 1865,27 is also less than charitable: ‘Our author’s novelty, however, is not in his words, but in the form of his writing. As we have said, it begins for all the world like verse and turns out to be arrant prose. It is more like Mr. Tupper’s proverbs than anything we have met’. Equally dismissive was Swinburne’s initial opinion of Whitman, voiced in The Fortnightly Review in 1887:

If then it appears that Mr. Walt Whitman has about as much gift of song as his precursors and apparent models in rhythmic structure and style, Mr. James Macpherson and Mr. Martin Tupper, his capacity for creation is the only thing that remains for us to consider. And on that score we find him, beyond all question, rather like the later than like the earlier of his masters. Macpherson could at least evoke shadows: Mr. Tupper and Mr. Whitman can only accumulate words.28

Of course Mr. Whitman’s accumulation of words turned out to have a considerably more lasting influence on modern poetry than anything ever written by either Tupper or Macpherson. Or by Swinburne, for that matter. Whitman never cites Tupper as an inspiration, although he does speak favourably of Tupper’s Probabilities, an Aid to Faith (1847) at one point.29 As for the possibility of influence from Macpherson on Whitman’s versification, Timothy Steele points us to an interesting, albeit decidedly marginal, remark by Whitman:

Whitman once clipped a paragraph in which Margaret Fuller recorded an “Ossianic” moment, and noted in the margin, “Don’t fall into the Ossianic, by any chance” but added further on, “Is it not Isaiah, Job, the Psalms, and so forth, transferred to the Scotch Highlands?”30

This little scrap of paratext is revealing: its self-admonitory tone suggests that Whitman was certainly aware of, if not downright anxious about, the possibility of influence from Ossian, but it also suggests that his real – real in the sense of acknowledged – source of inspiration was the Old Testament. This, of course, was also recognized by Whitman’s contemporaries: Robert Buchanan,

27 Reprinted in Hindus (ed.) (1971): 110-114. 28 Algernon Charles Swinburne, ‘Whitmania’, reprinted in Hindus (ed.) (1971), 199-207. 29 See Steele (1990): 298, n. 12. 30 Steele (1990): 154. 175

for instance, describes Whitman’s poetry as ‘very Biblical in form, and showing indeed on every page the traces of Biblical influence’.31 This point – that Hebrew is the most obvious and likely model for Whitman’s versification – has been made so frequently that it hardly needs further qualification. When Shin Shalom in 1950 undertook the task of translating Whitman into Hebrew, he thus explained that ‘Whitman’s pioneering is very close to us, and so are his Biblical rhythms. To translate him into Hebrew is like translating a writer back into his own language’.32

❦

When we reach Whitman, we have arrived at a corpus of English poetry that is, for all intents and purposes, consciously free of metrical constraints. Instead – as in the Hebrew versicle – lineation is subject to the syntactic constraint of having to co-occur with clausal or phrasal structures. In addition to making enjambment virtually impossible, this principle of composition also has implications for the status of the line as a unit: whereas in metrical verse the line-unit reflects the poem’s phonological features, Whitman’s loose line-units are determined principally by the poem’s syntactic structures. This fits well with what we observed earlier about the general relationship between metre and syntax: that syntactic freedom across line breaks tends to presuppose a strict and well-defined line structure, whereas a looser line structure tends to encourage syntactic reinforcement of line breaks. But it fits less well with Gasparov’s stipulation that ‘verse is segmented into equivalent and commensurable segments independent of syntax, and prose only in relation to syntax’.33 By Gasparov’s standard, Whitman’s poetry may seem to come dangerously close to falling into prose, but to this it must be objected that even though Whitman’s verse is never segmented into lines independently of syntax, syntax is not the sole principle by which his lineation is determined. Consider, for instance, the lineation in a short poem such as this one from Leaves of Grass:

When I heard the learn’d astronomer, When the proofs, the figures, were ranged in columns before me, When I was shown the charts and diagrams, to add, divide, and measure them,

31 In Broadway Magazine (Nov., 1867). Reprinted in Kenneth M. Price, Walt Whitman: The Contemporary Reviews (Cambridge: Cambridge University Press, 1996), 137-143. 32 Cited in Hindus (1971): 17. 33 Gasparov (1996): 97. 176

When I sitting heard the astronomer where he lectured with much applause in the lecture- room, How soon unaccountable I became tired and sick, Till rising and gliding out, I wander’d off by myself, In the mystical moist night-air, and from time to time, Look’d up in perfect silence at the stars.34

The first four lines are structured by the anaphoric ‘When’ and increase in length until they culminate in line 5, the exclamative clause of which lines 1-4 are paratactic adverbial clauses. Had the poem ended here, a good case might have been made for considering it a prose piece with the odd typographic feature of having its basic syntactic structure stamped out by lineation. However, the grammatical structure of line 6 is carried over the line break into line 7: the prepositional phrase ‘In the mystical moist night-air’ is clearly linked to the verb phrase ‘wander’d off’ in the preceding line. The second half of line 7, ‘and from time to time’, is, however, just as clearly linked, across another line break, to the verb phrase ‘Look’d up’ in line 8. None of these line breaks meets the condition of enjambment, of course, but had the final tercet of the poem followed the same syntactic principle of lineation that is applied in lines 1-5, we would have expected to see a couplet like this:

Till rising and gliding out, I wander’d off by myself, in the mystical moist night-air, And from time to time, look’d up in perfect silence at the stars.

The fact that line 7 in Whitman’s original poem does not constitute a syntactic structure in itself, but rather completes one structure and begins another is all the proof we need to establish that Whitman’s poetry does fall safely under the category of verse. His method of lineation is not one of automatized reproduction of syntactic structures, but rather a tool for juxtaposing and stacking those structures in various degrees of equilibrium. This occasionally produces individual verse instances that meet the requirements of traditional English verse designs, such as the last line in Whitman’s poem, an uncontroversial iambic pentameter by any standard: ‘Look’d up in perfect silence at the stars’.35

34 Cited from Walt Whitman, The Complete Poems, edited by Francis Murphy (London: Penguin, [1975] 1996). 35 The potential significance of such metrically regular lines in Whitman’s poetry is treated in some detail in Annie Finch, The Ghost of Metre: Culture and Prosody in American Free Verse (Ann Arbor, MI: University of Michigan Press, 1990), 31-56. 177

By abandoning the fixed line-length of conventional metre in favour of the malleable line- unit of syntactic parallelism, Whitman simultaneously, and paradoxically, reified and subverted the line’s status as the most basic unit of English versification. On the one hand, aligning lineation with syntactic structure has the effect of making the individual line seem less arbitrary and more autonomous in and by itself, but on the other hand the same procedure undermines the fact that it is on account of their very arbitrariness that metrical lines are capable of being employed as structurally equivalent building blocks in metrical verse. William Carlos Williams emphasized in particular Whitman’s subversive effect on the unit of the line:

Whitman to me was an instrument, one thing: he started us on the course of our researches into the nature of the line by breaking finally with English prosody. After him there has been for us no line.36

This estimation of Whitman’s contribution to the development of modernist poetry, the somewhat reserved acknowledgement of his pioneer status, is reminiscent of Pound’s unilateral peace-offering in ‘A Pact’:

I make a pact with you, Walt Whitman– I have detested you long enough. I come to you as a grown child Who has had a pig-headed father; I am old enough now to make friends. It was you that broke the new wood, Now is a time for carving. We have one sap and one root– Let there be commerce between us.37

Even more reluctant to embrace the legacy of Whitman was Eliot, who repeatedly and vehemently denied any influence from Whitman on his own work or on that of Pound.38 Thus, in ‘Ezra Pound:

36 The Selected Letters of William Carlos Williams, edited by John C. Thirlwall (New York: New Directions, 1957): 286-287. 37 Cited from Pound (2001). 38 On this, see James E. Miller, ‘Personal Mood Transmuted into Epic: T.S. Eliot’s “Waste Land”’, in Harold Bloom (ed.), T.S. Eliot’s The Waste Land, Updated Edition (New York: Chelsea House, 2007), 49-74. 178

His Metric and Poetry’ (1917) Eliot states that ‘Whitman is certainly not an influence; there is not a trace of him anywhere. Whitman and Mr. Pound are antipodean to each other’;39 and eleven years later, in his preface to Pound’s Selected Poems (1928), Eliot carefully distinguishes his own work from the kind of poetry that takes Whitman40 as a model: ‘To be more precise, there are, for instance, my own type of verse, that of Pound, and that of the disciples of Whitman’.41 On his own type of verse, Eliot comments:

I will not say that subsequently there have not appeared traces of reciprocal influence of several types [of free verse] upon one another, but I am here speaking of origins. My own verse is, so far as I can judge, nearer to the original meaning of vers libre than is any of the other types: at least, the form in which I began to write, in 1908 or 1909, was directly drawn from the study of Laforgue together with the later Elizabethan drama; and I do not know anyone who started from exactly that point. I did not read Whitman until much later in life, and had to conquer an aversion to his form, as well as to his matter, in order to do so.42

One very likely ‘reciprocal influence’ of Whitman on ‘the original meaning of vers libre’ – and one which Eliot does not mention here – is the influence that Whitman may have exerted on Jules Laforgue through the latter’s role as one of Whitman’s first translators into French.43 The importance of Laforgue and the French symbolists to Eliot and his contemporaries can hardly be overestimated, but it is important to realize that the relationship between Anglo-American and French versification during the late nineteenth and early twentieth centuries was not quite as unidirectional as is sometimes assumed. The musically tinted poetics of the symbolist movement owed much to the principles laid down in Edgar Allan Poe’s ‘The Philosophy of Composition’ (1846) and ‘The Poetic Principle’ (1848), so it is hardly surprising that among Poe’s earliest

39 Eliot (1965): 177. 40 In a remark to Sholom Kahn, Eliot stated the difference between Whitman’s verse and his own even more bluntly: ‘I must say frankly that it seems to me you are wasting your time in attempting to relate my work to that of Walt Whitman’. Cited in William R. Everdell, The First Moderns (Chicago, IL: Chicago University Press, 1997): 99. 41 The ‘disciples of Whitman’ Eliot refers to are most likely Chicago-based poets such as Carl Sandburg and Vachel Lindsay, all of whom were regularly published in Harriet Monroe’s magazine Poetry, of which Pound was foreign editor during its early years. In that capacity, Pound had secured Eliot his public debut as a poet with the publication of ‘Prufrock’ in the June issue of Poetry in 1915. 42 T.S. Eliot, ‘Introduction’ [1928] to Ezra Pound, Selected Poems, edited with an Introduction by T.S. Eliot (London: Faber and Faber, 1948): 8. 43 Whitman’s first French translator was Louis Etienne, whose translation was published in La Revue Européenne (November 1, 1861) under the remarkable title ‘Walt Whitman, poète, philosophe et “rowdy”’; see Roger Asselineau, ‘Whitman in France and Belgium’, in Gay Wilson Allen & Ed Folsom (eds), Walt Whitman & the World (Iowa City, IA: University of Iowa Press, 1995), 233-244. 179

translators into French we find such prominent symbolists as Charles Baudelaire and Stéphane Mallarmé.44 The latter’s lifelong involvement with the English language and its literature is particularly well attested: in addition to earning his livelihood as a teacher of English45 up until his retirement in 1893, Mallarmé had lived in London for a year as a student and had even published a treatise on English philology (Les Mots anglais, 1878). London was also where and had taken refuge in September 1872, spending there nine tumultuous months in hiding from Verlaine’s wife and in-laws. After a short but even more tumultuous interim on the European mainland, Rimbaud returned to London in 1874 – this time accompanied by Germain Nouveau – to complete Les Illuminations, which collection even includes a piece with an English title, ‘Being Beauteous’. London had also provided the setting for Mallarmé’s marriage to Marie Gerhard in the early 1860s, as it did indeed for Laforgue’s marriage to Leah Lee some twenty-five years later, at St. Barnabas’s Church in Kensington.46 A native of Teignmouth in Devon, Leah Lee had been teaching English to Laforgue on a regular basis in the year leading up to their marriage on the last day of 1886;47 her endeavours enabled him, in June of 1886, to commence work on a project entitled Brins d’herbe, a translation of Whitman’s Leaves of Grass.48 Laforgue’s premature death from tuberculosis the following year prevented him from completing the task, but his initial endeavours were published over three issues of Gustave Kahn’s literary periodical La Vogue during the summer of 1886, among them ‘Une Femme m’attend’ (‘A Woman Waits for Me’). Whitman was, in other words, far from unknown to the French symbolists, as L.E. Kastner pointed out as early as 1903 in his History of French Versification: ‘Walt Whitman’s irregular metres seem to have been the starting-point of some of the vers-libristes’.49 Kastner was being cautious: in point of fact, Whitman has been explicitly recognized as an influence by a long string of poets associated with French symbolism, among them

44 In a letter to Sara Rice, dated April 4, 1876, Mallarmé acknowledges his indebtedness to both Poe and Baudelaire: ‘I’ll ask you to leave room for one of those magnificent pages of prose that our great dead poet, my master Charles Baudelaire, wrote as a preface to his immortal translation of [Poe’s] Tales. You are certainly aware that he is the one who made Poe known and popular in France’ (Stéphane Mallarmé, Selected Letters, edited and translated by Rosemary Lloyd (Chicago, IL: University of Chicago Press, 1988): 115). 45 As did both Verlaine and Rimbaud in the years following their affair. 46 This, incidentally, is the same church in which T.S. Eliot married Valerie Fletcher in 1957, as Valerie Eliot explained in a letter to the Times Literary Supplement (July 1, 1967): ‘St Barnabas was chosen because the vicar was a friend of [Eliot’s] solicitor and we wanted privacy. Shortly before the wedding, while glancing through a book on his desk, my husband discovered to his pleasure that Laforgue had been married at this church. After our ceremony (at 6.15 a.m.) the Rev C. P. Wright gave us breakfast at his home, 10 Kensington Church Walk, where Ezra Pound had lived.’ Cited in David Arkell, Looking for Laforgue: An Informal Biography (Manchester: Carcanet, 1979): 244, n. 66. 47 Arkell (1979): 174-175, 223-225. 48 Everdell (1997): 81, 93. 49 Kastner (1903): 70, n. 1. 180

Kahn, Laforgue, Marie Krysinska, Jean Moréas, Maurice Maeterlinck, Guillaume Apollinaire, Francis Vielé-Griffin and Stuart Merrill.50 So pervasive was his influence that the Belgian Emile Verhaeren, whose work is perhaps closer to Whitman’s than is that of any of the above, is sometimes cited as an example of a symbolist poet who claimed not to have been influenced by Whitman.51 Eliot’s insistence that he owes nothing to Whitman, but much to the French vers libristes, must therefore be tempered by the recognition that the French vers libristes owed a great deal to Whitman; Roger Asselineau even goes as far as to assert that while ‘Vers libre caught on very slowly in France, ... it definitely had its origin in Leaves of Grass’.52 Asselineau’s assertion that French vers libre owed its very existence to Leaves of Grass may be stretching it in that it ignores the very likely possibility of influence from à la Aloysius Bertrand and Charles Baudelaire.53 Anne Holmes, however, makes a persuasive argument concerning the way in which translations of Whitman’s unmetered cadences may have contributed specifically to the development of vers libre in France:

A common vehicle for translation from English into French was prose—for example, Mallarmé’s translations of Poe—the assumption being that French verse imposed insuperable difficulties on the translator. Some poets, Elizabeth Browning and Rossetti among them, were rendered line by line, and so, faute de mieux, in a kind of free verse, but no significance was attached to the form. Whitman’s own free verse, however, removed the difficulty of choice, since it was obviously the natural medium, and this presumably meant that Laforgue could respond to the challenge of offering an equivalent on the stylistic as well as on the semantic level.54

50 Everdell (1997): 96. 51 See Huberta Frets Randall, ‘Whitman and Verhaeren – Priests of Human Brotherhood’, in The French Review, Vol. 16, No. 1 (Oct., 1942), 36-43: 36. 52 Asselineau (1995): 236. 53 For an account which calls Whitman’s influence on French vers libre into question, see P.M. Jones, ‘Influence of Walt Whitman on the Origin of the “Vers Libre”’, in Modern Language Review, 11 (1916), 186-194. 54 Holmes (1993): 99. For an example of the assumption that French is a particularly difficult target language as far as translation goes, we may turn to some remarks made by Jean le Rond d’Alambert, who was Diderot’s most important co-editor of the Encyclopédie. In the preface to his translation of Tacitus (‘Observations sur l’art de traduire’, 1758), d’Alambert notes: ‘The question has been raised whether poets can be translated in verse, especially in [French], which does not admit of unrhymed verse, unlike English and Italian, and which allows no leeway to either the poet or the translator. Some of our writers have maintained that poets cannot be translated into prose, either because they like poetry or because they like problems. Prose translations, they add, would disfigure poets and deprive them of their principal charm, measure, and harmony. … Can anyone seriously believe that [French] poetry with its rhymes, its half-lines that are always similar, the uniformity of its progression and its monotony—if one may say so—might be able to represent the varied cadences of 181

Laforgue’s 1886 translations of Whitman are not the earliest instances of French vers libre in the symbolist vein, but his posthumously published collection Derniers Vers (1890) was the first volume in French to be composed entirely in free verse.55 Three years earlier, in 1887, Laforgue’s friend and editor, Gustave Kahn, had published his own Les Palais nomades, the first collection of French poetry to contain a significant number of free verse poems.56 On that ground Kahn, unsurprisingly, claimed the honour of having invented vers libre, but letters exchanged between Laforgue and Kahn suggest a less clear-cut picture; of particular importance is a letter sent from Laforgue to Kahn in early August 1886, in which Laforgue, in his usual casual manner, informs his friend of certain new developments in his own approach to versification:

I forget to rhyme, I forget the number of syllables, I forget to set it in stanzas—the lines themselves begin in the margin just like prose. ... I’ll have a book like this ready when I come to Paris. I’m working on nothing else. ... And I’ll never write poetry any different from what I’m writing now.57

Regardless of who ‘invented’ vers libre, it is safe to say that it was Laforgue rather than Kahn who came to exercise the greatest influence on the ways in which free verse would develop in the Anglo-American world over the next decades. Part of the explanation for this is that Kahn was not accorded a chapter of his own in Arthur Symons’s seminal The Symbolist Movement in Literature (1899) whereas Laforgue was. In that work, Laforgue’s poetry is characterized as being ‘really vers libre, but at the same time correct verse, before vers libre had been invented’,58 and Symons’s use of the term vers libre is particularly noteworthy in that it predates the OED’s earliest attestation of this term in English by three years. In 1889 Arthur Symons had travelled to Paris in the company of Havelock Ellis, who introduced his travelling companion to a string of leading figures in the literary establishment, among them Verlaine, Mallarmé and Joris-Karl Huysmans.59 These encounters spawned a series of essays by Symons, which were subsequently collected and edited for the publication of his influential volume. Dedicated to William Butler Yeats, The

Greek and Latin poetry?’ (Jean le Rond d’Alambert, ‘Remarks on the Art of Translating’, in André Lefevere (ed.), Translation, History, Culture: A Sourcebook (London: Routledge, 1992), 105-115: 110) 55 Holmes (1993): 95. 56 Holmes (1993): 99, n. 19. 57 Cited in Arkell (1979): 196-197. 58 Arthur Symons, The Symbolist Movement in Literature (London: William Heinemann, 1899): 106. 59 John M. Munro, Arthur Symons (New York: Twayne Publishers, 1969): 30. 182

Symbolist Movement in Literature had a profound and lasting effect on the Anglo-American literary world and was even republished in an expanded edition as late as 1919.60 Among the earlier edition’s most ardent readers was T.S. Eliot, who first came across Symons’s book at Harvard in October 1908 and subsequently on several occasions credited it with having introduced him to Rimbaud, Verlaine and – most importantly – Laforgue.61 From a formal point of view, the kind of versification that Laforgue inspired in Eliot ‘in 1908 or 1909’ is not very free; his ‘Humouresque (After J. Laforgue)’, for instance, composed in November 1909, consists of neatly rhymed tetrameter quatrains, a form to which Eliot would subsequently return in Poems (1920). However, in some of the early poems that Eliot did not publish himself – especially those that remained unpublished until 1996 when Christopher Ricks edited Eliot’s note book, Inventions of the March Hare – the versification is looser and prefigures the type of verse that we find in, say, ‘The Love Song of J. Alfred Prufrock’. Written between 1910 and 1911, and first printed in Poetry in 1915, ‘The Love Song of J. Alfred Prufrock’ was the centrepiece of Eliot’s debut collection, Prufrock and Other Observations, published by the Egoist Press in June of 1917. Earlier that same year, Eliot had had a short essay in the New Statesman under the heading of ‘Reflections on Vers Libre’, an essay which has since proven to be among the most influential documents in the now century-old canon of free verse poetics. It is in ‘Reflections on Vers Libre’ that Eliot codifies what has subsequently become one of the most commonly held views on free verse: that good free verse is never really free, but depends for its success entirely on being in a parasitic relationship with traditional metrics. Eminently quotable throughout, Eliot’s ‘Reflections on Vers Libre’ begins with the provocative assertion that ‘Vers libre does not exist’; had it been ‘a genuine verse-form it [would] have a positive definition’. Eliot, however, ‘can define it only in negatives: (1) absence of pattern, (2) absence of rhyme, (3) absence of metre’. As for absence of metre, Eliot cannot say ‘What sort of a line that would be which would not scan at all’ since ‘Any line can be divided into feet and accents’. Elaborate metrical studies are, however, of limited value since ‘the most interesting verse which has yet been written in our language has been done either by taking a very simple form, like

60 Of comparable importance to the development of free verse based on French models in English was Adrien Van Bever and Paul Léautaud’s two-volume anthology Poètes d’aujourd’hui covering French symbolist verse published between 1880 and 1900. This anthology was tirelessly promoted by F.S. Flint in various London journals (especially The New Age) from 1908 onwards and is referred to favourably by poets such as Pound, Richard Aldington, H.D. and ; see Cyrena N. Pondrom, The Road from Paris: French Influence on English Poetry 1900-1920 (Cambridge: Cambridge University Press, 1974): 11. 61 A collection of Eliot’s most important references to Symons’s book and the influence of French symbolism on his own verse can be found in T.S. Eliot, Inventions of the March Hare: Poems 1909-1917, edited by Christopher Ricks (New York: Harcourt Brace & Company, 1996): 399-409. 183

the iambic pentameter, and constantly withdrawing from it, or taking no form at all, and constantly approximating to a very simple one’. That is to say that ‘the ghost of some simple metre should lurk behind the arras in even the “freest” verse; to advance menacingly as we doze, and withdraw as we rouse. Or, freedom is only truly freedom when it appears against the background of an artificial limitation’. In conclusion: ‘There is no escape from metre; there is only mastery’.62 Despite the regular frequency with which the main tenets of Eliot’s argument against free verse have been reiterated, cited and paraphrased throughout the twentieth century, the fundamental assumptions on which the argument rests have rarely been called into question. However, Eliot’s claim that free verse does not exist because it cannot be defined in other than negatives is unnecessarily pedantic, not only because free verse does exist, and did so in 1917, but especially because it presupposes that negative statements are inherently invalid for the purpose of taxonomy. This, of course, is not true. Remarking on an ostrich’s lack of flying-abilities is not necessarily more trivial than pointing out its ambipedal stature; of the two observations, the former may under certain circumstances even be the more salient. But what is particularly striking about Eliot’s discarding of negative definitions in the context of metrics is that traditional English metres rely so essentially on the principle of counterpointing prominent syllables with syllables that must be defined negatively, from the context in which they occur, as non-prominent. However, as outlined in Chapter 2, this insight was first given due consideration by Otto Jespersen, but in a wider perspective the recognition of negatively defined linguistic properties was, of course, to become a cornerstone of structuralism. The era of structuralism is conventionally viewed as beginning with the posthumous publication of Ferdinand de Saussure’s Cours de linguistique générale one year prior to Eliot’s ‘Reflections on Vers Libre’, in 1916. One of the most significant outcomes of Saussure’s work was arguably the development of phonology as an academic discipline in its own right, most prominently by Roman Jakobson and Nikolai Trubetzkoy. As established by the Prague School, the discipline of phonology rests on a finite inventory of systematic oppositions that take for granted – indeed depend upon – the validity of negatively defined linguistic features, as Trubetzkoy demonstrates in this excerpt from Grundzüge der Phonologie (1939):

The definition of the content of a phoneme depends on what position this phoneme takes in the given phonemic system, that is, in final analysis, with which other phonemes it is in

62 Eliot (1965): 183-189. 184

opposition. A phoneme can therefore sometimes be defined in purely negative terms. For example, if one considers all the optional and combinatory variants of the German phoneme r, the only way in which this phoneme can be defined is as a ‘nonlateral liquid.’ This is a purely negative definition since a ‘liquid’ itself is a ‘nonnasal sonorant,’ and a ‘sonorant’ is a ‘nonobstruent.’63

To account in more detail for the internal relationship between such polar oppositions, Jakobson and Trubetzkoy developed in tandem the concept of markedness, which posits that the poles of oppositional pairs are not equipollent but rather enter into a strict hierarchy of valorization, in which one pole is considered to be more general and inclusive than its opposite. In this hierarchy, the more general pole is termed ‘unmarked’ whereas its more specialized counterpart is termed ‘marked’.64 The notion of markedness is, however, not only applicable to phonological oppositions, but is pertinent to all levels of linguistic description: in Chapter 5 we noticed, for instance, that SVO is the unmarked word order of English declarative sentences. In fact, according to Jakobson, ‘every single constituent of a linguistic system is built on an opposition of two logical contradictories: the presence of an attribute (“markedness”) in contraposition to its absence (“unmarkedness”)’.65 Applied to the area of versification, these principles would stipulate that metrical utterances enter into an oppositional relationship with utterances that are not metrical, and that metrical utterances furthermore constitute the marked pole in that relationship. Metre, then, may be described as a marked mode of organizing one or more marked phonological features of a given language, as when stressed syllables are manipulated into apportioned iambic sequences in English. And while it is true that all viable metrical systems exploit naturally occurring phonological traits to achieve and sustain their effect, the (marked) regularity with which such (marked) traits are organized in metrical utterances sets them apart quite unambiguously from utterances in which this level of organization is absent. It is, in other words, the markedness of metrical utterances that allows us to define verse forms based on such utterances positively, just as it is the unmarkedness of non-metrical prosodic patterns that prevents us from defining the verse forms that may be built upon them positively. But that does not mean that verse forms based on unmarked prosodic patterns

63 N.S. Trubetzkoy, Principles of Phonology, translated by Christiane A.M. Baltaxe (Berkeley & Los Angeles, CA: University of California Press, 1969), second printing (1971): 67. 64 On the topic of markedness, see Edwin L. Battistella, Markedness: The Evaluative Superstructure of Language (Albany, NY: State University of New York Press, 1990). 65 Roman Jakobson, ‘Verbal Communication’, in Scientific American 227 (1972), 72-80; cited in Edwin Battistella, ‘Jakobson and Chomsky on Markedness’, in Eva Hajičovà et al (eds), Prague Linguistic Circle Papers, Vol. 1 (Amsterdam: John Benjamins, 1995), 55-72: 57. 185

do not, or cannot, exist. To say, as Eliot does, that ‘Any line can be divided into feet and accents’ is not only patently wrong (one has yet to see a metrical foot that can keep up with the erratic movements of E.E. Cummings’s Grasshopper, for instance), but also suggests a view of metre so wilfully misguided that we must attribute it to Eliot’s polemical style and, perhaps, to the defensiveness of a practising poet. Still, it must be admitted that Eliot’s reflections on vers libre quite adequately reflect the principles underlying his own type of free verse, as it unfolds in his debut collection published a few months after the ‘Reflections’:

Let us go then, you and I, When the evening is spread out against the sky Like a patient etherised upon a table; Let us go, through certain half-deserted streets, The muttering retreats Of restless nights in one-night cheap hotels And sawdust restaurants with oyster-shells: Streets that follow like a tedious argument Of insidious intent To lead you to an overwhelming question . . . Oh, do not ask, ‘What is it?’ Let us go and make our visit.66

This kind of verse lends itself quite easily to traditional scansion by foot, although not in a consistent manner. Even so, there appears to be some kind of effort present to organize the manner in which these twelve lines interact with one another. First of all, it may be noticed that the opening line and the concluding line of the verse paragraph can be scanned as iambic : both lines are acephalous, and the final line has a feminine ending:

66 Cited from T.S. Eliot, The Complete Poems and Plays (London: Faber and Faber, 2004). 186

/ . / . / . / Let us go then, you and I,

/ . / . / . / . Let us go and make our visit.

The occurrence of iambic tetrameters at both ends of the verse paragraph would seem to indicate the presence of a chiastic structure to the paragraph as a whole, and that indication is made stronger by the fact that the two middle lines are wrought in iambic pentameter:

. / . / . / . / . / Of restless nights in one-night cheap hotels

. / . / . / . / . / And sawdust restaurants with oyster-shells:

Using to denote unanalyzed lines and numerals to designate the number of feet in lines that have been analyzed, the verse paragraph is structured like this: 4oooo55oooo4. This chiastic pattern is underscored by the paragraph’s rhyme scheme – aabccddeefgg – in which the third and the third-to-last lines are the only ones that do not rhyme with another line. Both of these lines may, incidentally, be scanned as iambic pentameters with feminine endings; line 3 has a double anacrusis:

. . / . / . / . / . / . Like a patient etherised upon a table;

. / . / . / . / . / . To lead you to an overwhelming question . . .

This gives us the pattern 4o5oo55oo5o4 for the verse paragraph as whole: the chiastic plot thickens. Or does it? In order to justify the scansions presented above, we have had to call upon such decidedly specialized metrical terms as ‘double anacrusis’ and ‘acephalous’ to account for the ways 187

in which individual verse instances may be stretched and clipped into metrical patterns, and this after having looked at only half of the lines that make up Eliot’s verse paragraph. By doing so, we have been playing the game of a latter-day Procrustes and have ignored the arras while looking for the ghost’s foot prints. However, even the occasional occurrence of individual lines that meet the requirements of established metrical forms exactly – such as, for instance, the iambic pentameters of line 6 or 7 in ‘Prufrock’ – does not make a poem metrical. In metrical verse, metre functions as a structural device that delimits and controls the entire process of versification, but in Eliot’s verse metre has been demoted (or promoted) to a powerful ornamental device that may be called upon, hinted at or nudgingly alluded to at the poet’s whim. When executed skilfully, this mode of versification produces a kind of poetry which is continuously engaged in a dialectic relationship with the metrical tradition, but which is not metrical itself, at least not in the strict sense in which metre is defined in this dissertation. What John Hollander observes about metrical scansion of free verse in general applies to ‘Prufrock’, too:

Thus, to try to scan free verse by counting the number of stresses and concluding that, in any event, the poem under discussion is roughly assembled of four- and five-stressed lines, may be merely to assert a trivial correlation built into the structure of English ... the probability is that – unless his [the poet’s] syntax is most distorted, his density of short, emphatic monosyllabic imperatives and expletives, his latinate [sic] polysyllables likewise – most of his lines will have three or four stresses, if counted in accentual-syllabic terms. But it would surely be wrong to hold the stress-patterning to be the principle.67

Eliot’s verse in ‘Prufrock’ consequently fits the negative definition of free verse, even though it is in a certain sense closer to metrical verse than is Whitman’s versification in Leaves of Grass. Even so, it ought to be noticed that ‘Prufrock’ shares with Whitman’s verse a tendency to avoid enjambment. In the 132 lines that make up ‘Prufrock’ there are only four instances of enjambment; two of those are phrasal and occur in the opening verse paragraph (‘retreats / Of restless nights’ and ‘argument / Of insidious intent’); the other two are clausal (‘Do I dare / Disturb the universe?’ and ‘how should I begin / To spit out all the butt-ends of my days and ways?’). In Eliot’s later poems in free verse, however, the ratio of enjambment goes up (consider, for instance, the opening lines of The Waste Land): an indication, perhaps, of Eliot’s growing confidence in his own mastering of the

67 Hollander (1975): 235. 188

medium whose existence he had denied in 1917, when he stated that ‘There is no escape from metre; there is only mastery’. Metre and mastery were brought together again one year later with the publication by Robert Bridges of Gerard Manley Hopkins’s body of poetry in 1918. ‘Thou mastering me / God!’ is the startling opening of The Wreck of the Deutschland, Hopkins’s first major poem in what he termed Sprung Rhythm. But the idea of mastery also figures prominently in the poem which Hopkins considered his best, ‘The Windhover’:

I caught this morning morning’s minion, king- dom of daylight’s dauphin, dapple-dáwn-drawn Falcon, in his riding Of the rólling level úndernéath him steady áir, and stríding High there, how he rung upon the rein of a wimpling wing In his ecstasy! then off, off forth on swing, As a skate’s heel sweeps smooth on a bow-bend: the hurl and gliding Rebuffed the big wind. My heart in hiding Stirred for a bird,—the achieve of; the mastery of the thing!68

Hopkins’s poetry and the idiosyncratic metrics that underlies it have received much critical attention since 1918 and are frequently cited as a nineteenth-century precursor of free verse, almost on a par with Whitman.69 In 1933, , himself a prolific writer of modernist free verse, asserted that ‘the virtue of Hopkins’s theory of Sprung Rhythm is that it justifies every departure from regularity of rhythm: it justifies, in fact, free rhythm’.70 Less enthusiastic were Eliot’s remarks on the same subject in After Strange Gods (1934):

Hopkins is a fine poet, to be sure; but he is not nearly so much a poet of our time as the accidents of his publication and the inventions of his metric have led us to suppose. His innovations certainly were good, but like the mind of their author, they operate only within a narrow range, and are easily imitated though not adaptable for many purposes...71

68 Cited from Hopkins (2002). 69 On this, see Michael D. Hurley, ‘What Sprung Rhythm Really Is NOT’, in The Hopkins Quarterly 33 (2006), 71-94. 70 Herbert Read, ‘Gerard Hopkins’s Meters’, in TLS 1627 (1933): 127. 71 T.S. Eliot, After Strange Gods (London: Faber and Faber, 1934): 47. 189

Read’s comments on Hopkins’s versification are cited from a recent article by Peter L. Groves, in which ‘The Windhover’ is subjected to an interesting metrical analysis that treats Hopkins’s sonnet as an attempt to ‘open up’ the iambic pentameter. Taking his cue from Hopkins’s remark that ‘The development [of Sprung Rhythm] is mine but the beat is in Shakspere’, Groves sees Hopkins’s metrics as ‘an exploration and extension of some of Shakespeare’s more peculiar metrical practices, in particular his use of catalexis, or the representation of offbeats by syllabic gaps or absences’.72 Groves discards the common assumption that Hopkins’s notion of counterpoint primarily applies to the interaction between a fixed metre and the rhythm of individual lines; what in fact is being counterpointed in Hopkins’s work is traditional metrics and Sprung Rhythm. From the discussion of Milton’s choruses in Samson Agonistes earlier in this chapter it will be recalled that Hopkins’s primary criticism of Milton’s versification in that work was that Milton ‘does not let the reader clearly know what the ground-rhythm is meant to be’. The same charge cannot be levelled against Hopkins’s practice in ‘The Windhover’: that sonnet’s opening line is a formidable iambic pentameter whose line-final hyphen hurls the reader headlong into a succession of lines that, in Groves’s analysis, radically challenge the conventions of iambic pentameter without abandoning them completely. This analysis would seem to suggest that Hopkins’s aspirations with Sprung Rhythm were perhaps more congenial to Eliot’s poetics than Eliot himself suspected or allowed.

❦

In Chapter 3 we noted that it is not uncommon for European verse traditions to draw upon (at least) two structurally different verse designs and to assign specific roles to each. We also noted that, usually, the verse design with the longer line length is associated with more formal genres such as epic and tragedy, whereas the verse design with the shorter line length is associated with more demotic and lyric genres. In English, the long-line tradition is, of course, represented first and foremost by the iambic pentameter, whereas the short-line tradition is represented by the four-beat line. In Chapter 2 we outlined how Derek Attridge in his Rhythms of English Poetry (1982) has demonstrated the great versatility of the four-beat measure as the underlying rhythm of a host of English verse forms, ranging from the limerick to the iambic tetrameter. We also noted that this formal versatility of the four-beat line is in glaring contrast to the very limited metrical possibilities

72 Peter L. Groves, ‘“Opening” the Pentameter: Hopkins’s Metrical Experimentation’, in The Hopkins Quarterly, 38, 1- 2 (Winter/Spring, 2011), 93-110: 94. 190

offered by the five-beat line, which in English is almost exclusively represented by the iambic pentameter. However, the fundamental distinction between long-line and short-line poetry that is observed in metrical verse is equally if not more relevant to the province of free verse. Up until now, we have been focusing exclusively on free verse based on relatively long lines, but there is another tradition of free verse which is based on significantly shorter measures. In the context of English poetry this tradition may be said to go back to Milton’s choruses in Samson Agonistes, the structure of which was cited by Frank Sayers as a source of inspiration for the loosely structured choruses in his tragedy Moina (1790).73 Sayers’s work in turn served as a model for Robert Southey’s Madoc (1805) and Thalaba the Destroyer (1801), but none of these works bear comparison with the striking effects that the modernists were to achieve with their short-line free verse. A more compelling candidate for comparison with modernist free verse based on short measures is , whose idiosyncratic Black Riders and Other Lines (1895) contains just under sixty poems, of which this – Number III – is representative:74

In the desert I saw a creature, naked, bestial, who, squatting upon the ground, Held his heart in his hands, And ate of it. I said, “Is it good, friend?” “It is bitter—bitter,” he answered; “But I like it “Because it is bitter, “And because it is my heart.”75

In this poem, Crane succeeds in avoiding enjambment altogether, in spite of the relative shortness of his lines. Line 2 may be said to scan as a somewhat awkward iambic pentameter – ‘I sáw a créature, náked, béstiál’ – but the shortness and irregularity of the remaining lines suggest that this

73 Kirby-Smith (1996): 112-113. 74 An even earlier candidate would be William Ernest Henley, whose collection In Hospital (written in the 1870s) contains several poems without metre. 75 Cited from Stephen Crane, Poems and Literary Remains, edited by Fredson Bowers (Charlottesville, VA: University Press of Virginia, 1975). 191

is merely coincidental. In fact, the one feature which saves this poem from the banality to which it all but succumbs, is the absence of metre as that which controls the poem’s movement. Had the poem insisted on iambic patterning throughout, the crucial final line would have lost the fine ambiguity that it retains in free verse. As the final line stands, we are unable to determine conclusively whether it should be interpreted as ‘And because it is my heart’ or as ‘And because it is my heart’. If, on the other hand, this line had been the conclusion of a metrical iambic poem, we would have been forced into accepting the former interpretation as the more probable, on account of the fact that it may be scanned as an acephalous iambic tetrameter: ‘Ánd becáuse it ís my héart’. However, the absence of metre in the poem as a whole obliges us to wonder. This style of short-line free verse with a minimum of enjambment found its most iconic practitioner in Hilda Doolittle, who in 1958 recalled how it was Pound who was responsible for shortening her name as well as her verse lines: ‘“But Dryad ... this is poetry’, Pound reputedly told Doolittle in 1912 upon having read a poem of hers in the tea room of the British Museum, ‘Cut this out, shorten this line. ... Will this do?” And he scrawled “H.D. Imagiste” at the bottom of the page.’76 The poem that so excited Pound’s editorial disposition – and inspired the name of – was ‘Hermes of the Ways’, the opening lines of which demonstrate the difficulty of avoiding enjambment entirely in short-line free verse:

The hard sand breaks, And the grains of it Are clear as wine.

Far off over the leagues of it, The wind, Playing on the wide shore, Piles little ridges, And the great waves Break over it.

76 H.D. [Hilda Doolittle], End to Torment: A Memoir of Ezra Pound (New York: New Directions Books, 1979): 18. On having received the manuscript for this memoir, Pound commented: ‘Torment title excellent, but optimistic’; cited in H.D. (1979): xi. 192

In this excerpt, there are two instances of clausal S/V enjambment (between lines 2-3 and 8-9), but in both instances the subject is a fairly complex noun phrase whose boundaries coincide with the line breaks in such a manner that the disruptive effect of enjambment is minimal. It is, however, possible to find examples in H.D.’s early work of poems, in which enjambment is completely absent. The most commonly cited is probably ‘Oread’ –

Whirl up, sea— whirl your pointed pines, splash your great pines on our rocks, hurl your green over us, cover us with your pools of fir.

– but the most extreme example is ‘The Pool’:

Are you alive? I touch you. You quiver like a sea-fish. I cover you with my net. What are you – banded one?77

The question with which this last poem concludes may also be directed at the poem itself: what are you – lineated one? By bringing lineation into perfect alignment with sentence structure, ‘The Pool’ is not even verse by Gasparov’s definition, and perhaps this is why the poem, from a formal point of view, represents a cul-de-sac in the history of English poetry. ‘Oread’, too, with its total avoidance of both clausal and phrasal enjambment is too prosodically sterile to gain currency as a model for short-line verse on a larger scale.78 Both poems are thus exceptions to the general rule of short-line

77 All three poems by H.D. are cited from Hilda Doolittle, Collected Poems of H.D. (New York: Boni and Liveright, 1925). The selfsame poems that I have cited here are mentioned by Robert Duncan to exemplify ‘poems of the type approved by Pound and still favoured by anthologies at large’, in Robert Duncan, The H.D. Book, edited and with an Introduction by Michael Boughn & Victor Coleman (Berkeley & Los Angeles, CA: University of California Press, 2011): 397. 78 Robert Bridges’s comments on this bear repetition here: ‘The identification of the line unit with the grammatical unit must limit the varieties of line-structure. ... The grammatical forms of sentences in English are few, and must repeat themselves again and again; and each form has its proper and natural inflection of voice which, however overlaid, will 193

free verse in English, which is to allow for enjambment in order to counteract the prosodic monotony that permeates much of the earliest poetry in the imagist vein. H.D.’s use of enjambment is – as might be inferred from the excerpt from ‘Hermes of the Ways’ cited above – characterized by great restraint and unobtrusiveness. Clausal enjambment is permitted to occur sporadically, and so, to an even lesser extent, is phrasal enjambment, usually between a head word and a postmodifier. But other poets were less willing to let phrase structure determine their lineation to this degree, most prominent among them William Carlos Williams. His ‘Portrait of a Lady’, first published in 1920, may serve as an example of the sort of short-line free verse that works on the principle of consistently subverting the prosodic unity of the grammatical phrase:

Your thighs are appletrees whose blossoms touch the sky. Which sky? The sky where Watteau hung a lady’s slipper. Your knees are a southern breeze—or a gust of snow. Agh! what sort of man was Fragonard? —as if that answered anything. Ah, yes—below the knees, since the tune drops that way, it is one of those white summer days, the tall grass of your ankles flickers upon the shore— Which shore?— the sand clings to my lips—

impose its typical intonation on the sentence. Now if the grammatical forms are made coincident with the lines of the verse, they must impose the recurrence of their similar intonations upon the line. ... [M]onotony of this kind is often agreeable in itself, and sought for its special effect. None the less, one of the difficulties in writing good verse of any kind is to escape from the tyranny of these recurrent speech forms, and the restriction imposed by the rules of free verse must make that difficulty immeasurably greater.’ Robert Bridges, ‘A Paper on Free Verse’, in The North American Review, Vol. 216, No. 804 (Nov., 1922), 647-658: 656. 194

Which shore? Agh, petals maybe. How should I know? Which shore? Which shore? I said petals from an appletree.79

In addition to conventional clausal enjambment between both subject and verb (‘Your knees / are a southern breeze’) and verb and object (‘as if that answered / anything’), this short poem makes use of an extensive catalogue of stark phrasal enjambments, almost all of which even the late Shakespeare would have been hesitant to use: e,g, ‘a lady’s / slipper’, or ‘what / sort of man’. Enjambment begins to occur as soon as the rather high-flown lyric voice that opens the poem is interrupted by an intruding second voice in line 3, at which same point we also discover that the smooth iambic trimeter of the poem’s two opening lines was but a ruse to lure us into the pastoral mood. The opening of Williams’s poem may be said to exemplify the ‘ornamental’ use of metrical form that characterizes so much modernist free verse and of which Eliot is the great master and instigator in English. In Williams’s case, however, such citations of metricality usually serve the purpose of gently mocking the Eurocentric tradition, which he considered a hindrance to his own aspiration of giving voice to a distinctly American poetic idiom. Ideologically there is thus an obvious kinship between Whitman and Williams, but from a formal point of view they are as far removed from one another as can be imagined within the parameters of this investigation. Whereas Whitman’s verse combines the long-line tradition with syntactically motivated line breaks, Williams – at least in his most characteristic type of verse – employs continuous enjambment in combination with the short-line tradition. Both H.D. and Williams were featured in Pound’s seminal anthology Des Imagistes (1914), a volume whose quaint French title cannot conceal the fact that the most important influence for the most prominently featured poets was Greece. Thus, on the verso of the title page of Des Imagistes appears in Greek lettering a short quotation from the Eπιτάφιος Bίωνος with an English translation underneath it, and a glance at the contents pages reveals a catalogue of poems with titles such as ‘Choricos’, ‘Lesbia’ (Richard Aldington), ‘Priapus’, ‘Hermonax’ (H.D.) and ‘Δώρια’ (Pound). In view of this, Poetry’s early endorsement of the imagists as ‘a group of ardent Hellenists who are

79 Cited from William Carlos Williams, Collected Poems I: 1909-1939, edited by Walton Litz & Christopher MacGowan (Manchester: Carcanet Press, 2000). 195

pursuing interesting experiments in vers libre’80 seems fair, and resonates well with Pound’s own description of H.D.’s early poetry as ‘modern, for it is in the laconic speech of the Imagistes ... Objective—no slither; direct—no excessive use of adjectives, no metaphors that won’t permit examination. It’s straight talk, straight as the Greek!’81 H.D.’s predilection for Greek poetry is particularly well attested throughout her work and has received much critical attention.82 However, it is important to remember that, unlike earlier upsurges of Hellenism in English literature, the modernist brand of Hellenism was not the product of systematized philological training on the part of its practitioners. In this respect, the proponents of modernist Hellenism find an obvious predecessor in , whose ‘On First Looking into Chapman’s Homer’ is an early testimony of encountering the glory that was Greece primarily through translation. Somewhat surprisingly, Keats is singled out by William Carlos Williams as an inspiration for his own early poetry in a series of conversations with Edith Heal, during which Williams also comments on his own relationship with the Classical languages:

I had only a partial year of Latin at Horace Mann when the course was discontinued. No Greek. ... I remember casting envious eyes at a boy named Wise who was taking Greek, but I didn’t have the initiative to change my curriculum. Yet I was conscious that I was losing something that I might have had.83

In the same conversation, Williams also recalls that, at the time when he met her, ‘H.D. was studying Greek by herself’,84 an activity she would pursue for most of her life. Williams’s recollections are reflective of a decidedly new situation at the beginning of the twentieth century, which T.S. Eliot pinpointed quite accurately when he observed in 1920 that ‘The Classics have, during the latter part of the nineteenth century and up to the present moment, lost their place as a pillar of the social and political system’.85 In spite of this, the cultural prestige of classical learning remained high, and the aspiration of mastering Greek and Latin was even extended to include women. Thus, when Clarissa Dalloway exclaims ‘I’d give ten years of my life

80 ‘Notes and Announcements’, Poetry, Vol. 1, No. 2 (1912): 65. 81 Ezra Pound, Selected Letters 1907-1941, edited by D.D. Paige (New York: New Directions, 1971): 11. 82 See Eileen Gregory, H.D. and Hellenism: Classic Lines (Cambridge: Cambridge University Press, 1997) and Diana Collecott, H.D. & Sapphic Modernism (Cambridge: Cambridge University Press, 1999). 83 William Carlos Williams, I Wanted to Write a Poem: The Autobiography of the Works of a Poet, edited by Edith Heal [1958] (New York: New Directions: 1978): 6-7. 84 Williams [1958]: 7. 85 T.S. Eliot, ‘ and Professor Murray’. in Eliot (1920), 71-77: 72. 196

to know Greek,’ in Virginia Woolf’s The Voyage Out (1915),86 she is giving voice to an ambition that was not uncommon in the early decades of the twentieth century.87 1912 thus saw the publication of the first volumes of the Loeb Classical Library, the significance of which was summed up five years later by Woolf in the Times Literary Supplement:

The Loeb Library, with its Greek or Latin on one side of the page and its English on the other, came as a gift of freedom.... The existence of the amateur was recognised by the publication of this Library, and to a great extent made respectable.... The difficulty of Greek is not sufficiently dwelt upon, chiefly perhaps because the sirens who lure us to these perilous waters are generally scholars [who] have forgotten…what those difficulties are. But for the ordinary amateur they are very real and very great; and we shall do well to recognise the fact and to make up our minds that we shall never be independent of our Loeb.88

The distinction Woolf draws here between the professional scholar and the ‘ordinary amateur’, for whom the acquisition of Greek poses ‘very real and very great’ difficulties, is important. All of the ‘ardent Hellenists’ of imagism were, of course, amateur Hellenists, who very much depended on the availability of scholarly translations of Greek literature as a reference for their own dabbling with the language. Thus, in his memoir of Pound, Michael Reck notes that Pound read Greek mainly with the help of the Loeb Classical Library,89 an observation that lends credit to Pound’s own confession that he was ‘Too god damn iggurunt of Greek’ to translate Homer.90 In spite of such occasional admissions of their own philological limitations, the modernist poets of the early twentieth century were not above criticizing the available stock of scholarly translations, especially those in verse. Eliot’s review of Gilbert Murray’s translation of Euripides is particularly unkind: ‘Professor Murray has simply interposed between Euripides and ourselves a barrier more impenetrable than the . ... And it is inconceivable that anyone with a genuine feeling for the sound of Greek verse should deliberately elect the William Morris couplet,

86 Virginia Woolf, The Voyage Out (London: Hogarth Press, 1971): 46. 87 On Woolf’s own relationship to the Greek language, see Emily Dalgarno, Virginia Woolf and the Visible World (Cambridge: Cambridge University Press: 2001): 33-66. 88 Virginia Woolf, The Essays of Virginia Woolf, edited by Andrew McNeillie, Vol. 2: 1912-1918 (London & New York: Harcourt Brace Jovanovich, 1987): 114. 89 See Michael Reck, Ezra Pound, A Close-Up (New York: McGraw-Hill, 1968): 118-120. On Pound’s relationship with the classical languages, see Peter Liebregts, ‘The Classics’, in Ira B. Nadel (ed.), Ezra Pound in Context (Cambridge: Cambridge University Press, 2010), 171-180. 90 Pound (1971): 274. 197

the Swinburne lyric, as a just equivalent’.91 Eliot does not advocate directly the use of free verse to capture the sound of Greek verse, but towards the end of the review he opines, condescendingly, that ‘The choruses from Euripides by H. D. are, allowing for errors and even occasional omissions of difficult passages, much nearer to both Greek and English than Mr. Murray’s’.92 H.D.’s own view of her work with Greek translations is perhaps best captured in an episode of Bid Me to Live (1960), in which the protagonist observes:

Anyone can translate the meaning of the word. She wanted the shape, the feel of it, the character of it, as if it had been freshly minted. She felt that the old manner of approach was as toward hoarded treasures, but that treasure had passed through too many hands, too carefully assessed by grammarians.93

Grammarians, and by extension metrists, were also singled out by Pound as having done more harm than good to the understanding and appreciation of Greek verse:

When the Greek dramatists developed or proceeded from anterior Greek prosody, they arrived at forms which are to all extents ‘free’, though a superstructure of nomenclature has been gummed on to them by analysers whom neither Aeschylus nor Euripides would ever have bothered to read. These nomenclatures was probably invented by people who had never LISTENED to verse, and who probably wouldn’t have been able to distinguish Dante’s movement from Milton’s had they heard it read out aloud.94

What is needed, according to Pound, in order to bring alive an alien and ancient poetic idiom in translation is not magisterial scholarship, but rather a good ear on the part of the poet. And, of course, a certain familiarity with the poetic tradition of the language in question, as Pound lectured Iris Barry in a letter from 1916:

Really one DON’T need to know a language. One NEEDS, damn well needs, to know the few hundred words in the few really good poems that any language has in it. It is better to

91 Eliot (1920): 75. 92 Eliot (1920): 75. 93 H.D., Bid Me to Live (New York: Dial Press, 1983): 163. 94 Pound (1951): 204. 198

know [’s] POIKILOTHRON by heart than to be able to read Thucydides without trouble.95

There was among the poets featured in Des Imagistes at least one who would probably disagree with Pound on these matters, and that poet was F.S. Flint. With a command of more than ten languages, Flint is, perhaps, more remarkable for his linguistic aptitude than for his poetry; even so, his significance for the development of English free verse should not be underestimated. It was Flint, not Eliot, who introduced Pound to modern French poetry, a topic on which Flint was generally regarded as the greatest authority in London around 1910.96 Flint’s public promotion of French vers libre began in 1908 when he was made poetry reviewer for The New Age under the editorship of A.R. Orage. In this capacity Flint attracted the attention of T.E. Hulme and became a regular attendant at Hulme’s Thursday evening meetings at the Tour Eiffel, a small restaurant in Soho. When Pound published The Complete Poetical Works of T.E. Hulme (all five of them) as an appendix to Ripostes (1912), he referred to those Thursday gatherings as ‘the forgotten school of 1909’ of which ‘Les Imagistes [were] the descendants’.97 Pound himself was not in any strict sense a member of the Tour Eiffel group, but the group’s contribution must nonetheless be acknowledged in any serious attempt to trace the historical roots of free verse in English. To write off the whole school, with J.J Wilhelm, as ‘a few colorful songbirds’98 conceals the fact that it is in this group that we find the earliest collective effort to revitalize English verse along distinctly modernist lines. In addition to promoting French vers libre as a model for composing verse in English, Flint was also fascinated with the possibilities offered by Eastern poetry, particularly the Japanese form. Japanese was not among the languages that Flint mastered,99 but in his first signed article for The New Age, published July 11, 1908, Flint took under review a new volume of Japanese poems and lamented that they had been rendered into ‘heavy English rhymed quatrain[s]’ by the translators, Shōtarō Kimura and Charlotte M. A. Peake. Instead, Flint would have preferred ‘that the

95 Pound (1971): 93. 96 Richard Aldington, in 1914, offered this rather charming portrait of Flint: ‘Whenever I meet Mr Flint, I say to him, “Well, I’ve read the latest thing from Paris you told me to read the other day,” and he says, “My dear child, did I tell you to read that old-fashioned book? However, I am afraid I can’t stop now, because I have six new Fantaisiste authors, two volumes of Apollinaire and thirty-two other books by representatives of sixteen different schools to review by Saturday.”’ (The Egoist, No. 1, June 15, 1914: 221). Comparatively little is known about Flint’s background, but it was one that was very different from that of the other imagists, in that he grew up in extreme poverty and left school at the age of thirteen; see Helen Carr, The Verse Revolutionaries: Ezra Pound, H.D. and The Imagists (London: Jonathan Cape, 2009): 143-151. 97 Pound (2001): 266. 98 J.J. Wilhelm, Ezra Pound in London and Paris: 1908-1925 (Pennsylvania State University Press, 1990): 34. 99 J. B. Harmer, Victory in Limbo: A History of Imagism, 1908-1917 (New York: St. Martin’s, 1975): 127. 199

poems in this book had been translated into little dropping rhythms, unrhymed’, and he gave as an example of such a ‘dropping rhythm’ his own translation of Arakida Moritake’s famous haiku on the blossom and the butterfly:

A fallen petal Flies back to its branch: Ah! a butterfly!100

Of this poem and another haiku translation by Flint in the same article, Helen Carr has suggested that they ‘could lay claim to be the first published imagist poems’.101 She also makes the point that Flint’s translation is in fact a translation of a translation, in that the source text used by Flint was a French translation made by Paul-Louis Couchoud, published after a visit to Japan during the Russo- Japanese war. Couchoud’s translation of Moritake’s haiku, first printed in 1906 in Les Lettres, reads:

Un pétale tombé Remonte à sa branche: Ah! c’est un papillon!102

One can hardly miss the degree of faithfulness with which Flint follows his original: J.B. Harmer has pointed out that the faithfulness extends even to the level of punctuation, but at the same time Harmer overstates his case when he refers to Flint’s version of Moritake’s poem as a ‘literal rendering’ of Couchoud.103 Even without quibbling over the problematic word ‘literal’ used in relation to poetry, we should notice that Flint has chosen to translate the last line of the poem, not as an independent clause, but rather as a simple noun phrase. When Pound, in September 1914, tried his hand at translating the selfsame poem, he followed Flint rather than Couchoud in this:

100 F.S. Flint, ‘Book of the Week’, in The New Age, Vol. 3, No. 11 (1908), 212-213. 101 Helen Carr, ‘Imagism and Empire, in Howard H. Booth & Nigel Rigby (eds), Modernism and Empire (Manchester: Manchester University Press, 2000), 64-92: 70. 102 Cited from Harmer (1975): 130. 103 Harmer (1975): 129. 200

The fallen blossom flies back to its branch: A butterfly. 104

Even more significantly, Pound arranges his ‘hokku’ in just two lines instead of three, in all likelihood to facilitate the understanding of his notion of the ‘“one image poem”’, consisting of ‘one idea set on top of another’,105 and of which Pound’s own ‘In a Station of the Metro’ is the prime exemplar:

The apparition of these faces in the crowd : Petals on a wet, black bough .106

Whether or not Pound in fact first learned about from Flint and the other members of the Tour Eiffel group is for literary biographers to quarrel about. That there is a quarrel concerning this issue stems in no small part from a skirmish between Pound and Flint which arose in the wake of the publication of Flint’s ‘History of Imagism’ in The Egoist in May 1915. Flint’s rather sulky sketch of the events which led to the first imagist anthology presents imagism as little more than a late bud of the Tour Eiffel group’s endeavours and allocates to Pound only the role of usurper and publicist. Enraged, but ‘[a]fter deliberate consideration’, Pound dismissed the article as ‘BULLSHIT’,107 complaining in particular that Ford Madox Hueffer had been left out of the account and that Pound himself had not been named the main theorist of the movement. Flint, however, held his ground. In fact, the original draft to his historical sketch of imagism was even more dismissive of Pound’s importance for the development of modern English poetry than was the version which eventually saw print: in the original version Pound was presented as little more than a common thief of ideas.108 However, when Flint, in his history of imagism, lists first among the means by which the Tour Eiffel poets had sought to reform English poetry ‘pure vers libre’ and ‘Japanese tanka and haikai’,109 he neglects to mention that these techniques were by no means unknown in Britain at the

104 Cited from Ezra Pound, Gaudier-Brzeska: A Memoir (London: New Directions, 1970): 88. 105 Pound (1970): 89. 106 Cited from Poetry, Vol. 2, No. 1 (1913): 12. The following year, Richard Aldington published this poem in The Egoist: ‘The apparition of these poems in a crowd : / White faces in a black dead faint.’ (The Egoist, Vol. 1, No. 2 (1914): 36). 107 Unpublished letter to F.S. Flint, cited in Carr (2009): 765. 108 Carr (2009): 766. 109 Cited in Pondrom (1975): 301. 201

time. Earlier in this chapter we saw that a decade prior to the inaugural meeting at the Tour Eiffel the first edition of Arthur Symons’s The Symbolist Movement had already made use of the term vers libre without explaining it. This indicates that the concept, if not the practice, of vers libre had already been absorbed into the general nomenclature of English poetics before the turn of the twentieth century, and that Flint, in The New Age and elsewhere, was advocating a principle which, presumably, was already well-known to his readers. As for the Tour Eiffel group’s predilection for Japanese poetics in general and the haiku in particular, they were preceded by Basil Hall Chamberlain’s The Classical Poetry of the Japanese, which was published as early as 1880 and had served as an inspiration for Couchoud’s translations of into French.110 Again, the cross- linguistic path of origin is as conspicuous as it is significant: Chamberlain’s English treatment of the Japanese haiku prompted Couchoud’s French translation of Moritake, which in turn underlied Flint’s translation back into English. So, neither French vers libre nor Japanese haiku technique owes its introduction into English poetics to the Tour Eiffel group: knowledge of both forms was well-established among British literati from before the turn of the twentieth century. What was genuinely new and innovative about the Tour Eiffel group’s use of these techniques was rather their collective and programmatic approach to these forms and the audacity with which they sought to adapt them into English. Perhaps audacity as a characterization is too vague in its connotations: much of what Flint and his peers accomplished, especially in relation to the haiku, may better be described as reckless tampering. After all, and for all of Flint’s linguistic versatility, the Tour Eiffel poets’ knowledge of the haiku was mediated exclusively through French, for which reason they appear to have disregarded the fact that the haiku, in its vernacular form, is a strictly metrical form. It is commonplace to describe the haiku as a syllabic form of poetry consisting of seventeen syllables in three lines of five, seven and five syllables respectively.111 However, such a description fails to take into account that Japanese metrics is determined by morae rather than by our household Indo- European concept of syllables. More significantly, it obscures the fact that, as far as the Japanese haiku goes, the Western concept of the line-unit is not employed as a metrical constituent: in Japanese, haikus are conventionally printed in a single vertical line.112 This latter fact explains why English translators felt free to render haikus as octosyllabic couplets (Chamberlain), rhymed

110 Harmer (1975): 129. 111 E.g. J.A. Cuddon, Penguin Dictionary of Literary Terms and Literary Theory, Third Edition (London: Penguin, 1991): 399. 112 Harmer (1975): 127; Cor van den Heuvel, The Haiku Anthology, 2nd edition (New York: Simon & Schuster, 1986): 11. 202

quatrains (Kimura and Peake), free verse tercets (Flint, following Couchoud) and free verse couplets (Pound), without offering any justification. Pound, in describing ‘In a Station of the Metro’ as a ‘hokku-like sentence’, actually comes closest to the Japanese mark insomuch as this description evades any reference to lineation as a constituent of the haiku. There is, of course, a certain irony in the fact that an Eastern non-lineated form of poetry came to exert such an influence on the development of free verse: the one form of Western poetry which must take refuge in lineation to justify its poeticity. In this chapter we have attempted to outline a rough taxonomy of free verse in English, based on the distinction between long-line and short-line verse, and the ways in which lineation either reinforces or disrupts syntax. Convenient as such distinctions may be in a context such as this, it should, however, be remembered that a significant portion of free verse in English evades vain attempts at easy categorizing. Throughout the history of free verse, poets have been combining long lines with short lines within the same poem, with very little regard for the plight of the taxonomist. Generally, such mixed modes of free verse tend to avoid enjambment, but it is certainly possible to find poems in which this rule of thumb is violated:

When the moon falls on a man’s blood white and slippery, as on the black water in a port shaking asunder, and flicking at his ribs—

then the noisy, dirty day-world exists no more, nor ever truly existed; but instead this wet white gleam twitches, and ebbs hitting, washing inwardly, silverily against his ribs on his soul that is dark ocean within him.113

This is the opening of D.H. Lawrence’s ‘Moon Memory’, and we conclude our overview of English free verse forms by citing it as a reminder that, as far as free verse is concerned, exceptions far outnumber rules.

113 Cited from D.H. Lawrence, Complete Poems, collected and edited with an Introduction and Notes by Vivian de Sola Pinto & Warren Roberts (London: Heinemann, 1964). 203

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

The connection between numbering and knowing, the ability to count or measure and the ability to grasp, comprehend, or control, runs very deep in Greek thought about human cognition. Already in Homer, the poet associates knowing with the ability to enumerate: the Muses give him their knowledge of the warring armies by imparting a catalogue of their numbers and divisions.

Martha C. Nussbaum, The Fragility of Goodness

As was hinted in the previous chapter, D.H. Lawrence fits uneasily in the canon of modernist English free verse. In addition to being the only poet who was featured both in Edward Marsh’s anthologies of Georgian Poetry (five volumes, 1912-1922) and in Amy Lowell’s series Some Imagist Poets (three volumes, 1915-1917), Lawrence is also unusual in having embraced at a very early stage free verse as a valuable mode of writing poetry that requires no justification on formal grounds:

Much has been written about free verse. But all that can be said, first and last, is that free verse is, or should be direct utterance from the instant, whole man. ... It is no use inventing fancy laws for free verse, no use drawing a melodic line which all the feet must toe. Free verse toes no melodic line, no matter what drill-sergeant. Whitman pruned away his clichés—perhaps his clichés of rhythm as well as of phrase. And this is about all we can do, deliberately, with free verse. We can get rid of the stereotyped movements and the old hackneyed associations of sound or sense. We can break down those artificial conduits and canals through which we do so love to force our utterance. We can break the stiff neck of habit. We can be in ourselves spontaneous and flexible as flame, we can see that utterance rushes out without artificial foam or artificial smoothness. But we cannot positively prescribe any motion, any rhythm. All the laws we invent or discover—it amounts to pretty 204

much the same—will fail to apply to free verse. They will only apply to some form of restricted, limited unfree verse.1

As commonplace as Lawrence’s observations on free verse may seem today, they are cited here in some length because they represent a rare exception to the modernists’ general preoccupation with demonstrating theoretically that freedom in versification is merely superficial. Eliot’s contention that successful verse should always be haunted by the ghost of some simple metre represents one strategy of avoiding the embarrassment of free verse. Another strategy is that of claiming for free verse an underlying set of principles which – though impossible to explicate – are every bit as rigid as those determining traditional versification. William Carlos Williams’s obsession with carving out a place for his concept of the variable foot is a case in point;2 another is Amy Lowell’s experiments with establishing ‘cadence’ as an objective constituent of free verse by reading aloud poems by herself and H.D. to the phonetician William Patterson’s ‘sound-photographing machine’.3 This chapter investigates some of the most prevalent modernist justifications for abandoning traditional metrics and furthermore seeks to assess the validity of these claims by means of a corpus-based study of free verse in relation to, on the one hand, metrical verse and, on the other hand, novelistic prose.

❦

The modernists’ persistent attempts to explicate theoretically the constituent principles on which free verse operates were initially of a programmatic and practical nature: vague statements of intent aimed at establishing schools and isms for the purpose of attracting publicity. The best known – and surely the most influential – example of this is Pound’s ‘A Few Don’ts by an Imagiste’ published in Poetry in March 1913 alongside Flint’s three-point summation of the imagist principles:

1. Direct treatment of the “thing,” whether subjective or objective. 2. To use absolutely no word that [does] not contribute to the presentation.

1 ‘Introduction’ to New Poems (1920), in D.H. Lawrence, Selected Literary Criticism, edited by Anthony Beal (London: Heinemann, 1956), 84-89: 87-88. 2 See Stephen Cushman, William Carlos Williams and the Meanings of Measure (New Haven & London: Yale University Press, 1985) and Beyers (2001): 179-221. 3 See Michael Golston, Rhythm and Race in Modernist Poetry and Science (New York: Columbia University Press, 2008): 53-54. 205

3. As regarding rhythm: to compose in sequence of the musical phrase, not in sequence of a metronome.4

However, in response to the growing criticism of free verse by commentators from outside of the modernist fractions, theoretical treatments of the formal properties of free verse soon took on a more polemical and apologetic character. Sadly, Richard Aldington’s remarks in The Egoist in September 1914 may be considered as being fairly representative of this development:

What, then, is the difference between the new free verse and the old rhymed, accented verse? Not, obviously, the commonplace idiotic remark of the journalist that free verse is merely prose cut into different lengths. The man who says that has no ear for poetry; he is unable to distinguish between the free verse written by an artist and its imitation by an amateur.5

The tactic Aldington employs is the familiar one of denigrating an opponent’s competence in order to conceal a weak counter-argument: ‘free verse is not prose’, Aldington proceeds, since ‘The cadence [of free verse] is more rapid and more marked, its “rhythmic constant,” shorter, and more regular. It is—or it ought to be—about five times more concentrated than the best prose and about six times more emotional.’ Keeping in mind that this was an argument intended to be a defence of free verse, it is not surprising that a number of key modernist poets – most notably Eliot and Pound – eventually chose to distance themselves from the free verse controversy. One of the more curious episodes of the free verse controversy was played out in 1917, and it makes for an apt illustration of the problems with Aldington’s contention that only journalists with no ear for poetry could ever mistake ‘free verse written by an artist’ for an imitation. That year the New York-based magazine Others devoted their entire January issue to the promotion of the so- called Spectric School, whose artistic manifesto bore more than a passing resemblance to some of the imagist tenets. The prime exponents of this strikingly new movement were Anne Knish and Emanuel Morgan, and both of them favoured free verse as their principal medium of composition. Edgar Lee Masters – the revered author of the Spoon River Anthology (1915) – praised the spectrists for ‘getting to the core of things’, and William Carlos Williams – a regular contributor to Others – admitted to having been ‘quite taken in’ by their poems. It soon turned out, though, that the Spectric

4 F.S. Flint, ‘Imagisme’, in Poetry, Vol. 1, No. 6 (1913), 198-200: 199. 5 Richard Aldington, ‘Free Verse in England’, in The Egoist, Vol. 1, No. 18 (1914), 351-352. 206

School was little more than an elaborate literary hoax aimed at ridiculing modernist free verse, and that Anne Knish was, in fact, a pseudonym of Arthur Davison Ficke, just as Emanuel Morgan was the alter ego of Witter Bynner.6 Both poets of a fairly conservative disposition, Ficke and Bynner had set out to expose once and for all the pretentious mannerism of certain new schools of poetry by composing a bulk of deliberately bad poems in free verse and attaching to them a sufficiently self- assertive manifesto. Given that they had no problem in finding a publisher, Ficke and Bynner may be said to have succeeded, but when the hoax was eventually revealed, in 1918, it did not cause as much of a scandal as one might think. Pound was not amused and claimed to have known all along, and Alice Corbin Henderson, the assistant editor of Poetry, responded to the exposé of Spectra by observing that ‘there is no reason why one should not have supposed Emanual [sic] Morgan and Anne Knish to be real persons, for the poems in Spectra are no worse and no better, and hardly any less intelligible, than much of the free verse which has been thrust upon us’.7 As far as literary hoaxes go, the Spectra hoax was quite successful, but its implications for the status of free verse in the early twentieth century should not be unduly overemphasized. As Henderson’s comment shows, there was, even among avowed proponents of free verse, a widespread awareness of the fact that much free verse was simply bad. Thus, when The Little Review in 1916 announced ‘A Vers Libre Prize Contest’ to be judged by William Carlos Williams, Eunice Tietjens and Helen Hoyt, the editor eventually ended up by declaring: ‘I know very little about prize contests, but I imagine that there has never been one in the history of poetry which could boast so many really bad poems’.8 Bearing this in mind, it could well be argued that the literary scene of the mid-teens was in no immediate need of intentionally bad poems to add to the problem, and that Pound’s grumpiness therefore was well in order. However, for those involved in concocting it, the Spectra hoax was to have an interesting coda, in that both Bynner and Ficke later came to regard some of their Spectra mock poems as belonging to their best work and even took to writing free verse in earnest.9 Thus, as much as exposing the shortcomings of certain species of modernist free verse, the Spectra hoax can also, in retrospect, be construed as a reminder that there is no necessary link between a weak (or even insincere) theoretical foundation and the quality of the poetry that may be produced on its basis.

6 For references, see Suzanne W. Churchill, ‘The Lying Game: Others and the Great Spectra Hoax of 1917’, in American Periodicals, Vol. 15, No. 1, Little Magazine and Modernism (2005), 23-41. 7 Alice Corbin Henderson, ‘Baiting the Public’, in Poetry, Vol. 12, No. 3 (1918): 169-171. 8 The Little Review, Vol. 3, No. 10 (1917). The prize contest – ‘possibly the first prize extended to free verse’ – was first announced in The Little Review, Vol. 3, No. 1 (1916), and the winners turned out to be H.D. and Maxwell Bodenheim. 9 Churchill (2005): 38. 207

It would not be difficult to demonstrate the lack of theoretical rigour that permeates most of the early modernists’ attempts to justify their novel modes of versification. The literary magazines of the time were – on both sides of the Atlantic – littered with highly charged debates on the problems and properties of what was then called the New Poetry, in particular the abandonment of traditional accentual-syllabic metrics. In general, the theories put forward in these debates in favour of free verse have not aged well, so there is little to be gained from cataloguing their individual shortcomings.10 Rather, we must seek to reach beyond polemics in order to engage with the more fundamental issues that the emergence of free verse was a response to. If it is an oversimplification to say that English modernist verse represented a conscious rejection of the literary mores of the nineteenth century, it is at least an oversimplification that we owe to the modernists themselves: ‘As for the nineteenth century, with all respect to its achievements, I think we shall look back upon it as a rather blurry, messy sort of a period, a rather sentimentalistic, mannerish sort of a period.’11 The words are Pound’s, but the sentiment they express was widely shared by his contemporaries. The problem, according to Aldington, was that ‘the majority of the poetry of the last century had nothing to do with life and very little to do with poetry. There was a plague of prettiness and a plague of pomposity and several other minor diseases—such as over-much suavity, the cult of decorated adjectives.’12 The hostility towards the preceding century and its poetry was, importantly, not restricted to those literary groupings that we have now come to think of as distinctively modernist. Similar sentiments served as a major impetus for Rupert Brooke’s and Edward Marsh’s joint venture to issue the first volume of Georgian Poetry in 1912.13 As different from the imagist anthologies as the Georgian Poetry volumes admittedly were, both enterprises aimed to provide a platform for publishing contemporary verse that did not conform to the standards of mainstream publishers. Their standards were still firmly rooted in the nineteenth century, not only in Britain, but also in America, as Pound points out in Patria Mia:

It is well known that in the year of grace 1870 Jehovah appeared to Messrs Harper and Co. and to the editors of ‘The Century’, ‘The Atlantic’, and certain others, and spake thus: ‘The style of 1870 is the final and divine revelation. Keep things always just as they are now’. ...

10 A far more fruitful approach is that employed in Beyers (2001), which work aims to examine ‘the enabling myths...that Modern writers used without judging these ideas according to whether or not they are true’ (3). 11 ‘A Retrospect’, in Ezra Pound, Pavannes and Divisions (New York: Alfred A. Knopf, 1918): 95-111: 106. 12 Aldington, reviewing Hueffer in The Egoist, Vol. 1, No. 13 (1914). 13 On this, see Myron Simon, ‘The Georgian Poetic’, in The Bulletin of the Midwest Modern Language Association, Vol. 2, No. 1 (1969): 121-135. See also Pound’s letter to Harriet Monroe of 22 October 1913 208

And if you do not believe me, open a number of ‘Harpers’ for 1888 and one for 1908. And I defy you to find any difference, save on the page where the date is.14

That Pound has chosen the year 1908 as a reference date may well reflect the fact that 1908 was also the year in which Pound left America for Europe, and also the year in which his first collection of poetry, A Lume Spento, was published. After two short interims in Gibraltar and Venice, Pound arrived in London in August 1908, where he soon established an impressive network of literary contacts through which he finally got to meet Yeats. ‘London, deah old Lundon, is the place for poesy’, Pound complacently told William Carlos Williams in a letter of February 1909.15 If London indeed was the place for poesy around 1908 (and not all would agree that it was), Vienna was surely the place for music. That year, on the evening of December 21, Vienna witnessed the first public performance of Arnold Schoenberg’s String Quartet # 2, the fourth movement of which caused considerable outrage on account of the fact that it was not composed in a single fixed key.16 One year earlier, a comparable revolution may be said to have taken place within the pictorial arts when Pablo Picasso discarded the convention of maintaining a single fixed perspective in his Les Demoiselles d’Avignon. The pantonality of Schoenberg’s compositional technique (Schoenberg loathed the term atonality) and the use of multiple perspective characteristic of cubism are frequently cited as stock examples of distinctly modernist developments within the arts. And it is admittedly tempting to suggest that the rejection of fixed perspective in painting and of fixed keys in music finds its most obvious poetic analogy in the modernists’ rejection of fixed metres. However, it would perhaps be more accurate to say that modernist poetry in English first and foremost represents a rejection of a single fixed linguistic register, and that the abandonment of metre was a gambit to facilitate this. Once we accept that the modernist rejection of metre was, at least to some extent, contingent upon the rejection of fixed linguistic registers, it becomes easier to explain how free verse could be appropriated by the modernists to serve two almost opposite ideological agendas. On the one hand, there is Eliot’s and Pound’s shared agenda of absorbing and reinterpreting the classical heritage and to make it new. The previous chapter was devoted to demonstrating the extent to which translation, imitation and adaptation of foreign verse designs contributed to the development of free verse, based on the argument that traditional English metrics was not congenial with such foreign forms.

14 Ezra Pound, Patria Mia and the Treatise on Harmony (London: Peter Owen, 1962): 27. 15 Pound (1971): 7. 16 On the reception of Schoenberg’s String Quartet # 2, see Everdell (1997): 265-282. 209

On the other hand, there is Williams’s agenda of breaking with the classical heritage in favour of the local and the demotic, in search of a distinctly American kind of poetry: for this purpose, too, traditional English metrics was seen as a hindrance. Among those who shared Williams’s ambition, we find Mina Loy who, upon praising Pound as the great instigator of the modern movement, stresses Pound’s American roots and then proceeds to explain:

It was inevitable that the renaissance of poetry should proceed out of America, here latterly a thousand languages have been born, and each one, for purposes of communication at least, English—English enriched and variegated with the grammatical structure and voice- inflection of many races, in novel alloy with the fundamental time-is-money idiom of the United States, discovered by the newspaper cartoonists. This composite language is a very living language, it grows as you speak. ... And on the baser avenues of Manhattan every voice swings to the triple rhythm of its race, its citizenship and its personality. Out of the welter of this unclassifiable speech, while professors of Harvard and Oxford labored to preserve “God’s English,” the muse of modern literature arose, and her tongue had been loosened in the melting-pot.17

Loy’s argument that , on account of its composite nature, is incapable of being confined to the metrical patterns of ‘God’s English’ and therefore requires a break away from such patterns, is one with which Williams would have agreed. A similar argument was put forward by Amy Lowell, although she couches her argument in a more nationalistically tinted jargon:

Our environment is no longer exclusively English. Our immense immigration is at last beginning to be felt. We of the pure Anglo-Saxon stock are constantly coming into contact with people of other nationalities, and consciously or unconsciously we are being modified by them. We may not realise it, but slowly before our eyes, the American race is being born. And one of the evidences of it is that we are beginning to hew new pathways for ourselves in this most intimate thing, Poetry, and to free ourselves from the tutelage of another nation ... the New Poetry is blazing a trail towards Nationality far more intense than any settlement houses and waving the American flag in schools can ever achieve. I might say with perfect

17 Mina Loy, ‘Modern Poetry’ (1925), in Melissa Kwasny (ed.), Toward the Open Field (Middleton, CT: Wesleyan University Press, 2004), 269-273: 271. 210

truth that the most national things we have are skyscrapers, ice water, and the New Poetry.18

If the internationalist approach of Pound and Eliot can be accused of Eurocentrism, then comments such as those by Loy and Lowell are equally vulnerable to accusations of American exceptionalism.19 When Harriet Monroe in 1912 asked Pound to be a foreign correspondent for Poetry, she inadvertently turned her newly started magazine into a battlefield between these two ideological strands of the free verse movement. ‘Are you for or for poetry?’ Pound asked Monroe in the very letter in which he accepted the position; ‘The latter is more important,’ he proceeded, because ‘The glory of any nation is to produce art that can be exported without disgrace to its origin’.20 Given that Pound was hired by Monroe, we must assume that she initially was ‘for poetry’, un-modified by nationalist concerns, but over time Poetry increasingly came to promote the contention that free verse is a distinctly American art form. This process culminated, absurdly, in the February 1917 issue of Poetry, which was devoted to Native American poetry, or rather ‘poems from American-Indian motives’ by white American writers.21 Poems based on such motives, it turned out, bore a striking resemblance to imagist poems: ‘Suspicion arises definitely that the Red Man and his children committed direct plagiarisms on the modern imagists and vorticists’, Carl Sandburg noted, tongue-in-cheek.22 Tellingly, Pound’s sole contribution to the ‘Native American’ issue of Poetry was an obituary of Emile Verhaeren, in which he warned against attempts to pigeonhole Verhaeren exclusively as a Belgian poet. It was also in his correspondence with Harriet Monroe that Pound, in 1915, stated that ‘Poetry must be as well written as prose. Its language must be a fine language, departing in no way from speech save by a heightened intensity (i.e. simplicity).’23 In Missing Measures (1990), Timothy Steele devotes an entire chapter to analysing the ways in which novelistic prose came to serve as a model for the modernists’ experiments with free verse, following the ascendance of the novel during the nineteenth century.24 Steele emphasizes in particular the influence of Ford Madox Hueffer / Ford in promoting the clarity and alleged naturalness of prose as the yardstick by which

18 Cited in Carr (2009): 809. 19 As for the references to race in both Loy’s and Lowell’s comments, and for the ways in which race was used more generally in modernist accounts of rhythm and versification, see Golston (2008) and Anita Patterson, Race, and Transnational Modernisms (Cambridge: Cambridge University Press, 2008). 20 Pound (1971): 9. 21 Harriet Monroe, ‘Aboriginal Poetry’, in Poetry, Vol. 9, No. 5 (1917), 251-254: 251. 22 Poetry, Vol. 9, No. 5 (1917): 255. 23 Pound (1971): 48. 24 See Steele (1990), 69-108. 211

works in verse must be measured. The significance of Ford’s contribution to the development of modernist verse was certainly recognized by Pound: from the previous chapter it will be recalled that Pound was cross with F.S. Flint principally because Ford had been left out of Flint’s account of the history of imagism. In addition to its clarity, prose was also commended by Ford for its unfettered capacity to record contemporary everyday life, i.e. all the things that nineteenth-century verse had failed to capture. Thus, in much the same way that traditional English metre was presented as a hindrance to capturing the nuances of Greek or Japanese verse by some modernists, or the compositeness of American English by others, Ford presents metre as a hindrance to engaging with modernity. Among the prose writers most consistently recommended by Ford, Steele lists Henry James, whose review of Whitman’s Drum-Taps was quoted in the previous chapter. In that review, James characterizes Whitman’s verse as ‘arrant prose’, and then asks: ‘But what if, in form, it is prose? ... Very good poetry has come out of prose before this. To this we would reply that it must have gone into it. Prose, in order to be good poetry, must first be good prose’.25 The striking parallel between this statement and the shared maxim of Ford and Pound that ‘Poetry must be as well written as prose’ has been pointed out by Alan Holder, who also notes that James’s review had been reprinted as recently as 1908.26 In 1909, Eliot made his earliest published reference to James; it occurs in a review of James G. Huneker, of whom Eliot says that he possesses a style which ‘shares with that of Mr. Henry James ... what I should call a conversational quality; not conversational in admitting the slip-shod and maladroit, or a meagre vocabulary, but by a certain informality, abandoning all the ordinary rhetorical hoaxes for securing attention’.27 These comments on James are, of course, exactly contemporaneous with Eliot’s ‘study of Laforgue together with the later Elizabethan drama’, and the influence of Laforgue shines through. However, Eliot’s point that James’s prose style is characterized by its conversational quality also reminds us of the general capacity of the novel as a genre to accommodate speech across and even between linguistic registers. This insight into the novel’s special properties, we owe primarily to the works of M.M. Bakhtin, according to whom

25 Hindus (ed.) (1971), 110-114. 26 Alan Holder, ‘The Lesson of the Master: Ezra Pound and Henry James’, in American Literature, Vol. 35, No. 1 (March, 1963), 71-79: 73. 27 T.S. Eliot, untitled review, in Harvard Advocate, LXXXVIII (5 Oct. 1909); cited in Alan Holder, ‘T. S. Eliot on Henry James’, in PMLA, Vol. 79, No. 4 (Sep., 1964), 490-497: 490.

212

The novel can be defined as a diversity of social speech types (sometimes even diversity of languages) and a diversity of individual voices, artistically organized. The internal stratification of any single national language into social dialects, characteristic group behavior, professional jargons, generic languages, languages of generations and age groups, tendentious languages, languages of the authorities, of various circles and of passing fashions, languages that serve the specific socio-political purposes of the day, even the hour (each day has its own slogan, its own vocabulary, its own emphases)—this internal stratification present in every language at any given moment of its historical existence is the indispensible prerequisite of the novel as a genre.28

Bakhtin’s notion of internal linguistic strata, or heteroglossia, corresponds closely to what was earlier referred to as linguistic registers, when it was suggested that modernist poetry represents the rejection of a single fixed linguistic register. The work which most obviously suggests itself as an example of embracing a multitude of linguistic registers within a single poem is, of course, The Waste Land.29 Significantly, Eliot’s original title for the poem’s two first sections was ‘He Do the Police in Different Voices’, an allusion to Dickens’s Our Mutual Friend.

❦

Up until now M.L. Gasparov’s distinction between verse and prose – that verse is segmented into equivalent and commensurable segments independent of syntax, and prose only in relation to syntax – has stood us in good stead. However, for our present purpose it may be advantageous to limit ourselves to drawing a distinction between metrical verse and prose, and to do so in the strictest terms possible. Drawing on the Saussurean distinction between signifier and signified, we may begin by observing that whereas metrical verse is organized primarily at the level of the signifier (and only secondarily at the level of the signified), prose is organized primarily at the level of the signified (and only secondarily at the level of the signifier). That is to say that in metrical verse lexis is subordinated to and must be made to conform to the phonological exigencies of the metre. Metre, then, may be viewed as an especially salient particularization of Roman Jakobson’s notion of the poetic function, i.e. that aspect of verbal behaviour which is concerned not with the

28 M.M. Bakhtin, The Dialogic Imagination: Four Essays, edited by Michael Holquist, translated by Caryl Emerson & Michael Holquist (Austin: University of Texas Press, 1981): 262-263. 29 An even earlier example would be Hope Mirrlees’s Paris: A Poem, published by the Hogarth Press in 1919. 213

referential content of a given utterance but rather with its linguistic shape: ‘The poetic function’, Jakobson famously contends, ‘projects the principle of equivalence from the axis of selection into the axis of combination’, by which process ‘equivalence is promoted to the constitutive device of the sequence’.30 Elaborating this very principle some twenty years after it was originally put forward, Jakobson and Linda R. Waugh explain that

if a syllable is treated as a pertinent constituent of a verse-line, then one syllable is equated with any other syllable of the same sequence, whereas speakers do not measure the number of syllables in their ordinary speech. In a similar way, in certain verse-systems word-stress is assumed to equal word-stress, as unstress equals unstress, and word-stress becomes herewith a spontaneous unit of measure. ... Briefly, the verse pattern makes a choice of prosodic elements utilized by the meter, and following such a choice, syllables are converted into units of measure, as are stresses... .31

However, when equivalence is promoted to the constitutive principle of the syntagmatic axis by the introduction of metre, it affects the paradigmatic axis of selection, too. Whereas in prose a given lexical item is paradigmatically grouped with other lexical items with similar meanings, lexical items in metrical verse are first and foremost paradigmatically grouped with lexical items that have similar phonological properties. A dog in prose enters into a relatively unproblematic paradigmatic relationship with, say, canine quadruped; but in metrical verse a dog has more in common with, say, a moth than it has with a canine quadruped. It is, of course, the fact that dog is a monosyllable that precludes a five-syllable construct such as canine quadruped from taking its place in metrical discourse, in spite of the obvious semantic overlap between the two terms. This mundane observation is one very fundamental consequence of what was stated earlier, namely that metrical verse is organized primarily at the level of the signifier and only secondarily at the level of the signified, whereas in prose it is the other way around. This principal distinction between prose and metrical verse begs the question whether certain types of words, on account of their phonological properties, are better suited for metrical composition than are others. Or, alternatively, whether certain types of words, on account of their phonological properties, present special problems in metrical verse. In Chapter 1, we noted that the

30 Jakobson [1960]: 71. 31 Roman Jakobson & Linda R. Waugh, The Sound Shape of Language (Bloomington, IN & London: Indiana University Press, 1979): 216. 214

exigencies of the Greek dactylic hexameter in effect bar certain types of words – e.g. words in which a syllable with a long vowel is followed by one with a short vowel and one with a long syllable – from being used in any poem using that metre. No such strict processes of preclusion are likely to be at work in English iambic verse: quantity in Greek is, it will be recalled, a binary distinction, and that in and by itself makes Greek metres considerably less malleable than English metres based on stress. Even so, our investigation of blank verse in Chapters 4 and 5 certainly suggested that proparoxytones and monosyllabic function words can be problematic in certain positions of the iambic pentameter, although they are obviously not barred from occurring. These observations comport well with Jakobson’s claim that

A calculus of probability as well as an accurate comparison of poetic texts with other kinds of verbal messages demonstrates that the striking particularities in the poetic selection, accumulation, juxtaposition, distribution, and exclusion of diverse phonological and grammatical classes cannot be viewed as negligible accidentals governed by the rule of chance.32

In what follows we shall attempt to rise to the challenge that is implicit in Jakobson’s claim by comparing a corpus of metrical verse to a corpus of prose. What we are interested in discovering is whether the phonological profile of metrical verse in English can be shown to differ systematically from the phonological profile of English prose texts, and also – crucially – whether texts in free verse exhibit a phonological profile that can be distinguished from that of prose. After all, one of the most insistent accusations against free verse is that it is merely prose chopped into lines: if this accusation has any merit whatsoever, then the phonological profiles of free verse and prose ought to be virtually indistinguishable from one another. What exactly is entailed by the admittedly vague concept of a corpus’s phonological profile will be made explicit in due time. First, however, our investigation must be delimited in such a manner that the task of cataloguing the phonological properties of each corpus becomes both manageable and exhaustive. As I have been unable to procure any existing text corpora that are suitable for this kind of investigation, I decided to construct three corpora myself: one consisting of metrical verse, one consisting of free verse, and one consisting of novelistic prose. Each corpus was designed in such a manner that it contains a total of 20,000 running words taken from at least 100

32 Roman Jakobson, ‘Subliminal Verbal Patterning in Poetry’, in Poetics Today, Vol. 2, No. 1a (Autumn, 1980), 127- 136: 127. 215

different source texts in the relevant genre published between 1912 and 1922. Limiting ourselves to a single decade exempts us from having to take into account the possibility of diachronic interference on our results, and the decade in question has been chosen on the grounds that it represents the formative years of modernist free verse in English: 1912 saw the publication of Pound’s Ripostes and the establishment of Poetry in Chicago; 1922 is a year that requires no justification in the context of . However, even after having limited our investigation to a single decade, we are still left with the very real problem that it is virtually impossible to represent fairly and representatively any one literary genre with only 20,000 running words at one’s disposal. In recognition of this problem, I have deemed it prudent to outline briefly the exact composition as well as the principles of selection for each of the three corpora in Appendix A. Together the three corpora described in Appendix A make up a 60,000 running words Main Corpus of English literature published between 1912 and 1922. The structure of the Main Corpus may be schematized like this:

MAIN CORPUS

Metrical Verse Corpus Free Verse Corpus Novelistic Prose Corpus Number of running words 20,000 20,000 20,000 Number of texts 100 134 100 Number of authors 27 26 73

The decision to make each corpus the exact same size – i.e. 20,000 running words – was made in order to facilitate direct comparison of raw data extracted from one corpus with results extracted from the other two corpora. Each of the three corpora was first formatted as a single text file, and these three separate text files were then merged to form a new single text file, i.e. one containing the 60,000 running words of the Main Corpus. The Main Corpus text file was then subjected to a lengthy process in which each single word was annotated with a set of codes containing information about the word’s phonological properties, most importantly its number of syllables and its stress pattern. To facilitate this process, I first extracted a complete list of word-types, and it turned out that the 60,000 running words of the Main Corpus contained a total of 9,806 different word-types. By word-type should properly be understood something like unique alphanumeric string in that text 216

editing software does not register the connection between child and children and simply counts them as two different word-types: and . This is convenient for our purposes because child has very different phonological properties from children, for which reason the two word-types must be annotated with each their code. However, a far more serious problem with cataloguing word-types as alphanumeric strings is that presented by homographs and heteronyms. Text editing software does not distinguish between like as a preposition and like as a verb and simply counts the string as a single word-type. Similarly, no distinction is drawn between record as a verb and record as a noun; this latter shortcoming is highly problematic for our purposes in that the verb is an oxytone whereas the noun is a paroxytone, for which reason they must be annotated with two different codes. In order to eliminate the nuisance caused by homographs and heteronyms, I could think of no better course of action than to begin the annotation process by reading through the entire Main Corpus and annotate each occurrence of heteronyms individually. As a checklist I used John Higgins’s list of more than 600 English heteronyms, which is available online.33 The task of annotating the entire Main Corpus with information on each word’s phonological properties presents another methodological problem, one that is directly linked to the principles on which the source texts were selected for inclusion in the corpus to begin with. As ought to be obvious from the description of each corpus’s composition provided in Appendix A, no attempt has been made to limit the investigation along lines of nationality. Modernism was – to a far higher degree than any previous artistic movement – distinctly trans-national in its scope, and the same holds true of free verse as an aesthetic ideal. For this reason the Main Corpus consists of source texts that are written by authors with very different geographical backgrounds, and it consequently covers a large number of varieties of English. This fact presents us with two methodological options: 1) to analyze each source text individually according to the variety of English that its author most likely used, or 2) to analyze all the source texts according to one common standard variety of English. I have opted for the latter method, and the common standard variety of English that I have employed is that variant of which is usually referred to as Received Pronunciation. My reason for discarding option 1 is not only that it would be an extremely time-consuming venture, but also, and more importantly, that such an approach spuriously assumes that it is even possible to identify and delimit an author’s linguistic range to a single dialect of a single language (and, furthermore, that this linguistic construct is of relevance to

33 http://myweb.tiscali.co.uk/wordscape/wordlist/homogrph.html 217

an author’s work).34 As for my choice of Received Pronunciation as the common standard, it is motivated by the fact that this variety of English was named and thoroughly mapped by Daniel Jones in his English Pronouncing Dictionary from 1917, i.e. right in the middle of the period that is covered by the corpus. Jones’s dictionary thus provides us with a key to a variety of English that is contemporaneous with the publication of our corpus texts, and it has consequently been employed as a work of reference throughout the process of annotation. The choice of RP as the standard variety of English on which the annotation of the corpus is based is, in other words, a choice that gives priority to consistency over variation. However, at the same time, it should be pointed out that the implications of this methodological choice for an investigation such as this are relatively minor. Our main concern, after all, is with the number of syllables in words and their stress profile, and in these matters there is, in fact, very little variation across different varieties of English. The vast majority of English words are pronounced with the same number of syllables and the same basic stress profile in both British and American varieties of English. Furthermore, when there are differences pertaining to these matters, they tend to affect only words that occur with a comparatively low frequency. A word such as cemetery, for instance, is an example of a word-type that is typically pronounced as a proparoxytonic trisyllable in RP – [ˈsemətri] – but as a tetrasyllable with stress on the first syllable in American English: [ˈseməteri]. However, in the 60,000 running words of the Main Corpus, there are less than five occurrences of words with this particular feature. Again: the overwhelming majority of English words are pronounced with the exact same number of syllables and the exact same basic stress profile across all varieties of English. The codes that have been used for the annotation of the corpus are embarrassingly simple: monosyllables are marked with ; disyllables are marked with ; trisyllables are marked with ; tetrasyllables are marked with ; pentasyllables are marked with

; hexasyllables are marked with ; and heptasyllables are marked with . This initial letter of the code is then, whenever necessary, repeated to indicate which syllable in the word receives primary stress. For example: a word like congratulations – [kəŋɡræʧʊˈleɪʃənz] – will be annotated with the code to indicate that it is a pentasyllabic word which is stressed on its fourth syllable, whereas a word like pharmacy – [ˈfɑːməsi] – will be annotated with the code to indicate that it is a trisyllabic word that is stressed on its first syllable.

34 Cp. Bakhtin’s point that ‘monoglossia is always in essence relative. After all, one’s own language is never a single language: in it there are always survivals of the past and the potential for other-languagedness that is more or less sharply perceived by the working literary and language consciousness’. Bakhtin (1981): 66. 218

In addition to providing information on the number of syllables and the pattern of stress, the codes also contain information about the properties of the individual syllables that make up a given word. In our discussion of blank verse we noted that syllables containing a reduced vowel occasionally cause problems in metrical verse because such syllables do not easily carry a beat. This, of course, is directly linked to the fact that syllables containing a reduced vowel are by definition unstressed when they occur in English words. The most common reduced vowel in English syllables is schwa – [ə] – a mid-central vowel sound that is produced with a minimum of muscular activity, such as in the second syllable of wither: [ˈwɪðə].35 Schwa also occurs in a fronted version – [ɪ] – such as in the first syllable of syringe: [sɪˈrɪnʤ], as well as in a back, rounded version – [ʊ] – such as in the second syllable of into: [ˈɪntʊ]. When the fronted version of schwa occurs word-finally, it usually undergoes tensing and becomes [i], a phenomenon known as happy tensing on account of the fact that it occurs in the second syllable of the word happy: [ˈhæpi]. Finally, a group of syllables have such severely reduced sonority peaks that it hardly makes sense to talk of vowel sounds: these are syllables in which a sufficiently sonorous consonant – most commonly /m/, /n/ or /l/ – functions as the syllable’s peak, e.g. the second syllable of button: [ˈbʌtn ]. Syllabic consonants are marked by a vertical stroke beneath the baseline in IPA. The reduced vowels of English are distinguished from the full vowels, which are capable of occurring both in syllables that receive stress as well as in syllables that do not receive stress. Full vowels can furthermore be divided into two discreet groups on the basis of their ability to occur in open syllables, i.e. syllables that do not terminate in a consonant. Full vowels that may occur in both open and closed syllables are conventionally termed free vowels, whereas vowels that are precluded from occurring in open syllables are termed checked vowels. English RP has a phonemic inventory of six checked vowels; they are /ɪ/ as in kit; /e/ as in pet; /æ/ as in cat; /ɒ/ as in cod; /ʌ/ as in cut and /ʊ/ as in foot (note the impermissibility of *[kɪ], *[pe], *[kæ], *[kɒ], *[kʌ], *[fʊ]). The inventory of free vowels is somewhat larger and consists in RP of /iː/ as in key or beat; /ɑː/ as in spa or palm; /ɔː/ as in raw or call; /uː/ as in true or cool; /ɜː/ as in fur or churl; as well as of the /eɪ/ as in fey or pain; /əʊ/ as in low or cold; /aɪ/ as in pie or trite; /aʊ/ as in cow or Pound; /ɔɪ/ as in boy or coil; /ɪə/ as in fear or beard; /eə/ as in hair or cared; and /ʊə/ as in pure or cured.

35 While schwa is most commonly realized without lip-rounding – i.e. as [ə] – it is occasionally realized with lip- rounding, in which case it should be transcribed as [ɵ]; cp. a fence [ə fens] vs. offence [ɵˈfens]. 219

The free vowels’ ability to occur in both open and closed syllables is one that they share with the reduced vowels of English. This means that it is possible to draw a formal distinction between three separate types of vowels in English, based solely on their phonological properties:

Stressed Unstressed Open Closed syllable syllable syllable syllable Reduced vowels X X X Checked vowels X X X Free vowels X X X X

Whereas reduced vowels can occur in both open and closed syllables on the condition that the syllable in question is unstressed, checked vowels can occur in both stressed and unstressed syllables on the condition that the syllable in question is closed. Free vowels, on the other hand, can occur in both stressed and unstressed syllables, regardless of whether the syllable is open or closed. Using to designate a syllable with a reduced vowel, to designate a syllable with a checked vowel, and to designate a syllable with a free vowel, we are now in a position to expand the annotation codes used in the corpus. The full code for pharmacy is thus , i.e. a trisyllable with lexical stress on the first syllable, a free vowel in the first syllable and a reduced vowel in the second and third syllables: [ˈfɑː.mə.si]. Similarly, the full code for congratulations is , i.e. a pentasyllable with lexical stress on the fourth syllable, a reduced vowel in the first, third and fifth syllables, a checked vowel in the second syllable and a free vowel in the fourth syllable: [kəŋ.ɡræʧ.ʊ.ˈleɪ.ʃənz]. The actual ‘design’ of the codes is, quite evidently, not aimed at pleasing the reading eye, but rather at facilitating computational analysis of the corpus. Combining three separate yet interrelated parameters – number of syllables, stress assignment and vowel composition – the code system explained above provides us with a convenient matrix for distinguishing between a finite number of phonological word categories in English. Disyllables, for instance, are neatly divided into exactly twelve categories: , , , , , , , , , , and . The first six of these – those beginning with a single – are the six possible categories for disyllables with stress on the first syllable, and these six categories allow us to register the subtle differences, as well as the similarities, within a parliament of fowls such as this:

220

petrel [ˈpet.rl ] lapwing [ˈlæp.wɪŋ] magpie [ˈmæɡ.paɪ] raven [ˈreɪ.vn ] goldcrest [ˈɡəʊld.krest] skylark [ˈskaɪ.lɑːk]

For obvious reasons, the number of permissible phonological word categories increases significantly for words with more than two syllables: trisyllables, for instance, are divided into as many as 54 different categories. However, the point is that the number of possible categories remains finite and fully calculable. Before moving on to reporting the actual corpus findings, a few remarks need to be made about monosyllables. Words of one syllable are, of course, extremely frequent in Germanic languages in general, but particularly so in English. In Chapter 4, we saw how the loss of final e in a great many words in the early fifteenth century made Chaucer’s metrics incomprehensible to his successors, and this process naturally increased the number of monosyllables in the language: Chaucer’s yongë sonnë is our young sun. This process affected in particular such syllables that were the result of suffixation. All inflections of adjectival forms (except the comparative -er and -est) had thus disappeared by the Early Modern English period, a peculiar development that is unique to English among the Germanic languages.36 Moreover, the few suffixed morphemes that remained active in Early Modern English grammar and onwards are of a special kind in that only under certain phonological circumstances do they add an extra syllable to a word stem. The marker for plural, for genitive and for present tense (3rd person singular), -s, adds an extra syllable only to stems that end in a sibilant, just as the marker for past tense of weak verbs, -ed, adds an extra syllable only to stems that terminate in an alveolar plosive, i.e. /t/ or /d/. Furthermore, English does not distinguish between the infinitive form and the present tense form (except for 3rd person singular) of verbs. And, finally, it should be remembered that the English definite article is a (monosyllabic) word in its own right in contrast to the Scandinavian enclitics.37 What all of this

36 Cp. Danish en ung mand (‘a young man’) vs. den ungë mand (‘the young man’) and de ungë mænd (‘the young men’). See also Barbara A. Fennell, A History of English: A Sociolinguistic Approach (Malden, MA & Oxford: Blackwell Publishing, 2006): 142-143. 37 Cp. Danish en mand (‘a man’) vs. mandën (‘the man’). 221

means is that a monosyllabic word in English enjoys an unusually high probability rate of surviving as a monosyllable regardless of the shifting syntactic environments it may find itself in. The extreme minimalism that appears to characterize English morphology is, however, merely superficial. As noted in Chapter 5, English is a highly analytical language, for which reason it relies on a finite inventory of so-called function words to express a great many grammatical distinctions which in synthetic languages are expressed by means of morphemic suffixation. The vast majority of these function words are monosyllabic in English: articles, pronouns, prepositions, conjunctions; and the same goes for auxiliary verbs and modal verbs. Words such as these carry very little semantic weight in and by themselves; rather they function as a sort of syntactic mortar that binds together content words in various syntactic patterns. For this reason monosyllabic function words very rarely attract stress, and this makes them prone to undergo reduction: all the so- called ‘weak form words’ in English that were discussed in Chapter 5 belong to the category of monosyllabic function words. Whether a given monosyllabic function word is pronounced with a full vowel or a reduced vowel depends entirely on the context in which it occurs and it is therefore impossible to predict the exact phonological properties of monosyllabic function words in a corpus study such as this. We are consequently unable to encode the monosyllabic function words in the corpus with as detailed phonological information as that which we can assign to the content words. It should, however, be noted that the most frequent monosyllabic function words tend to undergo reduction quite systematically. This is particularly true of the articles – the [ðə]/[ði], a [ə], an [ən] – and of certain very frequent prepositions, e.g. of [əv], in [ɪn] and to [tʊ]. Auxiliary verbs and modal verbs also tend to undergo reduction, to the extent that they’re frequently contracted, even in writing. However, in recognition of the fact that the exact syllable quality of monosyllabic function words cannot be accurately predicted on phonological principles alone, I have restricted myself to encoding monosyllabic function words only with the information that they consist of a single syllable in a particular word class. Monosyllabic content words, on the other hand, have been subjected to the full treatment of code annotation, and yet they fall into only two separate categories based on the criteria used in this investigation: those with a checked vowel, , and those with a free vowel, . This digression devoted to the special challenges that English monosyllables present is justified overwhelmingly by the corpus results. If we begin by focusing only on the number of syllables that the 60,000 running words of the Main Corpus contain, we find that close to three quarters of them are, in fact monosyllables: 222

MAIN CORPUS (60,000 running words)

Words Syllables Number Percentage Number Percentage Monosyllables 44281 73.802% 44281 54.479% Disyllables 11796 19.660% 23592 29.193% Trisyllables 2933 4.888% 8799 10.888% Tetrasyllables 825 1.375% 3300 4.083% Pentasyllables 148 0.247% 740 0.916% Hexasyllables 16 0.027% 96 0.119% Heptasyllables 1 0.002% 7 0.009% 60000 80815

In the columns under the heading ‘Words’, the raw number of occurrences for each word category is given alongside the corresponding relative frequency. The columns under the heading ‘Syllables’ show how many syllables a given word category accounts for and its corresponding frequency in relation to the total number of syllables. For example: the Main Corpus yields a total of 11,796 disyllables; this means that 19.66 per cent of all the words in the corpus are disyllables. The 11,796 disyllables account for 23,592 syllables; this means that 29.193 per cent of all the syllables in the corpus come from disyllables. By comparing the number of words and the number of syllables, we can see that the 60,000 running words contain a total of 80,815 syllables, corresponding to a syllable per word ratio of 1.347:1. This initial corpus result bears out and makes explicit what intuition and common sense would also seem to promise: that in English the frequency of a given word category is inversely proportional to its number of syllables. Together the categories of monosyllables, disyllables and trisyllables account for more than 59,000 of the 60,000 running words and for close to 95 per cent of the 80,815 syllables in the corpus. Paying attention to the number of syllables in a text corpus is, of course, highly unusual in corpus studies, but for our purposes this parameter is of crucial importance, as Jakobson & Waugh have reminded us already: ‘if a syllable is treated as a pertinent constituent of a verse-line, then one syllable is equated with any other syllable of the same sequence, whereas speakers do not measure the number of syllables in their ordinary speech’. In this study we have been measuring the number of syllables in three corpora – one consisting of metrical verse, one consisting of free verse and one consisting of novelistic prose – and the figures given in Table 7.1 represent the findings from these three corpora taken together. What we are interested in discovering now is whether there are systematic differences between each 223

of the three corpora in terms of the frequency with which each word category is represented. For example: does the fact that disyllables occur with a frequency of 19.66 per 100 words in the Main Corpus allow us to estimate the number of disyllables in each of our three sub-corpora as somewhere in the vicinity of 0.1966 X 20,000 running words, i.e. 3,932? As it turns out, it does: the Novelistic Prose Corpus contains 4,151 disyllables (ratio 20.755 per 100 words); the Metrical Verse Corpus contains 3,829 disyllables (ratio 19.145 per 100 words); and the Free Verse Corpus contains 3,816 disyllables (ratio 19.08 per 100 words).

Similarly with monosyllables: given their ratio of 73.802 per 100 words in the Main Corpus, we would expect to find around 14,760 monosyllables in each sub-corpus, and what we find is that the Novelistic Prose Corpus contains 13,897 monosyllables (ratio 69.485 per 100 words); the Metrical Verse Corpus contains 15,262 monosyllables (ratio 76.31 per 100 words); and the Free Verse Corpus contains 15,108 monosyllables (ratio 75.54 per 100 words).

More than anything else, the relative uniformity of these findings serves as an illustration of Hollander’s important point cited earlier that foot scansion of free verse in English tends to assert ‘a trivial correlation built into the structure of English’. None of the results reported above diverge from the mean values of the Main Corpus by more than some 6 per cent. At the same time, however, it should be noted that the reported figures for the Free Verse Corpus are strikingly close to those for the Metrical Verse Corpus, and that the Novelistic Prose Corpus is consistently the odd one out. This general tendency is also reflected in the total number of syllables that each corpus contains: the Novelistic Prose Corpus contains a total of 28,699 syllables (ratio 1.435 syllables per word); the Metrical Verse Corpus contains a total of 26,272 syllables (ratio 1.314 syllables per word); and the Free Verse Corpus contains a total of 25,710 syllables (ratio 1.286 syllables per word).

However, when we look at the distribution of trisyllables across our three sub-corpora, a far more diversified pattern begins to emerge. In the Main Corpus, trisyllables occur with a frequency of 4.888 per 100 words, and, given the very even distribution of monosyllables and disyllables, this 224

ought to allow us to assume that each sub-corpus would contain some 978 trisyllables. However, what we find is that the Novelistic Prose Corpus contains 1,378 trisyllables (ratio 6.89 per 100 words); the Metrical Verse Corpus contains 730 trisyllables (ratio 3.65 per 100 words); and the Free Verse Corpus contains 825 trisyllables (ratio 4.125 per 100 words).

This tendency – i.e. a considerably higher frequency in the Prose Corpus than in the Verse Corpora – becomes even more pronounced in the distributional patterns of words with more than three syllables: more than half of the 825 tetrasyllables in the Main Corpus occur in the Novelistic Prose Corpus, and the same is true of two thirds of the 148 pentasyllables in the Main Corpus. Hexasyllables and heptasyllables occur almost exclusively in the Novelistic Prose Corpus. Thus, the overall tendency of our study so far is that all word categories, except monosyllables, consistently occur more frequently in the Prose Corpus than they do in the Verse Corpora, and that this tendency increases with the number of syllables in the word category. Conversely, monosyllables occur more frequently in the Verse Corpora than they do in the Prose Corpus. Turning our attention to the distribution of the monosyllables in each of our three sub- corpora, a remarkable pattern emerges. During the process of corpus annotation, a distinction was made between monosyllabic function words and monosyllabic content words, but so far both kinds have been counted simply as monosyllables. However, when the distinction between lexical and functional monosyllables is taken into account, we find that whereas the ratio of monosyllabic function words remains relatively uniform across the three sub-corpora, there is considerable variation in the distribution of monosyllabic content words:

Monosyllables Lexical Function Ratio Total Monosyllables Monosyllables Lex./Funct. Novelistic Prose 13911 5208 8703 0.599:1 Metrical Verse 15262 7315 7947 0.920:1 Free Verse 15108 6219 8889 0.699:1

Given the analytical nature of English, it is hardly surprising to find that some 40 to 45 per cent of all words in the corpora belong to the category of monosyllabic function words. It should, however, be noted that the rate of monosyllabic function words is somewhat lower in the Metrical Verse Corpus (ratio 39.72 per 100 words) than in the Novelistic Prose Corpus (ratio 43.45 per 100 words) 225

and in the Free Verse Corpus (ratio 44.45 per 100 words). However, a far more significant difference characterizes the ratio with which monosyllabic content words occur: in the Novelistic Prose Corpus monosyllabic content words occur with a ratio of 26.04 per 100 words, whereas in the Metrical Verse Corpus the ratio is as high as 35.56 per 100 words; the Free Verse Corpus is poised in between (ratio 31.10 per 100 words). This means that there are over 40 per cent more monosyllabic content words in the Metrical Verse Corpus than in the Novelistic Prose Corpus. If we look more closely at the population of monosyllabic content words in the three sub- corpora and take into account the distinction between those with a checked vowel and those with a free vowel, we find that the increase of monosyllabic content words in metrical verse is most significant for those that contain a free vowel:

Lexical Checked Free Ratio Monosyllables Vowel Vowel Checked/Free Novelistic Prose 5208 1841 3367 0.547:1 Metrical Verse 7315 2403 4912 0.489:1 Free Verse 6219 2060 4159 0.495:1

The Metrical Verse Corpus contains some 30 per cent more monosyllabic content words with a checked vowel than does the Novelistic Prose Corpus, but almost 46 per cent more monosyllabic content words with a free vowel. A similar tendency can also be detected if we compare The Novelistic Prose Corpus to the Free Verse Corpus: the Free Verse Corpus contains some 12 per cent more monosyllabic content words with a checked vowel than does the Novelistic Prose Corpus, but over 23 per cent more monosyllabic content words with a free vowel. The category of monosyllabic content words is unique in that it does not contain any reduced vowels. Based on the 18,742 occurrences of monosyllabic content words in the totality of the Main Corpus, it would seem that free vowels are considerably more prevalent than checked vowels in this category of words, almost exactly twice as common. Disyllables, on the other hand, may contain all three types of vowels: reduced, checked and free. Given that all English words of more than one syllable are lexically stressed on one of their syllables, and that reduced vowels are precluded from occurring in stressed syllables, it goes without saying that we will not find disyllables with the codes *, *, *, *, * or *. However, all of the twelve permissible types of disyllables listed earlier occur in each of the three sub-corpora, albeit with very different frequencies:

226

Novelistic Prose Metrical Verse Free Verse Main Corpus Corpus Corpus Corpus 1630 1401 1430 4461 45 41 41 127 153 159 197 509 1328 1402 1277 4007 46 62 57 165 165 117 171 453 289 246 241 776 422 349 361 1132 3 10 6 19 25 20 8 53 22 11 16 49 23 11 11 45 Total 4151 3829 3816 11796

In all the sub-corpora, the frequency of paroxytonic disyllables is some five to six times higher than the frequency of oxytonic disyllables: another trivial correlation built into the structure of English. But it is noteworthy that the four types of disyllables in which the unstressed syllable contains a reduced vowel account for as many as 10,376 of the 11,696 occurrences of disyllables, i.e. 88 per cent. Even so, full vowels still outnumber reduced vowels in the category of disyllables: of the 23,592 syllables yielded by the 11,796 disyllables in the corpus, 13,216 contain a full vowel, and 10,376 contain a reduced vowel. Of the 13,216 full vowels, 6,305 are checked, and 6,911 are free. Thus, if we compare the vowel profile of the population of disyllables with that of the population of monosyllabic content words, we find that the presence of reduced vowels in the disyllables affects the proportion of free vowels far more dramatically than it does the proportion of checked vowels:

Reduced vowels Checked vowels Free vowels Lexical monosyllables 0% 33.64% 66.36% Disyllables 43.98% 26.73% 29.29%

If we extend our survey to include the category of trisyllables, it should first be noted that reduced vowels outnumber full vowels in the population of trisyllables when taken as a whole. Of the 8,799 syllables yielded by the 2,399 trisyllables that occur in the corpus, 4,853 contain a reduced vowel, and 3946 contain a full vowel, of which 2229 are checked, and 1717 are free:

227

Lexical monosyllables Reduced vowels Checked vowels Free vowels Trisyllables 55.15% 25.33% 19.51% Polysyllables 60.17% 22.42% 17.41%

As can be seen, the higher proportion of reduced vowels in the category of trisyllables amplifies what we observed concerning disyllables: that the ratio of checked vowels remains relatively unaffected by the influx of syllables with reduced vowels, while the ratio of free vowels drops, so that for trisyllables the ratio of free vowels is even lower than that of checked vowels. The same pattern can be observed in the category of polysyllables, i.e. those words in the corpus that contain more than three syllables. This has a rather interesting effect when we compare the relative frequency of syllables with a checked vowel across our three corpora. The codes with which the corpora were annotated allow us to extract a complete list of occurring syllables for each corpus, from each category of words. Leaving the monosyllabic function words out of the account for now, the three corpora yield the following results:

Monosyllablesbubvvvv Reduced Checked Free Total vowels vowels vowels Novelistic Prose

Monosyllables 0 1835 3373 5208 Disyllables 3669 2261 2372 8302 Trisyllables 2314 1070 750 4134 Polysyllables 1424 530 412 2366 Total 7407 5696 6907 20010 Ratio 0.3702 0.2847 0.3452

Metrical Verse

Monosyllabic 0 2403 4911 7314 Disyllables 3398 2001 2259 7658 Trisyllables 1198 542 450 2190 Polysyllables 442 164 128 734 Total 5038 5110 7748 17896 Ratio 0.2815 0.2855 0.4329

Free Verse

Monosyllables 0 2060 4159 6219 Disyllables 3309 2043 2280 7632 Trisyllables 1341 617 517 2475 Polysyllables 636 234 173 1043 228

Total 5286 4954 7129 17369 Ratio 0.3043 0.2852 0.4104

In the three middle columns are given the number of syllables with reduced, checked and free vowels respectively in each corpus and for each category of words. As can be seen, the ratio of syllables with checked vowels is remarkably stable: in each corpus some 28.5 per cent of all occurring syllables (except for the monosyllabic function words) contain a checked vowel. The ratio of syllables with long vowels, on the other hand, differs markedly: in the Prose Corpus 34.52 per cent of all syllables (except for the monosyllabic function words) contain a long vowel, whereas in the Metrical Verse Corpus this is the case for 43.29 per cent. Conversely, the ratio of syllables with a reduced vowel is significantly lower in Metrical Verse than in Novelistic Prose.38 In the Prose Corpus as many as 3,534 of the syllables with free vowels are yielded by words with more than one syllable; in the Metrical Verse Corpus the corresponding figure is 2,837, and in the Free Verse Corpus it is 2,970. This means that the higher ratio of syllables with free vowels in the two Verse Corpora is caused solely by the very high proportion of monosyllabic content words with free vowels in those two corpora. When the monosyllabic function words are taken into account, and we change our focus to consider the frequency of monosyllabic content words with free vowels per word instead of per syllable, we find that

16.865 per cent of the 20,000 running words in the Novelistic Prose Corpus; 24.555 per cent of the 20,000 running words in the Metrical Verse Corpus; and 20.795 per cent of the 20,000 running words in the Free Verse Corpus are monosyllabic content words with free vowels. Or, in more readily comprehensible numbers: in metrical verse one in four words is a monosyllabic content word with a free vowel; in free verse it is one in five, and in novelistic prose it is one in six.

38 If we take into account the category of monosyllabic function words (MFW), the same tendency can be detected, although a bit less conspicuously: the ratio of syllables with checked vowels remains relatively stable across the three corpora, whereas the ratios of syllables with reduced vowels and syllables with long vowels fluctuate:

Reduced Checked Long MFW Novelistic Prose 0.2580 0.1984 0.2406 0.3031 Metrical Verse 0.1949 0.1977 0.2998 0.3075 Free Verse 0.2013 0.1887 0.2715 0.3385

229

In order to test the consistency of this distributional pattern, I decided to split up each sub- corpus into five even-sized text files of 4,000 running words and calculate the relative frequency of monosyllabic content words with free vowels for each text file. At the same time I also calculated the relative frequency of monosyllabic content words with checked vowels. In the table below, the percentage of running words that each category accounts for in the fifteen text files are given (monosyllables with free vowels are in the middle row; those with checked vowels are in the bottom row):

NP1 NP2 NP3 NP4 NP5 MV1 MV2 MV3 MV4 MV5 FV1 FV2 FV3 FV4 FV5 16.6 16.125 16.475 18.35 16.625 24.1 24.25 24.25 25.65 24.55 18.75 20.025 20.575 21.875 22.75 8.925 8.9 9.4 9.6 9.2 13.65 12.075 11.125 11.65 11.1 10.75 10.45 11.25 10.15 8.9

A more convenient method for comparing these figures is to insert them into a coordinate system, in which the horizontal axis represents the percentage of monosyllabic content words with free vowels, and the vertical axis represents the percentage of monosyllabic content words with checked vowels:

14

13

12

11

10

9

8 16 17 18 19 20 21 22 23 24 25 26

The light grey diamonds in the left bottom corner are the five text files of Novelistic Prose; the dark grey diamonds in the top right corner are the five text files of Metrical Verse; and the black diamonds in between are the five text files of Free Verse. As can be seen, the distributional patterns of monosyllabic content words appear to be relatively stable, at least from a bird’s eye perspective. The text files containing metrical verse are clearly distinguishable from the text files containing novelistic prose, whereas the text files containing free verse fall somewhere in the middle. This 230

distribution calls to mind Ford Madox Ford’s comment in his memoir Thus to Revisit (1921): ‘And so the case for vers libre is made. ... For who in his senses will deny that, between the entrenched lines of Prosaists and Versificators lies a No Man’s Land that is the territory of Neither-Prose-Nor- Verse?’39 The distribution of monosyllabic content words is obviously not a sufficient criterion for assessing the relationship between prose, verse and free verse. However, the pronounced difference between the frequency of monosyllabic content words with free vowels in metrical verse (1 in 4 words) and in novelistic prose (1 in 6 words) allows us to venture the hypothesis that in order to sustain iambic metre in English, it is expedient to keep the ratio of reduced vowels low and the ratio of free vowels high. This at least is what the corpus results have consistently indicated. The Metrical Verse Corpus contained fewer monosyllabic function words than the other two corpora: such words are likely to undergo reduction and thereby contribute to increasing the ratio of reduced vowels. The Metrical Verse Corpus also contained fewer trisyllables and polysyllables than did the other two corpora: such words contain many reduced vowels and therefore also contribute to increasing the ratio of reduced vowels. Given that syllables with reduced vowels never attract stress, they are far more likely to occupy the weak syllable positions of the metre than the strong syllable positions. When they do occupy a strong syllable position, however, metrists speak of promotion. Conversely, demotion occurs when a syllable which receives stress (and consequently has a full vowel) occupies a weak position in the metre. Both promotion and demotion are, of course, common enough features of English iambic verse; the two phenomena are, in fact, responsible for creating much of the subtle variation that characterizes successful poetry in metre.40 However, the general tendency of English iambic verse is that reduced vowels occur relatively infrequently in the strong positions, and this may be one of the reasons for their lower frequency in the Metrical Verse Corpus. The increased frequency of syllables with free vowels in the Metrical Verse Corpus also merits a few comments. Up until now, we have been careful to distinguish between full vowels only on the basis of their ability to occur in open syllables. However, the checked/free distinction corresponds closely to the less technical distinction between ‘short’ and ‘long’ vowels in English: the checked vowel in pull is notably shorter than the vowel in pool or the diphthong in pole. Furthermore, short vowels are lax, whereas their long counterparts are tense; that is to say that short vowels in general are more centralized and therefore less distinct than long vowels. Vowel length,

39 Ford Madox Hueffer, Thus to Revisit (New York: Dutton, 1921): 194. 40 See, for example, Steele (1999): 8-11. 231

of course, is not a structural constituent in English accentual-syllabic versification, but D.S. Carne- Ross still ventures to ask the obvious question:

What part does quantity play in English verse? And (even harder to answer), how does one distinguish between a stressed and a quantitatively long vowel? I think one might begin by suggesting that the role of quantity in English verse is roughly analogous to that of pitch in Greek and accent in Latin verse. That is, in every developed metrical system there exists, alongside the “official” determinant (quantity in Greek and Latin, stress in English), one or more other elements which play an important but not easily definable part. Take—in its context, of course—Milton’s line (PL 1.599): / / / / Perplexes Monarchs. Dārk’n’d so, yet shon

where my scansion marks point to the syllables on which the main stresses fall and also to the fact that the a of “Dark’n’d” is quantitatively long in addition to being stressed. No less certain is it that the a of “Monarchs” is quantitatively long and yet unstressed.41

This is certainly not the place to re-open the century-old feud between ‘stressers’ and ‘timers’, but that is not Carne-Ross’s purpose either. Rather, his point seems to be that the fact that only one phonological feature – stress – is used as a constituent in English metre makes possible an interesting interplay with other phonological features, including vowel length. The very high ratio of monosyllables with long vowels in the Metrical Verse Corpus suggests that metre, at least to some extent, depends for its effects on a higher proportion of long vowels than is found in prose. Aldington’s contention cited earlier that free verse ‘ought to be about five times more concentrated than the best prose and about six times more emotional’ can obviously not be supported by the findings of this (or any other) corpus investigation. However, the findings do indicate that free verse, in its phonological texture, is at least as far removed from prose as it is from metrical verse: in fact, the Free Verse Corpus quite consistently occupies an intermediate position between the Metrical Verse Corpus and the Novelistic Prose Corpus with regard to almost all of the features examined here.

41 D.S. Carne-Ross, ‘New Metres for Old: A Note on Pound’s Metric’, in Arion, Vol. 6, No. 2 (Summer, 1967), 216- 232: 223. 232

Concluding Remarks

The year 1912 is an important one in the history of English modernist verse: it saw the publication of Pound’s Ripostes (and, with it, Hulme’s Complete Poetical Works), as well as the establishment of Poetry, the magazine which arguably did more than any other to further the cause of free verse. But 1912 was also the year in which W.P. Ker’s English Literature: Medieval was first published. This work is alluded to by Pound in the January 1914 issue of Poetry: ‘Dr. Ker has put an end to much babble by showing us Summer is ycummen in written beneath the Latin words of the first known example of a canon’.1 Ker’s comments on this matter are worth quoting in a little more detail:

[The Cuckoo song] is not an improvisation, but the newest kind of art, one of the most ingenious things of its time. Further, the words that belong to it are Latin words, a Latin hymn; the Cuckoo song, which appears so natural and free, is the result of deliberate study; syllable for syllable it corresponds to the Latin, and to the notes of the music.2

It is easy to see why such an argument might resonate with Pound. The emphasis on deliberate study of foreign verse traditions as a necessary precondition for freedom and innovation has a decidedly Poundian ring to it: ‘The tradition’, Pound says at the opening of the essay in which he alludes to Ker, ‘is a beauty which we preserve and not a set of fetters to bind us. This tradition did not begin in A. D. 1870, nor in 1776, nor in 1632, nor in 1564. It did not begin even with Chaucer’.3

This study has only intermittently ventured further back than Chaucer: it is, after all, to him that we owe the invention of the iambic pentameter, the quintessential English metre which the modernists singled out for breaking. However, as we saw in Chapter 3, the iambic pentameter is first and foremost characterized by its decidedly mixed lineage, and its structure echoes and preserves traits from Italian, Occitan, French and even Greek verse forms. In this respect, the iambic pentameter as a metrical form could well be said to embody many of the virtues that the proponents of free verse claimed to be seeking, and yet Pound was satisfied to represent its basic features as ‘ti tum ti tum ti tum ti tum ti tum’.4 This caricature hardly does justice to the versatility with which the iambic pentameter has been employed in English verse since the late sixteenth

1 Ezra Pound, ‘The Tradition’, in Poetry, Vol. 3, No. 4 (1914), 137-141: 139. 2 W.P. Ker, English Literature: Medieval [1912] (London: Williams & Norgate, 1926): 75. 3 Pound (1914): 137. 4 Pound (1951): 203. 233

century, and Pound, of course, was well aware of this. We must therefore interpret Pound’s caricature first and foremost as a comment on what the iambic pentameter – and, by extension, accentual-syllabic metre as such – was habitually reduced to by the turn of the twentieth century. Perusing the poems in iambic pentameter from the Georgian Poetry series, it is not difficult to sympathize with the modernist aspiration to come up with an alternative, but at the same time we need to remind ourselves of the truly radical nature of this alternative. In spite of earlier attempts to challenge the hegemony of iambic verse in English poetry, the modernist rejection of metre was unprecedented, not only in scope, but also in terms of the success with which it has subsequently met. Few would dispute the claim that some of the twentieth century’s finest poetry in English owes its existence to the advent of modernist free verse, and even fewer would deny that the absence of metre is a defining characteristic of much modern poetry. It has not been my intention in this dissertation to pass judgement on the merits of either metrical verse or free verse, but as a way of concluding it seems pertinent to raise the question of how the emergence of free verse has influenced the field of metrics. In Chapter 2 we noted that free verse rose to prominence simultaneously with the emergence of modern linguistics, and in certain respects linguistics has taught us more about metre than philology ever did. However, because metre is no longer viewed as a requirement for poetry, the subject of metrics is in danger of being relegated to the margins of literary studies, especially within the educational system.5 If metrics is reduced to a catalogue of arbitrary patterns with exotic names, we lose sight, not only of important historical aspects of versification, but also of the crucial ways in which metre interacts with other levels of organization such as diction and syntax. This is not only detrimental to the understanding and appreciation of metrical verse; it also diminishes our capacity to engage fruitfully with free verse. After all, one of the most insistent claims about free verse is that it presupposes a metrical tradition from which to depart, and it is this tradition that the present study has sought to elucidate.

5 For evidence of this, see Bill Overton, ‘“People Have Forgotten How to Hear the Music”: The Teaching of Poetry and Prosody’, in English (2008), 57 (219): 266-282. 234

Appendix A: The Composition of the Corpora

The Metrical Verse Corpus The period 1912-1922 coincides exactly with the publication dates of the five volumes of Georgian Poetry, which series is generally viewed as the most significant outlet for traditionalist versification in English during a period that we have now come to think of as anti-traditionalist. For this reason, the Georgian Poetry series has been chosen as the basis for the Metrical Verse Corpus, although it should be noted that I have decided to exclude two Georgian Poetry texts from consideration, namely Lascelles Abercrombie’s The Sale of Saint Thomas and Gordon Bottomley’s King Lear’s Wife, from the first and second anthology respectively. Both texts are quite lengthy plays in blank verse and thus clearly depart from the general lyric mode which characterizes the Georgian project as a whole. Only texts written wholly – or overwhelmingly so – in iambic pentameter have been included in the Metrical Verse Corpus, but in order to keep the corpus at a manageable size, some texts have not been included in their entirety. Each of the five anthologies is represented equally in the corpus – that is with 4,000 words each – even though the number of poems in iambic pentameter differs significantly from volume to volume. In the first anthology, less than ten poems are in iambic pentameter, whereas in the fourth anthology the number of poems in that measure exceeds thirty. For this reason the procedure of inclusion has differed slightly from the first anthologies to the last ones, although the same basic principles have been applied. From the first three anthologies, all poems in iambic pentameter have been included in the corpus, but only texts with less than 1,000 words have been included in their entirety. For example: the second anthology contains ten poems in iambic pentameter, eight of which have less than 1,000 words. Together these eight source texts contain 1,726 words, leaving us 2,274 words short of the desired 4,000 in total. The 2,274 missing words have then been split equally between the two source texts with more than 1,000 words so that each of these texts is represented in the sub-corpus with 1,137 words. However, the two final anthologies contain a considerably higher number of relatively short poems in iambic pentameter, for which reason the cutting-off point for full inclusion needed to be set at 300 words so as to not exceed the 4,000 words limit. For example: the fifth anthology contains 27 poems in iambic pentameter, 18 of which contain 300 words or less. Together these 18 poems contain 2,530 words, leaving us 1,470 words short of the desired 4,000 in total. These 1,470 missing words have 235

then been split equally between the nine source texts with more than 300 words so that each of these texts is represented in the sub-corpus with 163 words. All in all, the corpus of metrical verse comprises 78 full poems and 22 excerpts from poems, amounting to a total of 100 source texts by 27 different poets. The featured poets are Lascelles Abercrombie, Martin Armstrong, Herbert Asquith, Edmund Blunden, Gordon Bottomley, Rupert Brooke, William H. Davies, John Drinkwater, John Freeman, Wilfred Wilson Gibson, Robert Graves, Richard Hughes, William Kerr, Francis Ledwidge, Walter de la Mare, John Masefield, Harold Monro, Robert Nichols, Frank Prewett, V. Sackville-West, Siegfried Sassoon, Edward Shanks, Fredegond Shove, J.C. Squire, James Stephens, W.J. Turner and Francis Brett Young. A complete list of the poems featured in the Metrical Verse Corpus is given in Appendix B.

The Free Verse Corpus For the Free Verse Corpus, all source texts, except one, have been compiled from three important anthologies of modernist verse; the anthologies in question are Des Imagistes (1914), edited by Ezra Pound; Some Imagist Poets (1915), edited by Amy Lowell; and Others for 1919 (1920), edited by . While the two imagist anthologies are widely viewed as representing two different phases of imagism – Pound referred to Lowell’s species of imagism as ‘Amygism’ – they were both directly inspired by the great success that the first volumes of Georgian Poetry had enjoyed. One year prior to the actual publication of Des Imagistes, Edward Marsh already suspected as much, as is evident from a letter he sent to Rupert Brooke in 1913: ‘Wilfrid [Scawen Blunt] tells me there’s a movement for a “Post-Georgian” Anthology, of the Pound-Flint-Hulme school, who don’t like being out of GP, but I don’t think it will come off.’1 The anthology did, of course, eventually come off, but its moderate commercial success soon led to open disagreement between Pound and his fellow imagists. Lowell especially was keen to expand the market for imagist poetry and suggested publishing a new anthology under that aegis for five consecutive years; such an enterprise, she pointed out, ‘would enable us, by constant iteration, to make some impression on the reading public ... it is a method which the editors of the “Georgian anthology” have found most satisfactory’.2 Pound refused to take part in this venture, leaving Lowell to issue, not five, but three

1 Cited in Joy Grant, Harold Monro & the Poetry Bookshop (Berkeley & Los Angeles: University of California Press, 1967): 99. 2 Cited in Carr (2009): 677. 236

yearly anthologies between 1915 and 1917 under the heading of Some Imagist Poets.3 It is the first of these volumes that has been selected as a basis text for the Free Verse Corpus. Des Imagistes contains 35 poems by eleven different poets, to which are appended a few so- called ‘Documents’ at the end of the volume. Of the 35 poems in the main body of the anthology, 28 have been included in the corpus in their entirety (the remaining seven poems being metrical), and most of them are by poets who have been discussed at length already: Richard Aldington, H.D., F.S. Flint, Amy Lowell, William Carlos Williams, Ezra Pound and Ford Madox Hueffer. Only Skipwith Cannell, who is represented in the corpus with a single poem, requires introduction here: he was an American poet who in 1913 had been introduced to Pound in Paris through John Gould Fletcher, and whose main source of inspiration for writing free verse appears to have been the Old Testament.4 The 28 poems from Des Imagistes selected for inclusion in the Free Verse Corpus account for a total of 3,429 running words. The first volume of Some Imagist Poets contains 37 poems by six different poets: Aldington, H.D., John Gould Fletcher, F.S. Flint, D.H. Lawrence and Amy Lowell. With a few obvious exceptions – Lowell’s prose piece ‘The Bombardment’ and Lawrence’s and Flint’s poems in metre – the anthology is included in its entirety in the Free Verse Corpus, accounting for a total of 29 poems that yield 5,033 running words all in all. Whereas both Des Imagistes and Some Imagist Poets are sufficiently brief to have been included almost in their entirety, Others for 1919 is considerably more comprehensive in its scope and contains over 100 poems by 26 different poets. Thus, in order not to exceed the 20,000 words limit for the corpus as a whole, it has been necessary to exclude a number of poems from this anthology. Some poems exclude themselves by being metrical – e.g. the contributions by Robert Frost, Conrad Aiken, H.L. Davis and – while other poems have been excluded on account of the fact that they are bordering on being prose pieces, e.g. the contributions by Robert Alden Sanborn, Fenton Johnson, Arturo Giovannitti and Orrick Johns. In the end, a total of 76 poems from Others for 1919 were included in the corpus in their entirety, representing the work of sixteen different poets. Of these the most widely known today are probably William Carlos Williams, Wallace Stevens, Mina Loy and Lola Ridge. Also included are poems by the founder of Others, Alfred Kreymborg, as well as by two of the Spectra hoaxers, namely Witter Bynner and Marjorie Allen Seiffert. However, some of the Others poets that are featured in the Free Verse Corpus have received more recognition for their work in the visual arts than for their poetic

3 For an account of the schism between Pound and Lowell, see Carr (2009): 674-713. 4 See Carr (2009): 551-552. 237

experiments. These include Marsden Hartley, a prominent member of the Regionalist Movement in American painting, whose work was exhibited at the famous 1913 Armory Show in New York,5 and Jeanne d’Orge, an eccentric painter and poet, who played a central role in the artists’ community in Carmel-by-the-Sea in California during the 1920s.6 The remaining poets from Others for 1919 that have been included in the corpus are Evelyn Scott, a native of Tennessee who achieved some success as a novelist in the 1920s and 1930s; Emanuel Carnevali, an Italian immigrant who returned to his native country after having published only a single volume of verse in his own name;7 Donald Evans, the founder of the Claire Marie Press and the first publisher of Gertrude Stein’s Tender Buttons;8 Haniel Long, a protégé of Witter Bynner and a peripheral member of the Santa Fe writers’ community during the 1920s and 1930s;9 Wallace Gould, a former child prodigy at the piano,10 who gave piano lessons to William Carlos Williams’s son Bill during the winter of 1918;11 William Saphier, a Romanian immigrant who worked as a co-editor for Others; and, finally, Max Michelson, of whom there is nothing interesting to report that I have been able to dig up. Altogether the 76 poems from Others for 1919 selected for inclusion in the Free Verse Corpus amount to a total of 10,501 running words. The selections from the two imagist anthologies and those from the Others anthology provide us with 133 texts for the Free Verse Corpus and yield a total of 18,963 running words. In order to fill the gap between this number and the requirement of 20,000 running words in the corpus as a whole, I chose to include the opening 1,037 words from Eliot’s ‘The Love Song of J. Alfred Prufrock’. Eliot is not featured in any of the three anthologies on which the Free Verse Corpus is based, but given his pivotal role in the modernist movement he ought to be represented in the corpus. His poem is the only text in the Free Verse Corpus that is not featured in its entirety, and with its inclusion in the corpus there are 26 different poets represented. For details, see Appendix C.

5 Townsend Ludington, Marsden Hartley: The Biography of an American Artist (New York: Cornell University Press, 1998): 99. 6 Jeanne d’Orge, whose real name was Lena Yates, was the mother of the children’s author Virginia Lee Burton as well as the wife of Carl Cherry, the inventor of the Cherry friction-lock rivet, which revolutionized the manufacture of aircraft during World War II and made him a millionaire. On this, see Barbara Elleman, Virginia Lee Burton: A Life in Art (New York: Houghton Mifflin Harcourt, 2002): 7-12. 7 Tales of a Hurried Man (1925). 8 Jerome McGann, Black Riders: The Visible Language of Modernism (Princeton, NJ: Princeton University Press, 1993): 19. 9 Lynn Cline, Literary Pilgrims: The Santa Fe and Taos Writers’ Colonies 1917-1950 (Albuquerque, NM: University of New Mexico Press, 2007): 132-143. 10 Marsden Hartley, Somehow a Past: Prologue to Imaginative Living, edited with an Introduction by Susan Elizabeth Ryan (Cambridge, MA: M.I.T. Press, 1997): 54. 11 William Carlos Williams, The Autobiography of William Carlos Williams (New York: New Directions, 1967): 165. 238

The Novelistic Prose Corpus In his Preface to The Oxford Book of English Prose (1925), Arthur Quiller-Couch asks the pertinent question whether ‘any anthology of short passages [can] rightly illustrate an art of which the property is to be long?’12 What Quiller-Couch points out with reference to prose in general applies to novelistic prose even more saliently. Thus, whereas the two Verse Corpora – on account of the relative brevity of poems in general – mostly contain complete texts, the Novelistic Prose Corpus, for quite obvious reasons, features only excerpts from novels and short stories. Furthermore, the Novelistic Prose Corpus is different from the two Verse Corpora in that it is not put together on the basis of contemporary anthologies, for the simple reason that no such relevant anthologies of novelistic prose exist. Instead, the required 20,000 running words of the Novelistic Prose Corpus have been put together by including the opening 200 words from a random selection of 100 novels and short stories that were first published in the period between 1912 and 1922. The texts chosen for inclusion are by 73 different authors; they are Sherwood Anderson, Arnold Bennett, J.D. Beresford, John Buchan, Frances Hodgson Burnett, James Branch Cabell, Abraham Cahan, Willa Cather, Agatha Christie, Marie Corelli, Joseph Conrad, E.E. Cummings, Norman Douglas, Arthur Conan Doyle, Theodore Dreiser, Edna Ferber, F. Scott Fitzgerald, Ford Madox Ford, John Fox Jr., John Galsworthy, David Garnett, Charlotte Perkins Gilman, Ellen Glasgow, Zane Grey, H. Rider Haggard, Georgette Heyer, William Dean Howells, E.M. Hull, Aldous Huxley, Jerome K. Jerome, James Weldon Johnson, James Joyce, Robert Keable, Sheila Kaye-Smith, D.H. Lawrence, Stephen Leacock, Sinclair Lewis, David Lindsay, Jack London, Rose Macaulay, Compton Mackenzie, Katherine Mansfield, W. Somerset Maugham, Harold McGrath, H.C. McNeile, Abraham Merritt, A.A. Milne, L.M Montgomery, George Moore, Christopher Morley, Baroness Orczy, John Dos Passos, Eden Phillpotts, Ernest Poole, Arthur Ransome, Ernest Raymond, Henry Handel Richardson, Mary Roberts Rinehart, Sax Rohmer, Saki, Ruth Sawyer, Booth Tarkington, Mark Twain, Edgar Wallace, Hugh Walpole, Mary Webb, Jean Webster, H.G. Wells, Rebecca West, Edith Wharton, P.G. Wodehouse, Virginia Woolf and E.H. Young. The complete list of featured authors is included here, not merely for the sake of consistency, but also as a reminder of the great variety of genres for which novelistic prose is employed. The composition of the Novelist Prose Corpus is meant to reflect this variety, which is why a modernist masterpiece such as Joyce’s Ulysses (1922) is featured alongside forgotten fantasy novels such as Merritt’s The Metal Monster (1920) and hardboiled detective stories such as

12 Arthur Quiller-Couch (ed.), The Oxford Book of English Verse (Oxford: At the Clarendon Press, 1925): ix. 239

Rohmer’s The Yellow Claw (1915). It is the fact that a given work is composed in prose (as well its year of publication) that makes it eligible for inclusion in the Novelistic Prose Corpus, not its literary merits. A full list of the featured novels and short stories is provided in Appendix D. 240

Appendix B: The Metrical Verse Corpus

Please note that texts marked with an asterisk are not featured in their entirety.

GEORGIAN POETRY 1 (1912) Words Gordon Bottomley The End of the World 514 Babel: The Gate of the God 701 Rupert Brooke Town and Country 242 William H. Davies The Child and the Mariner* 676 John Masefield Biography* 676 Harold Monro Lake Leman 383 James Stephens In the Cool of Evening 132 The Lonely God* 676 Total 4000

GEORGIAN POETRY 2 (1915) Rupert Brooke The Great Lover 592 Clouds 113 The Soldier 115 Walter de la Mare Music 93 John Drinkwater The Carver in Stone* 1137 Wilfred Wilson Gibson The Gorse 500 Francis Ledwidge The Wife of Llew 103 The Lost Ones 104 John Masefield The ‘Wanderer’* 1137 Total 4000

GEORGIAN POETRY 3 (1917) W.J. Turner Ecstasy 284 James Stephens The Turn of the Road 125 Siegfried Sassoon The Dragon and the Undying 109 ‘They’ 107 ‘In the Pink’ 154 Haunted 345 The Death-Bed 331 Robert Nichols To— 99 Harold Monro Week-End* 240 John Masefield Poem I 109 Poem II 117 Poem III 107 Poem IV 110 Poem V 113 Robert Graves It’s a Queer Time 259 Wilfred Wilson Gibson Rupert Brooke 110 Tenants 112 Sea-Change 100 John Freeman Happy Is England Now 193 John Drinkwater The Midlands 388 Reciprocity 77 William H. Davies The White Cascade 67 Gordon Bottomley Atlantis 218 Herbert Asquith The Volunteer 126 Total 4000 241

GEORGIAN POETRY 4 (1919) Lascelles Abercrombie Witchcraft: New Style* 85 Gordon Bottomley Littleholme* 82 Francis Brett Young Lochanilaun 111 William H. Davies Lovely Dames 121 When Yon Full Moon 137 Birds 177 Oh, Sweet Content! 104 England 209 Walter de la Mare The Tryst 195 John Freeman The Wakers 193 The Herd 218 Wilfred Wilson Gibson Wings 102 The Cakewalk 108 Driftwood 100 Quiet 111 Robert Nichols The Sprig of Lime* 82 Seventeen* 82 Siegfried Sassoon Sick Leave 106 Banishment 120 Repression of War Experience* 82 The Portrait 198 Thrushes 76 Edward Shanks In Absence 118 The Cataclysm 109 Fredegond Shove A Dream in Early Spring 141 The World* 82 J.C. Squire Sonnet 107 The Birds* 82 W.J. Turner Silence 218 The Princess 150 Peace 194 Total 4000

GEORGIAN POETRY 5 (1922) Lascelles Abercrombie Ryton Firs 300 Martin Armstrong The Buzzards 236 Edmund Blunden The Poor Man’s Pig 113 Almswomen* 166 Perch-Fishing* 163 The Giant Puffball 225 April Byeway* 163 William H. Davies The Captive Lion 98 A Bird’s Anger 97 John Drinkwater Persuasion* 163 John Freeman Moon-Bathers 158 In Those Old Days 204 Change 109 Wilfred Gibson Fire 97 Barbara Fell 105 Philip and Phoebe Ware 127 Robert Graves The Pier-Glass* 163 Richard Hughes Vagrancy* 163 William Kerr Past and Present 94 242

The Audit 114 The Apple Tree 91 Harold Monro Unknown Country* 163 Robert Nichols Night Rhapsody* 163 Frank Prewett To My Mother in Canada 168 V. Sackville-West Sailing Ships* 163 Evening 92 Edward Shanks The Rock Pool 102 Total 4000 243

Appendix C: The Free Verse Corpus xxxxxxxxxxxxxxxxxxxxxxxxxxx DES IMAGISTES Words Richard Aldington Choricos, To a Greek Marble, Au Vieux Jardin, Lesbia, Beauty Thou Hast 1169 Hurt Me Overmuch, Argyria, In the Via Sestina, The River, Bromios, To Atthis H.D. Sitalkas, Hermes of the Ways I, Hermes of the Ways I, Priapus, Acon, 582 Hermonax F.S. Flint I-V 483 Skipwith Cannell Nocturnes 175 Amy Lowell In a Garden 148 William Carlos Williams Postlude 133 Ezra Pound Δώρια, The Return, After Ch’u Yuan 204 Ford Madox Hueffer In the Little Old Market-Place 535 Total 3429 SOME IMAGIST POETS Richard Aldington Childhood, The Poplar, Round-Pond, Daisy, The Faun Sees Snow For the 1426 First Time, Lemures H.D. The Pool, The Garden, Sea Lilly, Sea Iris, Sea Rose, Oread, Orion Dead 515 John Gould Fletcher The Blue Symphony, London Excursion 1493 F.S. Flint Trees, Lunch, Malady, Accident, Fragment, Houses 671 D.H. Lawrence Ballad of Another Ophelia, Illicit 387 Amy Lowell Venus Transiens, The Travelling Bear, The Letter, Grotesque, Bullion, 541 Solitaire Total 5033 OTHERS FOR 1919 Witter Bynner Veils, Lightning, Horses, Crystal, Nakedness, Singing, Dusk, Weariness, 494 The Moon, Leer Emanuel Carnevali The Apprentice 512 Jeanne d’Orge Joy, The Enchanted Castle, To a Fumbling Lover, The Prayer Rug, 506 Defeat, The Sewing Bee, Annabel Donald Evans Portrait of Nancy Trevors, Dinner at the Hotel De La Tigresse Verte, 925 Mary Douglass Bruiting the Beauty of the Hands of Monsieur Y. Wallace Gould Intermezzo, Vigil, En Route, To a Weakling, Communion, After 1330 Tschaikowsky Marsden Hartley ‘Local Boys and Girls, Small Town Stuff’, Salutation to a Mouse, 484 Fishmonger, The Flatterers Alfred Kreymborg Dorothy, Grasses, Dust, Indian Summer, Phallic, Initials, Poetry 795 Haniel Long Students 972 Mina Loy The Black Virginity, The Dead 445 Max Michelson Masks 689 Lola Ridge The Song of Iron I 150 William Saphier Childhood Memories, Flamingo Dreams, Sadness, Rain, Before Sunrise, 579 The Night Shift, Mood Evelyn Scott Viennese Waltz, The Death of Columbine, Pietà, Rainy Twilight, Tropic 915 Winter, Lullaby, Japanese Moon, The Naid, Night Music, Stars, Venus’ Fly Trap, Monochrome, The Red Cross, Crowds, The Long Moment, Autumn Night Marjorie Allen Seiffert Love Poems in Autumn 496 Wallace Stevens Of the Surface of Things, The Paltry Nude Starts On a Spring Voyage 199 William Carlos Williams Flowers of August I-VII 1010 Total 10501

T.S. Eliot Prufrock* 1037 244

Appendix D: The Novelistic Prose Corpus

Each of the following works is featured with 200 running words:

Joseph Conrad, The Secret Sharer, 1912 Theodore Dreiser, The Financier, 1912 Zane Grey, Riders of the Purple Sage, 1912 Willa Cather, Alexander’s Bridge, 1912 James Weldon Johnson, The Autobiography of an Ex-Colored Man, 1912 Compton Mackenzie, Carnival, 1912 Jack London, The Scarlet Plague, 1912 Jack London, A Son of the Sun, 1912 D.H. Lawrence, Sons and Lovers, 1913 Saki, When William Came, 1913 Arthur Conan Doyle, The Poison Belt, 1913 Ellen Glasgow, Virginia, 1913 Willa Cather, O Pioneers!, 1913 Baroness Orczy, El Dorado, 1913 John Fox, Jr., The Heart of the Hills, 1913 Mary Roberts Rinehart, The Case of Jennie Brice, 1913 L.M Montgomery, The Golden Road, 1913 James Joyce, Dubliners, 1914 (‘The Dead’) H.G. Wells, The World Set Free, 1914 Stephen Leacock, Arcadian Adventures with the Idle Rich, 1914 Theodore Dreiser, The Titan, 1914 Sinclair Lewis, Our Mr. Wrenn, 1914 Ford Madox Ford, The Good Soldier, 1915 Virginia Woolf, The Voyage Out, 1915 Charlotte Perkins Gilman, Herland, 1915 P.G. Wodehouse, Something New, 1915 John Buchan, The Thirty-Nine Steps, 1915 Joseph Conrad, Victory, 1915 D.H. Lawrence, The Rainbow, 1915 Jean Webster, Dear Enemy, 1915 Ruth Sawyer, The Primrose Ring, 1915 Sax Rohmer, The Yellow Claw, 1915 James Joyce, A Portrait of the Artist as a Young Man, 1916 Mark Twain, The Mysterious Stranger, 1916 Booth Tarkington, Seventeen, 1916 William Dean Howells, The Leatherwood God, 1916 Sherwood Anderson, Windy McPherson’s Son, 1916 George Moore, The Brook Kerith, 1916 E.H. Young, Moor Fires, 1916 John Buchan, Greenmantle, 1916 Arthur Ransome, Old Peter’s Russian Tales, 1916 Abraham Cahan, The Rise of David Levinsky, 1917 Joseph Conrad, The Shadow-Line, 1917 Ernest Poole, His Family, 1917 Norman Douglas, South Wind, 1917 Sinclair Lewis, The Job, 1917 245

A.A. Milne, Once on a Time, 1917 Edith Wharton, Summer, 1917 Mary Webb, Gone to Earth, 1917 Henry Handel Richardson, Australia Felix, 1917 H.C. McNeile, No Man’s Land, 1917 P.G. Wodehouse, Piccadilly Jim, 1917 Christopher Morley, Parnassus on Wheels, 1917 Rebecca West, The Return of the Soldier, 1918 Booth Tarkington, The Magnificent Ambersons, 1918 Willa Cather, My Ántonia, 1918 Katherine Mansfield, ‘Bliss’, 1918 Arnold Bennett, The Roll-Call, 1918 Edna Ferber, Cheerful – by Request, 1918 Edgar Wallace, The Man Who Knew, 1918 Virginia Woolf, Night and Day, 1919 W. Somerset Maugham, The Moon and Sixpence, 1919 James Branch Cabell, Jurgen, 1919 Sherwood Anderson, Winesburg, Ohio, 1919 Jerome K. Jerome, All Roads Lead to Calvary, 1919 Hugh Walpole, The Secret City, 1919 J.D. Beresford, The Jervaise Comedy, 1919 Christopher Morley, The Haunted Bookshop, 1919 E.M. Hull, The Sheik, 1919 John Dos Passos, Three Soldiers, 1920 F. Scott Fitzgerald, This Side of Paradise, 1920 John Galsworthy, In Chancery, 1920 D.H. Lawrence, Women in Love, 1920 Sinclair Lewis, Main Street, 1920 Edith Wharton, The Age of Innocence, 1920 D.H. Lawrence, The Lost Girl, 1920 David Lindsay, A Voyage to Arcturus, 1920 Abraham Merritt, The Metal Monster, 1920 Agatha Christie, The Mysterious Affair at Styles, 1920 Harold McGrath, The Drums of Jeopardy, 1920 Sherwood Anderson, Poor White, 1920 Aldous Huxley, Crome Yellow, 1921 H. Rider Haggard, She and Allan, 1921 F. Scott Fitzgerald, Flappers and Philosophers, 1921 Marie Corelli, The Secret Power, 1921 Georgette Heyer, The Black Moth, 1921 Rose Macauley, Dangerous Ages, 1921 Eden Phillpotts, The Grey Room, 1921 Robert Keable, Simon Called Peter, 1921 Sheila Kaye-Smith, Joanna Godden, 1921 Booth Tarkington, Alice Adams, 1921 James Joyce, Ulysses, 1922 David Garnett, Lady into Fox, 1922 Ernest Raymond, Tell England, 1922 Virginia Woolf, Jacob’s Room, 1922 Sinclair Lewis, Babbitt, 1922 246

E.E. Cummings, The Enormous Room, 1922 F. Scott Fitzgerald, The Beautiful and Damned, 1922 Frances Hodgson Burnett, The Head of the House of Combe, 1922 D.H. Lawrence, Aaron’s Rod, 1922 247

Bibliography

Adams, Stephen. (1997). Poetic Designs: An Introduction to Meters, Verse Forms and Figures of Speech (Ontario: Broadview Press). Agamben, Giorgio. (1999). The End of the Poem: Studies in Poetics, translated by Daniel Heller- Roazen (Stanford, CA: Stanford University Press). Agard, Frederick B. (1984). A Course in Romance Linguistics, Vol. 2, A Diachronic View (Washington, D.C.: Georgetown University Press). Ainger, A. C. & Wintle, H. G. (1890). An English-Latin Gradus or Verse Dictionary (London: Murray). Akehurst, F.R.P. & Judith M. Davis (eds). (1995). A Handbook of the Troubadours (Berkeley & Los Angeles, CA: University of California Press). Allen, Gay Wilson & Ed Folsom (eds). (1995). Walt Whitman & the World (Iowa City, IA: University of Iowa Press). Allen, W. Sidney. (1968). Vox Graeca: A Guide to the Pronunciation of Classical Greek (Cambridge: Cambridge University Press). Arber, Edward. (1897). Tottel’s Miscellany, English Reprints (Westminter: A. Constable). Arkell, David. (1979). Looking for Laforgue: An Informal Biography (Manchester: Carcanet). Aspland, C.W. (ed.). (1979). A Medieval French Reader (Oxford: Oxford at the Clarendon Press). Asselineau, Roger. (1995). ‘Whitman in France and Belgium’. In Allen & Folsom (eds) (1995), 233-244. Attridge, Derek. (1974). Well-Weighed Syllables: Elizabethan Verse in Classical Metres (Cambridge: Cambridge University Press). Attridge, Derek. (1982). The Rhythms of English Poetry (New York: Longman). Attridge, Derek. (1990). ‘Rhythm in English Poetry’. In New Literary History, Vol. 21, No. 4 (Autumn, 1990), 1015-1037. Attridge, Derek. (1995). Poetic Rhythm: An Introduction (Cambridge: Cambridge University Press). Aubrey, Elizabeth. (1996). The Music of the Troubadours (Bloomington, IN: Indiana University Press). Augustine. (1991). Confessions, translated by Henry Chadwick (Oxford: Oxford University Press). Ayres-Bennett, Wendy. (1996). A History of the French Language through Texts (London: Routledge). Baker, David (ed.). (1996). Meter in English: A Critical Engagement (Fayetteville, AR: The University of Arkansas). Bakhtin, M.M. (1981). The Dialogic Imagination: Four Essays, edited by Michael Holquist, translated by Caryl Emerson & Michael Holquist (Austin, TX: University of Texas Press). Barber, Charles. (1996). Early Modern English (Edinburgh: Edinburgh University Press). Barber, Nicholas & Charles Barber. (1990-91). ‘The Versification of the Canterbury Tales: A Computer-Based Statistical Study’. In Leeds Studies in English, 21 (1990), 81-103, & 22 (1991), 57-84. Battistella, Edwin L. (1990). Markedness: The Evaluative Superstructure of Language (Albany, NY: State University of New York Press). Beare, William. (1957). Latin Verse and European Song. A Study in Accent and Rhythm (London: Methuen). Becker, Ph. Aug. (1888). Zur Geschichte der Vers libres in der neu-französischen Poesie (Halle: Erhard Karras). 248

Berlin, Isaiah. (2000). Three Critics of the Enlightenment: Vico, Hamann, Herder (Princeton, NJ: Princeton University Press). Bermann, Sandra & Michael Wood (eds). (2005) Nation, Language, and the Ethics of Translation (Princeton, NJ: Princeton University Press). Bernhardt-Kabisch, Ernest. (2003). ‘“When Klopstock England Defied”: Coleridge, Southey, and the German/English Hexameter’. In Comparative Literature, Vol. 55, No. 2 (Spring, 2003), 130-163. Bernhart, A. Walter. (1974). ‘Generative Metrics’. In Poetics 12 (April, 1974). Beyers, Chris. (2001). A History of Free Verse (Fayetteville, AR: The University of Arkansas Press). Blackwell, Thomas. (1735). An Enquiry into the Life and Writings of Homer. (London). Blake, William. (1977). The Complete Poems, edited by Alicia Ostriker (London: Penguin Books). Blake, William. (1988). The Complete poetry and Prose, Newly revised edition, edited by David V. Erdman (New York: Doubleday). Bliss, Alan J. (1958). The Metre of Beowulf (Oxford: Blackwell, 1958). Bliss, Alan J. (1962). An Introduction to Old English Metre (Oxford: Blackwell). Bloch, R. Howard. (2003). The Anonymous Marie de France (Chicago, IL: The University of Chicago Press). Bloom, Harold (ed.). (2007). T.S. Eliot’s The Waste Land, Updated Edition (New York: Chelsea House, 2007). Booth, Howard H. & Nigel Rigby (eds). (2000). Modernism and Empire (Manchester: Manchester University Press). Borrow, George. (1923-24). The Works of George Borrow, edited by Clement Shorter, Vol. VI (London & New York: Constable & Co. & Gabriel Wells). Boswell, James. (2008). The Life of Samuel Johnson, edited with an introduction by David Womersley (London: Penguin). Bowers, Fredson (ed.) (1976). The Dramatic Works in the Beaumont and Fletcher Canon, Vol. III (Cambridge: Cambridge University Press). Bradley, A.C. (1919). Shakespearean Tragedy, Second Edition (London: MacMillan and Co.). Bridges, Robert. (1921). Milton’s Prosody with a Chapter on Accentual Verse & Notes, Revised, Final Edition (Oxford: Oxford at the Clarendon Press). Bridges, Robert. (1922). ‘A Paper on Free Verse’. In The North American Review, Vol. 216, No. 804 (Nov., 1922), 647-658. Brogan, T.V.F. (1981). English Versification, 1570-1980: A Reference Guide with a Global Appendix (Baltimore, MD: Johns Hopkins University Press). Brooke-Rose, Christine. (1971). A ZBC of Ezra Pound (Berkeley & Los Angeles, CA: University of California Press). Brower, Reuben A. (ed.). (1971). Twentieth-Century Literature in Retrospect (Cambridge, MA: Harvard University Press). Bysshe, Edward. [1702]. The Art of English Poetry (Menston: Scolar Press, 1968). Bysveen, Josef. (1982). Epic Tradition and Innovation in James Macpherson’s Fingal (Uppsala). Cable, Thomas M. (1974). The Meter and Melody of ‘Beowulf’ (Urbana, IL: University of Illinois Press). Carne-Ross, D.S. (1967). ‘New Metres for Old: A Note on Pound’s Metric’, in Arion, Vol. 6, No. 2 (Summer, 1967), 216-232. Carr, Helen. (2000). ‘Imagism and Empire. In Booth & Rigby (eds) (2000), 64-92: 70. Carr, Helen. (2009). The Verse Revolutionaries: Ezra Pound, H.D. and The Imagists (London: Jonathan Cape). 249

Chambers, E.K. (1930). William Shakespeare: A Study of Facts and Problems, Vol. I (London: Oxford University Press). Chaucer, Geoffrey. (1972). The Complete Works of Geoffrey Chaucer, edited by Walter W. Skeat [1894], Second edition [1899] (Oxford: Oxford University Press). Chaucer, Geoffrey. (1987). The Riverside Chaucer, Third edition, edited by Larry D. Benson (Oxford: Oxford University Press). Churchill, Suzanne W. (2005). ‘The Lying Game: Others and the Great Spectra Hoax of 1917’. In American Periodicals, Vol. 15, No. 1, Little Magazine and Modernism (2005), 23-41. Cline, Lynn. (2007). Literary Pilgrims: The Santa Fe and Taos Writers’ Colonies 1917-1950 (Albuquerque, NM: University of New Mexico Press). Cole, Thomas. (1969). ‘The Saturnian Verse’. In Studies in Latin Poetry, Yale Classical Studies, 21 (London: Cambridge University Press: 1969), 1-73. Coleridge, Samuel Taylor. (1983). The Collected Works, Vol. 7, Biographia Literaria, edited by James Engell & W. Jackson Bate (Princeton, NJ: Princeton University Press). Coleridge, Samuel Taylor. (2001). The Collected Works, Vol. 16, Poetical Works, edited by J.C.C. Mays (Princeton, NJ: Princeton University Press). Colish, Marcia L. (1990). The Stoic Tradition from Antiquity to the Early Middle Ages (Leiden: E.J. Brill). Collecott, Diana. (1999). H.D. & Sapphic Modernism (Cambridge: Cambridge University Press). Congreve, William. (1706). ‘A Discourse on the Pindarique Ode’ (London). Cooper, G. Burns. (1998). Mysterious Music: Rhythm and Free Verse (Stanford, CA: Stanford University Press). Cowley, Abraham. (1905). Poems, edited by A.R. Waller (Cambridge: Cambridge University Press). Crane, Stephen. (1975). Poems and Literary Remains, edited by Fredson Bowers (Charlottesville, VA: University Press of Virginia). Creaser, John. (2007). ‘“Service is Perfect Freedom”: Paradox and Prosodic Style in Paradise Lost’. In The Review of English Studies, New Series, Vol. 58, No. 235 (June, 2007), 268-315). Croce, Benedetto. (1967). Poesia popolare e poesia d'arte. Studi sulla poesia italiana dal Tre al Cinquecento (Bari: Laterza). Crombie, Winifred. (1987). Free Verse and Prose Style: An Operational Definition and Description (London: Croon Helm). Cuddon, J.A. (1991). Penguin Dictionary of Literary Terms and Literary Theory, Third Edition (London: Penguin, 1991). Culler, A. Dwight. (1948). ‘Edward Bysshe and the Poet’s Handbook’. In PMLA, Vol. 63, No. 3 (Sep. 1948), 858-885. Cureton, Richard D. (1992). Rhythmic Phrasing in English Verse, English Language Series No. 18 (London & New York: Longman). Curtius, Ernst Robert. (1953). European Literature in the Latin Middle Ages, translated by Willard R. Trask (London & Henley: Routledge & Kegan). Cushman, Stephen. (1985). William Carlos Williams and the Meanings of Measure (New Haven & London: Yale University Press, 1985. Dalgarno, Emily. (2001). Virginia Woolf and the Visible World (Cambridge: Cambridge University Press). Danielson, Dennis. (2008). Paradise Lost by John Milton, Parallel Prose Edition (Vancouver: Regent College Publishing). Dante Alighieri (1939). The Divine Comedy, Vol. 2, Inferno, Italian text with translation and comment by John D. Sinclair (New York: Oxford University Press). 250

Dante Alighieri. (1980). The Divine Comedy, translated by C.H. Sisson (Oxford: Oxford University Press). Doolittle, Hilda. (1925). Collected Poems of H.D. (New York: Boni and Liveright). Doolittle, Hilda. (1979). End to Torment: A Memoir of Ezra Pound (New York: New Directions Books). Doolittle, Hilda. (1983). Bid Me to Live (New York: Dial Press). Draugsvold, Ottar G. (ed.). (2000). Nobel Writers on Writing (Jefferson, NC: McFarland & Co.). Dryden, John. (1961). The Poems and Fables, edited by James Kinsley (London: Oxford University Press). Dryden, John. (2007). Selected Poems, edited by Paul Hammond & David Hopkins. Duff, William. (1767). Essay on Original Genius (London). Duffell, Martin J. (2000). ‘“The Craft So Long to Lerne”: Chaucer’s Invention of the Iambic Pentameter’. In The Chaucer Review, Vol. 34, No. 3 (2000), 269-288. Duffell, Martin J. (2005). ‘Some Phonological Features of Insular French: A Reconstruction’. In Wright & Ricketts (eds) (2005): 103-125. Duffell, Martin J. (2008). A New History of English Metre (London: Modern Humanities Research Association and Maney Publishing). Duffell, Martin J. & Dominique Billy. (2004). ‘From Decasyllable to Pentameter: Gower’s Contribution to English Metrics’. In The Chaucer Review, Vol. 38, No. 4 (2004), 383-400. Duggan, Hoyt N. (1996). ‘Libertine Scribes and Maidenly Editors: Meditations on Textual Criticism and Metrics’. In McCully & Anderson (eds) (1996), 219-237. Duncan, Robert. (2011). The H.D. Book, edited and with an Introduction by Michael Boughn & Victor Coleman (Berkeley & Los Angeles, CA: University of California Press). Easthope, Anthony. (1983). Poetry as Discourse (London: Methuen). Edwards, A.S.G. (1983). Stephen Hawes (Boston, MA: Twayne Publishers). Eliot, T.S. (1920). The Sacred Wood (London: Methuen). Eliot, T.S. [1921] ‘The Metaphysical Poets’. In Keast (ed.) (1962), 22-30. Eliot, T.S. [1928]. ‘Introduction’ to Ezra Pound, Selected Poems, edited with an Introduction by T.S. Eliot (London: Faber and Faber, 1948): 8. Eliot, T.S. (1934). After Strange Gods (London: Faber and Faber). Eliot, T.S. (1957). On Poetry and Poets (London: Faber and Faber). Eliot, T.S. (1969). To Criticize the Critic (London: Faber and Faber). Eliot, T.S. (1996). Inventions of the March Hare: Poems 1909-1917, edited by Christopher Ricks (New York: Harcourt Brace & Company, 1996). Eliot, T.S. (2004). The Complete Poems and Plays (London: Faber and Faber). Elleman, Barbara. (2002). Virginia Lee Burton: A Life in Art (New York: Houghton Mifflin Harcourt). Epstein, Edmund L. & Terence Hawkes. (1959). Linguistics and English Prosody (Buffalo, NY: University of Buffalo Department of Anthropology and Linguistics). Everdell, William R. (1997). The First Moderns (Chicago, IL: Chicago University Press). Fabb, Nigel & Morris Halle. (2008). Meter in Poetry: A New Theory (Cambridge: Cambridge University Press). Fennell, Barbara A. (2006). A History of English: A Sociolinguistic Approach (Malden, MA & Oxford: Blackwell Publishing). Finch, Annie. (1990). The Ghost of Metre: Culture and Prosody in American Free Verse (Ann Arbor, MI: University of Michigan Press). Fisiak, Jacek. (ed.). (1984). Historical Syntax, Trends in Linguistics: Studies and Monographs 23 (Berlin: Walter de Gruyter & Co.). 251

Foerster, Donald M. (1969). Homer in English Criticism. The Historical Approach in the Eighteenth Century (Hamden, CT: Archon Books). Fortson IV, Benjamin W. (2010). Indo-European Language and Culture: An Introduction, Second edition (London: Wiley-Blackwell). Gaimar, Geffrei. (2009). Estoire des Engleis, edited and translated by Ian Short (Oxford: Oxford University Press). Gascoigne, George. ‘Certayne Notes of Instruction concerning the making of verse or ryme in English’ [1575], in Vickers (ed.) (1999): 162-171. Gaskill, Howard. (ed.). (2004). The Reception of Ossian in Europe (London & New York: Thoemmes Continuum). Gasparov, M.L. (1996). A History of European Versification, translated by G.S. Smith & Marina Tarlinskaja, edited by G.S. Smith with Leofranc Holford-Strevens (Oxford: Clarendon Press). Gerritsen, Marinel. (1984). ‘Divergent Word Order Developments in Germanic Languages’. In Fisiak (ed.) (1984), 107-135. Getty, Paul. (2002). The Metre of ‘Beowulf’: A Constraint-Based Approach (Berlin & New York: Mouton de Gruyter). Gildon, Charles. (1718). The Complete Art of Poetry, Vol. I (London). Gillingham, John. (1997). ‘Kingship, Chivalry and Love. Political and Cultural Values in the Earliest History Written in French: Geoffrey Gaimar’s Estoire des Engleis’. In Hollister (ed.) (1997), 33-58. Golston, Michael. (2008). Rhythm and Race in Modernist Poetry and Science (New York: Columbia University Press). Grant, Joy. (1967). Harold Monro & the Poetry Bookshop (Berkeley & Los Angeles, CA: University of California Press). Gray, Thomas. (1955). Gray’s Poems, Letters and Essays (London: J.M. Dent & Sons). Gregory, Eileen. (1997). H.D. and Hellenism: Classic Lines (Cambridge: Cambridge University Press). Gross, Harvey (ed.). (1979). The Structure of Verse. Modern Essays on Prosody, Revised edition (New York: The Ecco Press). Groves, Peter L. (1999). ‘The Chomsky of Grub Street: Edward Bysshe and the Triumph of Classroom Metrics’. In Versification [An Electronic Journal of Literary Prosody], Vol. 3 (1999). Groves, Peter L. (2005). ‘Finding his Feet: Wyatt and the Founding of English Pentameter’. In Versification [An Electronic Journal of Literary Prosody], Vol. 4 (2005). Groves, Peter L. (2011). ‘“Opening” the Pentameter: Hopkins’s Metrical Experimentation’. In The Hopkins Quarterly, 38, 1-2 (Winter/Spring, 2011), 93-110. Gummere, Francis B. (1886). ‘The Translation of Beowulf, and the Relations of Ancient and Modern English Verse’. In The American Journal of Philology, Vol. 7, No. 1 (1886). Hajičovà, Eva et al (eds). (1995). Prague Linguistic Circle Papers, Vol. 1 (Amsterdam: John Benjamins). Halle, Morris & Samuel Jay Keyser. (1971). English Stress: Its Form, Its Growth, and Its Role in Verse (New York: Harper & Row). Hammerstein II, Oscar. (1985). Lyrics, edited by William Hammerstein (Milwaukee, WI: Hal Leonard Books). Hanford, James Holly. (1919). ‘Coleridge as a Philologian’. In Modern Philology, Vol. 16, No. 12 (Apr., 1919), 615-636. 252

Hardison Jr., O.B. (1984). ’Blank Verse before Milton’. In Studies in Philology, Vol. LXXXVI, No. 3 (Summer, 1984), 253-274. Harmer, J. B. (1975). Victory in Limbo: A History of Imagism, 1908-1917 (New York: St. Martin’s). Harmon, William. (1997). ‘English Versification: Fifteen Hundred Years of Continuity and Change’. In Studies in Philology, Vol. XCIV, No.1 (Winter, 1997), 1-37. Hartley, Marsden. (1997). Somehow a Past: Prologue to Imaginative Living, edited with an Introduction by Susan Elizabeth Ryan (Cambridge, MA: M.I.T. Press). Hartman, Charles O. (1980). Free Verse: An Essay on Prosody (Princeton, NJ: Princeton University Press). Hartman, Herbert (ed.). (1933). Surrey’s Fourth Boke of Virgill (London & New York: Oxford University Press). Haslewood, Joseph. (1815). Ancient Critical Essays upon English Poets and Poësy (London: Robert Triphook). Haugen, Kristine Louise. (1998). ‘Ossian and the Invention of Textual History’. In Journal of the History of Ideas, Vol. 59, No. 2 (Apr., 1998), 309-327. Haven, Cynthia L. (2002) Joseph Brodsky: Conversations (Jackson, MS: University Press of Mississippi). Hayes, Bruce. (1995). Metrical Stress Theory: Principles and Case Studies (Chicago, IL: University of Chicago Press, 1995). Herzman, Ronald B., Graham Drake & Eve Salisbury (eds). (1999). Four Romances of England (Kalamazoo, MI: Medieval Institute Publications). Heuvel, Cor van den. (1986). The Haiku Anthology, 2nd edition (New York: Simon & Schuster). Hindus, Milton (ed.). (1971). Walt Whitman: The Critical Heritage (London: Routledge and K. Paul). Holder, Alan. (1963). ‘The Lesson of the Master: Ezra Pound and Henry James’. In American Literature, Vol. 35, No. 1 (March, 1963), 71-79. Holder, Alan. (1964). ‘T. S. Eliot on Henry James’. In PMLA, Vol. 79, No. 4 (Sep., 1964), 490-497. Holder, Alan. (1995). Rethinking Meter: A New Approach to the Verse Line (Lewisburg, PA: Bucknell University Press / London: Associated University Presses). Hollander, John. (1956). ‘The Music of Poetry’. In The Journal of Aesthetics and Art Criticism, Vol. 15, No. 2 (Dec., 1956), 232-244. Hollander, John. (1975). Vision and Resonance: Two Senses of Poetic Form (New York: Oxford University Press). Hollister, C. Warren (ed.). (1997). Anglo-Norman Political Culture and the 12th-Century Renaissance (Woodbridge, The Boydell Press). Holmes, Anne. (1993). Jules Laforgue and Poetic Innovation (Oxford: Clarendon Press). Hopkins, Gerard Manley. (2002). The Major Works, edited with an Introduction and Notes by Catherine Phillips (Oxford: Oxford University Press). Houston, John Porter. (1988). Shakespearean Sentences: A Study in Style and Syntax (Baton Rouge, LA: Louisiana State University Press). Hrushovski, Benjamin. (1960). ‘On Free Rhythms in Modern Poetry’. In Sebeok (ed.) (1960), 173- 190. Hueffer, Ford Madox. (1921). Thus to Revisit (New York: Dutton). Hulle, Dirk Van & Wim Van Mierlo (eds). (2004). Variants 2/3: Reading Notes (Amsterdam & New York: Editions Rodopi). Hulst, Harry van der (ed.). (1999). Word Prosodic Systems in the Languages of Europe (Berlin & New York: Mouton de Gruyter). 253

Hurd, Richard. (1767). Letters on Chivalry and Romance (London). Hurley, Michael D. (2006). ‘What Sprung Rhythm Really Is NOT’. In The Hopkins Quarterly 33 (2006), 71-94. Ingram, John K. (1874). ‘On the “Weak Endings” of Shakspere with Some Account of the History of the Verse-Tests in General’. In The New Shakspere Society’s Transactions (London: Trubner & Co., 1874), 442-464. Jakobson, Roman. [1960] ‘Linguistics and Poetics’. In Roman Jakobson, Language in Literature, edited by Krystyna Pomorska & Stephen Rudy (Cambridge, MA & London: The Belknap Press of Harvard, 1987), 62-94. Jakobson, Roman. (1972). ‘Verbal Communication’. In Scientific American 227 (1972), 72-80. Jakobson, Roman. (1980). ‘Subliminal Verbal Patterning in Poetry’, in Poetics Today, Vol. 2, No. 1a (Autumn, 1980), 127-136. Jakobson, Roman & Linda R. Waugh. (1979). The Sound Shape of Language (Bloomington, IN & London: Indiana University Press). James I of Scotland. (1971). The Kingis Quair, with an Introduction, Notes and Glossary by John Norton-Smith (Oxford: Oxford University Press). Jespersen, Otto. (1933). ‘Notes on Metre’. In Gross (ed.) (1979), 105-128. Johnson, Samuel. (1905). Lives of the English Poets, edited by George Birkbeck Hill (Oxford: At the Clarendon Press). Jones, P.M. (1916). ‘Influence of Walt Whitman on the Origin of the “Vers Libre”’. In Modern Language Review, 11 (1916), 186-194. Kastner, L.E. (1903). A History of French Versification (Oxford: At the Clarendon Press). Kay, Dennis. (1984). ‘Wyatt and Chaucer: They Fle From Me Revisited’. In Huntington Library Quarterly, 47 (1984), 211-225. Keast, William R. (ed.). (1962). Seventeenth-Century English Poetry: Modern Essays in Criticism (New York: Oxford University Press). Keller, Stefan Daniel. (2009). The Development of Shakespeare’s Rhetoric: A Study of Nine Plays (Tübingen: Francke Verlag). Ker, W.P. [1912]. English Literature: Medieval (London: Williams & Norgate, 1926). Kidwell, Carol. (2004). Pietro Bembo: Lover, Linguist, Cardinal (Montreal, QC: McGill-Queen’s University Press). Kiparsky, Paul. (1975). ‘Stress, Syntax, and Meter’. In Language, Vol. 51, No. 3 (Sep., 1975), 576- 616. Kirby-Smith, H.T. (1996). The Origins of Free Verse (Ann Arbor, MI: University of Michigan Press). Krier, Theresa M. (ed.). (1998). Refiguring Chaucer in the Renaissance (Gainesville, FL: University Press of Florida). Kwasny, Melissa (ed.). (2004). Toward the Open Field (Middleton, CT: Wesleyan University Press, 2004). Lapidge, Michael. (1979). ‘Aldhelm’s Latin Poetry and Old English Verse’. In Comparative Literature, Vol. 31, No. 3 (Summer, 1979), 209-231. Laskaya, Anne & Eve Salisbury (eds). (1995). The Middle English Breton Lays (Kalamazoo, MI: Medieval Institute Publications). Laver, John. (1994). Principles of Phonetics (Cambridge: Cambridge University Press). Lawrence, D.H. (1956). Selected Literary Criticism, edited by Anthony Beal (London: Heinemann). Lawrence, D.H. (1964). Complete Poems, collected and edited with an Introduction and Notes by Vivian de Sola Pinto & Warren Roberts (London: Heinemann). Lennard, John. (1996). The Poetry Handbook (Oxford: Oxford University Press). 254

Lefevere, André (ed.). (1992). Translation, History, Culture: A Sourcebook (London: Routledge, 1992). Lewis, C.S. (1960). A Preface to Paradise Lost (London: Oxford University Press). Lewis, C.S. (1965). English Literature in the Sixteenth Century (Oxford: Oxford University Press). Lock, Charles. (2006). ‘Heterographics: Towards a History and Theory of Other Lettering’. In Angles on the English-Speaking World, New Series, Vol. 6 (2006), 97-112. Lock, Charles. (2009). ‘Inhibiting the Voice: Thomas Hardy and Modern Poetics’. In Keith Wilson (ed.) (2009), 450-464. Longenbach, James. (2004). The Resistance to Poetry (Chicago & London: University of Chicago Press). Longenbach, James. (2007). The Art of the Poetic Line (St. Paul, MN: Graywolf). Lowth, Robert. [1787]. Lectures on the Sacred Poetry of the Hebrews, translated by George Gregory, New Edition with Notes by Calvin E. Stowe, Vol. II (Andover, MA: Codman Press, 1829). Ludington, Townsend. (1998). Marsden Hartley: The Biography of an American Artist (New York: Cornell University Press). Lydgate, John. (1998). Troy Book: Selections, edited by Robert R. Edwards (Kalamazoo, MI: Western Michigan University Press). Maas, Paul. [1903]. ‘Der byzantinische Zwölfsilber’. In Paul Maas, Kleine Schriften (München: C.H. Beck’sche Verlangsbuchhandlung: 1973), 242-288. Mallarmé, Stéphane. (1988). Selected Letters, edited and translated by Rosemary Lloyd (Chicago, IL: University of Chicago Press). Marlowe, Christopher. (1981). Tamburlaine the Great, edited by J.S. Cunningham (Manchester: Manchester University Press, 1981). Mason, John. (1761). Two Essays on the Power of Numbers (London). Maurer, Oscar. (1952). ‘Swinburne vs. Furnivall: A Case Study in “Aesthetic” vs. “Scientific” Criticism’. In The University of Texas Studies in English, Vol. 31 (1952), 86-96. McCully, C.B. & J.J. Anderson (eds). (1996). English Historical Metrics (Cambridge: Cambridge University Press). McGann, Jerome. (1993). Black Riders: The Visible Language of Modernism (Princeton, NJ: Princeton University Press). Miller, James E. (2007). ‘Personal Mood Transmuted into Epic: T.S. Eliot’s “Waste Land”’. In Bloom (ed.) (2007), 49-74. Milton, John. (1957). Complete Poems and Major Prose, edited by Merritt Y. Hughes (New York: Odyssey Press). Milton, John. (2000). Paradise Lost, edited with an Introduction and Notes by John Leonard (London: Penguin). Minkova, Donka. (1991). The History of Final Vowels in English: The Sound of Muting (Berlin & New York: Mouton de Gruyter). Moore, Dafydd. (2004). ‘The Reception of The Poems of Ossian in England and Scotland’. In Gaskill (ed.) (2004), 21-39. Munro, John M. (1969). Arthur Symons (New York: Twayne Publishers). Murphy, James J. (1974). Rhetoric in the Middle Ages: A History of Rhetorical Theory from Saint Augustine to the Renaissance (Berkeley, Los Angeles & London: University of California Press). Nadel, Ira B. (ed.). (2010). Ezra Pound in Context (Cambridge: Cambridge University Press). Ostler, Nicholas. (2007). Ad Infinitum: A Biography of Latin (London: Harper Press). 255

Overton, Bill. (2008). ‘“People Have Forgotten How to Hear the Music”: The Teaching of Poetry and Prosody’, in English (2008), 57 (219): 266-282. Patterson, Anita. (2008). Race, American Literature and Transnational Modernisms (Cambridge: Cambridge University Press). Peabody, Berkley. (1975). The Winged Word: A Study in the Technique of Ancient Greek Oral Composition As Seen Principally through Hesiod’s Works and Days (New York: State University of New York Press). Pearsall, Derek. (1992). The Life of Geoffrey Chaucer: A Critical Biography (Oxford: Oxford University Press). Pearsall, Derek (ed.). (1999). Chaucer to Spenser: An Anthology (Oxford: Blackwell). Philips, John. (1762). Poems Attempted in the Style of Milton (London). Pondrom, Cyrena N. (1974). The Road from Paris: French Influence on English Poetry 1900-1920 (Cambridge: Cambridge University Press). Pope, Alexander. (1715). The Iliad of Homer (London). Pope, Alexander. (2006). The Major Works, edited by Pat Rogers (Oxford: Oxford University Press). Pope, Mildred K. [1932]. From Latin to Modern French (Manchester: Manchester University Press, 1973). Posner, Rebecca. (1997). Linguistic Change in French (Oxford: Oxford University Press). Pound, Ezra. (1914). ‘The Tradition’, in Poetry, Vol. 3, No. 4 (1914), 137-141. Pound, Ezra. (1915). ‘The Renaissance: I – The Palette’. In Poetry, Vol. 5, No. 5 (Feb., 1915), 227- 234. Pound, Ezra. (1918). Pavannes and Divisions (New York: Alfred A. Knopf). Pound, Ezra. (1951). ABC of Reading (London: Faber and Faber). Pound, Ezra. (1962). Patria Mia and the Treatise on Harmony (London: Peter Owen). Pound, Ezra. (1970). Gaudier-Brzeska: A Memoir (London: New Directions). Pound, Ezra. (1971). Selected Letters 1907-1941, edited by D.D. Paige (New York: New Directions). Pound, Ezra. (1996). The Cantos of Ezra Pound (New York: New Directions). Pound, Ezra. (2001). Personæ. Collected Shorter Poems, edited by Lea Baechler & A. Walton Litz (London: Faber and Faber). Preminger, Alex & T.V.F. Brogan (eds). (1993). The New Princeton Encyclopedia of Poetry and Poetics (Princeton, NJ: Princeton University Press). Price, Kenneth M. (1996). Walt Whitman: The Contemporary Reviews (Cambridge: Cambridge University Press, 1996). Prins, Yopie. (2000). ‘Nineteenth-Century Homers and the Hexameter Mania’. In Bermann & Wood (eds) (2005), 229-256. Purcell, Henry. (2000). The Fairy Queen in Full Score (Toronto: Dover Publications). Puttenham, George. [1589]. The Arte of English Poesie (Menston: Scolar Press, 1968). Quiller-Couch, Arthur (ed.). (1925). The Oxford Book of English Verse (Oxford: At the Clarendon Press). Randall, Huberta Frets. (1942). ‘Whitman and Verhaeren – Priests of Human Brotherhood’. In The French Review, Vol. 16, No. 1 (Oct., 1942), 36-43. Raven, D.S. (1965). Latin Metre: An Introduction (London: Faber and Faber). Reck, Michael. (1968). Ezra Pound, A Close-Up (New York: McGraw-Hill). Ricks, Christopher. (1963). Milton’s Grand Style (London: Oxford at the Clarendon Press). Roach, Peter. (2001). Phonetics (Oxford: Oxford University Press). 256

Roach, Peter. (2009). English Phonetics and Phonology: A Practical Course, 4th edition (Cambridge: Cambridge University Press). Russell, D.A. & Michael Winterbottom (eds). (1989). Classical Literary Criticism (Oxford: Oxford University Press). Russom, Geoffrey. (1987). Old English Meter and Linguistic Theory (Cambridge: Cambridge University Press). Russom, Geoffrey. (1998). Beowulf and Old Germanic Metre (Cambridge: Cambridge University Press). Saintsbury, George. (1898). A Short History of English Literature (London: Macmillan). Saintsbury, George. (1923). A History of English Prosody from the Twelfth Century to the Present Day, Second edition (London: Macmillan). Sandras, E.-G. (1859). Étude sur Chaucer (Paris: Auguste Durand). Sandys, J.E. [1908]. A History of Classical Scholarship, Vol. III (New York & London: Hafner Publishing Company, 1967). Say, Samuel. (1745). Poems on Several Occasions (London). Schick, J. (ed.). (1891). Lydgate’s Temple of Glas (London). Sebeok, Thomas A. (ed.). (1960). Style in Language (Cambridge, MA: M.I.T. Press). Shannon III, Richard Stoll. (1975). The Arms of Achilles and Homeric Compositional Technique (Leiden: E.J. Brill). Shaw, Robert B. (2007). Blank Verse: A Guide to Its History and Use (Athens, OH: Ohio University Press). Shelley, P.B. (1904). The Complete Poetical Works, edited by Thomas Hutchinson (Oxford: At the Clarendon Press). Sherbo, Arthur. (1967). Christopher Smart: Scholar of the University (East Lansing, MI: Michigan State University Press). Sheridan, Thomas. (1775). Lectures on the Art of Reading; Second Part: Containing The Art of Reading Verse (London: J. Dodsley et al.). Short, Ian & Brian Merrilees (eds). (1979). Benedeit: The Anglo-Norman Voyage of St Brendan (Manchester: Manchester University Press). Sievers, Eduard. (1885). ’Zur Rhythmik des Germanischen Alliterationverses’. In Pauls und Braunes Beiträge, 10 (1885). Simon, Myron. (1969). ‘The Georgian Poetic’. In The Bulletin of the Midwest Modern Language Association, Vol. 2, No. 1 (1969), 121-135. Skelton, John. (1983). The Complete English Poems, edited by John Scattergood (Harmondsworth: Penguin). Smart, George K. (1937). ‘English Non-dramatic Blank Verse in the Sixteenth Century’. In Anglia, LXI (1937), 370-397. Spiller, Michael R.G. (1992). The Development of the Sonnet: An Introduction (London: Routledge). Steadman, John M. (1964). ‘Verse without Rime: Sixteenth-Century Italian Defences of Versi Sciolti’. In Italica, Vol. 41, No. 4 (Dec., 1964), 384-402. Steele, Timothy. (1990). Missing Measures: Modern Poetry and the Revolt against Meter (Fayetteville, AR & London: University of Arkansas Press). Steele, Timothy. (1999). All the Fun’s in How You Say a Thing (Athens, OH: Ohio University Press). Stevens, Martin. (1979). ‘The Royal Stanza in Early English Literature’. In PMLA, Vol. 94, No. 1 (Jan., 1979), 62-76. Strachan, John & Richard Terry. (2000). Poetry (Edinburgh: Edinburgh University Press). 257

Swinburne, Algernon Charles. (1971). ‘Whitmania’. In Hindus (ed.) (1971), 199-207. Symonds, John Addington. (1895). Blank Verse (New York: Scribner). Symons, Arthur. (1899). The Symbolist Movement in Literature (London: William Heinemann). Tarlinskaja, Marina. (1976). English Verse: Theory and History (The Hague & Paris: Mouton). Taylor, William. (1796). ‘English Hexameter Exemplified’. In Monthly Magazine, Vol. I (London: 1796), 404-405. Thomas, Walter. (1907). ‘Milton’s Heroic Line Viewed from an Historical Standpoint’. In The Modern Language Review, Vol. 2, No. 4 (Jul., 1907): 289-315. Thompson, John. (1961). The Founding of English Metre (London: Routledge & Kegan Paul). Thorpe, James (ed.). (1951). Milton Criticism: Selections from Four Centuries (London: Routledge & Kegan Paul). Tranel, Bernard. (1987). The Sounds of French: An Introduction (Cambridge, New York: Cambridge University Press). Trask, R.L. (1996). A Dictionary of Phonetics and Phonology (London: Routledge). Traugott, Elizabeth. (1972). The History of English Syntax (New York: Holt, Rinehart and Winston). Troyes, Chrétien de. (1989). Le Chevalier de la charrette (Lancelot), Texte établi, annoté et présenté avec variantes par A. Foulet et K.D. Uitti (Paris: Bordas). Trubetzkoy, N.S. (1969). Principles of Phonology, translated by Christiane A.M. Baltaxe (Berkeley & Los Angeles: University of California Press, second printing (1971). Tsur, Reuven. (1997). ‘Poetic Rhythm: Performance Patterns and Their Acoustic Correlates’. In Versification [an electronic journal of literary prosody], Vol. 1 (1997). Tupper, Martin Farquhar. (1849). Tupper’s Proverbial Philosophy: A Book of Thoughts and Arguments, Originally Treated, First and Second Series (Milwaukie [sic], WI: I.A. Hopkins). Turco, Lewis Putnam. (1996). ‘Verse vs. Prose/Prosody vs. Meter’. In Baker (ed.) 1996, 249-263. Turco, Lewis [Putnam]. (2000). The Book of Forms: A Handbook of Poetics (Hanover, NH: University Press of New England). Valéry, Paul. (1960). Tel Quel II, Pléiade II (Paris: Gallimard). Vaughan, Henry. (1963). Poetry and Selected Prose, edited by L.C. Martin (London: Oxford University Press). Vickers, Brian (ed.). (1999). English Renaissance Literary Criticism (Oxford: Oxford University Press). Virgil. (1908). The Æneid, translated into English Verse by Theodore C. Williams (Boston & New York: Houghton Mifflin Company). Vising, Johan. (1923). Anglo-Norman Language and Literature (London: Oxford University Press). Vitz, Evelyn Birge. (1999). Orality and Performance In Early French Romance (Cambridge: D.S. Brewer). Walling, Robert Alfred John. (1908). George Borrow, The Man And His Work (London: Cassell). Warren, F.M. (1890). Review of Ph. Aug. Becker’s Zur Geschichte der Vers libres in der neu- französischen Poesie. In Modern Language Notes, No. 2 (February, 1890). Watkins, Calvert. (1963). ‘Indo-European Metrics and Archaic Irish Verse’. In Celtica, 6 (1963): 194-209. Watkins, John. (1998). ‘“Wrastling for this World”: Wyatt and the Tudor Canonization of Chaucer’. In Krier (ed.) (1998), 21-39. Webb, Daniel. (1769). Observations on the Correspondence between Poetry and Music (London). Weeks, L.T. (1910). ‘The Order of Rimes of the English Sonnet’. In Modern Language Notes, Vol. 25, No. 6 (Jun., 1910), 176-180. 258

Wesling, Donald. (1971). ‘The Prosodies of Free Verse’. In Brower (ed.) (1971), 155-187. Wesling, Donald. (1985). The New Poetries: Poetic form since Coleridge and Wordsworth (Lewisburg, PA: Bucknell University Press). Wesling, Donald. (1996). The Scissors of Meter: Grammetrics and Reading (Ann Arbor, MI: University of Michigan Press). West, M.L. (1973a). ‘Indo-European Metre’. In Glotta, 51, Bd., 3./4. H. (1973), 161-187. West, M.L. (1973b). ‘Greek Poetry 2000-700 B.C.’. In The Classical Quarterly, New Series, Vol. 23, No. 2 (Nov., 1973), 179-192. West, M.L. (1982). Greek Metre (Oxford: Oxford at the Clarendon Press). Westphal, Rudolf. (1861). Die Fragmente und die Lehrsätze der Griechischen Rhythmiker (Leipzig: Druck und Verlag von B.G. Teubner). Whalen, Logan E. (2008). Marie de France & the Poetics of Memory (Washington, D.C.: The Catholic University of America Press). White, Carolinne. (2000). Early Christian Latin Poets (London: Routledge). Whitman, Walt. (1996). The Complete Poems, edited by Francis Murphy (London: Penguin). Wilhelm, J.J. (1990). Ezra Pound in London and Paris: 1908-1925 (Pennsylvania State University Press). Williams, William Carlos. (1957). The Selected Letters of William Carlos Williams, edited by John C. Thirlwall (New York: New Directions). Williams, William Carlos. [1958]. I Wanted to Write a Poem: The Autobiography of the Works of a Poet, edited by Edith Heal (New York: New Directions: 1978). Williams, William Carlos. (1967). The Autobiography of William Carlos Williams (New York: New Directions). Williams, William Carlos. (2000). Collected Poems I: 1909-1939, edited by Walton Litz & Christopher MacGowan (Manchester: Carcanet Press). Wilson, Keith (ed.) (2009). A Companion to Thomas Hardy (Oxford: Blackwell). Wimsatt, James I. (1991). Chaucer and His French Contemporaries: Natural Music in the Fourteenth Century (Toronto: University of Toronto Press). Wood, Robert. (1775). An Essay upon the Original Genius of and Writings of Homer (London). Woolf, Virginia. (1971). The Voyage Out (London: Hogarth Press). Woolf, Virginia. (1987). The Essays of Virginia Woolf, edited by Andrew McNeillie, Vol. 2: 1912- 1918 (London & New York: Harcourt Brace Jovanovich). Wormser, Baron & David Capella. (2000). Teaching the Art of Poetry: The Moves (Mahwah, NJ: Lawrence Erlbaum Associates). Woudenberg, Maximiliaan van. (2004). ‘Coleridge’s Reading Notes in Göttingen: The Intertextual Research of the Projected Life of Lessing in 1799’. In Van Hulle & Van Mierlo (eds) (2004), 245-270. Wrenn, C.L. (1973). Beowulf and the Finnesburg Fragment, 1973 Edition, fully revised by W.F. Bolton (London: Harrap). Wright, George T. (1985). ‘Wyatt’s Decasyllabic Line’. In Studies in Philology, Vol. 82, No. 2 (Spring, 1985), 129-156. Wright, Roger & Peter Ricketts (eds). (2005). Studies on Ibero-Romance Linguistics Dedicated to Ralph Penny (Newark, DE: Juan de la Cuesta). Wyatt, Thomas. (1969). Collected Poems of Sir Thomas Wyatt, edited by Kenneth Muir & Patricia Thomson (Liverpool: Liverpool University Press). Yeats, W.B. (ed.). (1936). The Oxford Book of Modern Verse (Oxford: At the Clarendon Press). Young, Edward. (1759). Conjectures on Original Composition (London). 259

Resume Denne ph.d.-afhandling undersøger forholdet mellem metrisk vers og frie vers i engelsk poesi med særlig vægt på den modernistiske lyriks kritik af det jambiske pentameter. Dette metrum og de principper, der underligger det, dominerede engelsk poesi fra slutningen af det 14. århundrede frem til begyndelsen af det 20. århundrede, hvor en række fremtrædende digtere bevidst søgte at undslippe dets dominans. Med inspiration fra fransk symbolisme, japansk haiku og en lang række andre mere eller mindre eksotiske kilder, skabte digtere såsom Ezra Pound og T.S. Eliot på tærsklen til det tyvende århundrede en ny type engelsk poesi, der ikke i traditionel forstand kan beskrives som metrisk. Dette nybrud medførte en ophedet debat om, hvorvidt verskompositioner uden metrum overhovedet kan betragtes som digte, og det er denne debat, der danner rammen for nærværende afhandling. I modsætning til tidligere undersøgelser af debatten omkring frie vers fokuserer denne afhandling primært på den tradition, modernistisk poesi i frie vers kan betragtes som et opgør med. Den jambiske traditions historie, forudsætninger og indvirkning på det engelske sprog kortlægges, og det påvises, at metrik influerer både syntaks og diktion. Jambisk metrum er med andre ord hverken et neutralt medium eller et fænomen, der kan anskues isoleret fra et digts øvrige egenskaber og karakteristika. Både historisk og strukturelt er der god grund til at skelne mellem forskellige typer af jambisk vers: Selv inden for det urimede jambiske pentameter (blankvers) påvises der således væsentlige forskelle mellem dets egenskaber i episk og dramatisk digtning, eksemplificeret ved henholdsvis Milton og Shakespeare. Især undersøges forholdet mellem sætningsstruktur og linjering, og disse parametre anvendes også til at skelne imellem forskellige typer af frivers- kompositioner. Afhandlingen afsluttes med en kvantitativ undersøgelse af en række fonologiske forskelle mellem frie vers, metrisk vers og prosa, der antyder, at modernistisk frivers – på trods af fraværet af metrisk struktur – har væsentlige fonologiske lighedstræk med den metriske tradition, modernismen forsøgte at frigøre sig fra, og som samtidig klart adskiller frivers fra prosa. 260

Abstract This dissertation examines the relationship between metrical verse and free verse in English poetry, with particular emphasis on the modernist critique of the iambic pentameter. This metre and its underlying principles were dominant in English poetry from the late fourteenth century up until the early twentieth century, at which point a number of key poets deliberately sought to free themselves from its shackles. Seeking inspiration in French symbolism, Japanese haiku and many other exotic sources, poets such as Ezra Pound and T.S. Eliot produced a body of verse that was not metrical in the traditional sense of the word, giving rise to a heated debate concerning the exact nature and properties of verse that is free from metre. Unlike previous studies of this debate, the present study is mainly concerned with that tradition which the emergence of modernist free verse was a response to. It maps the history and the preconditions of the iambic tradition as well as its influence on the English language, and it shows that metre influences both syntax and diction in a number of ways. Iambic metre, in other words, is neither a neutral medium nor a phenomenon that can be viewed in isolation. Consequently, for reasons both historical and structural, we must distinguish between different types of iambic verse, and it is shown that unrhymed iambic pentameter (i.e. blank verse) has very different properties, depending on whether it is used for epic (Milton) or for dramatic purposes (Shakespeare). Particular attention is accorded to the relationship between syntactic structure and lineation, and these criteria are also used to distinguish between different species of free verse. The dissertation concludes with a quantitative study of a number of phonological features in free verse, metrical verse and prose; the findings of this study suggest that modernist free verse – in spite of its lack of metre – exhibits a number of phonological characteristics that are closer to metrical verse than they are to prose.