The organization of sound in “kubla khan” is impossible to ignore—and yet generations of critics have studiously done just that: ignored it. Naturally, readers have found many things to say about the poem’s abundant local effects (assonance, alliteration, rhyme, etc.), effects which are then made to consort with its famously disputed content, to more or less persuasive effect. But if by “organization of sound” we mean something more than such isolated instances, namely, the manner in which they interweave to form a particular metrical structure, or prospective generic affiliation, the critical silence is striking—all the more so given that “Kubla Khan” can make a creditable claim to being the best-thumbed poem in the vernacular.
Sometimes this silence appears to indicate the conviction that there is really nothing to say, no puzzle in need of solution. That view emerges in George Saintsbury’s reverential yet truncated account of the poem, in the third volume of his History of English Prosody:
… it is not easy to think of a greater piece of poetry than Kubla Khan, and the comparison of the opening strophe with its mother-passage in Purchas is almost a complete object-lesson in the difference between prose and poetry. But though it has not exactly the least prosodic interest, it has the least prosodic interest for us. It is, in point of form, simply an example, immensely improved in form itself, and charged with a double and tenfold portion, of the half-regular ode or lyric, the ‘broken and cuttit’ verse. … In other words it is a satura, composed of batches of octosyllabic and decasyllabic verse, with rhyme arranged at discretion, and sometimes doubled; with rhythm varying, but not beyond the ranges of iamb and trochee. Such fingering of the general scheme had hardly been seen since Comus and Lycidas and the Arcades; [End Page 243] but the scheme could not, even to Coleridge himself, have seemed ‘new.’2
Saintsbury’s work is today frequently dismissed as “mere” appreciation, or, worse still, as ideologically suspect—verdicts that offer a poor tribute to his erudition and palpable relish for verse.3 Nonetheless I must confess that the above description chimes very poorly with my own sense of “Kubla Khan.” “Half-regular ode or lyric” hedges its bets from the start. Many of the lines of the poem are neither octosyllabic nor decasyllabic. The specification of “Broken and cuttit” verse is, moreover, more than a little misleading. The term, now sadly obsolete, was coined by King James vi of Scotland, later England’s James i. His Majesties poeticall exercises at vacant houres (1591) defines “Broken and cuttit” as a verse form that tolerates liberal amounts of syllabic variation.4 Yet several of King James vi’s own examples from the Scottish court—such as Alexander Montgomerie’s “The Cherry and the Slae”—prove neither lyric nor ode, as Saintsbury suggests; rather, they take the form of the irregular ballad.5
The History of English Prosody, then, begs quite as many questions as it answers regarding the prosodic structure of “Kubla Khan.” This being the case, it is all the more surprising to find that subsequent critical literature, which has prosecuted many recondite topics with unstinting abandon, has nonetheless drawn a blank regarding the meter of Coleridge’s elusive poem. There is to my knowledge only one dedicated treatment of the topic, which extends to a five-page article in the Spring 1962 issue of Studies in Romanticism, derived from Alan C. Purves’s doctoral dissertation on Coleridge’s prosody. “[Cjritical attention,” Purves observes, “has centered on the preface and not on the poem, particularly not on the poem’s structure. Most critics have described it as ‘free or irregular’ and have let the matter rest there.”6 Purves continues, however, to draw a contrast between those critics who stress the fragmentary character of “Kubla Khan,” and those, like himself, who see rather an organic whole that is metrical as much as thematic: [End Page 244]
… For I believe those champions of the poem’s completion have been struck—as was I—by the fact that “Kubla Khan” is much less irregular, particularly in its versification, than would first appear. It has less metrical variation than would be expected from the author of “Christabel” and curiously few examples of the major Coleridgean device, the anapest. “Kubla Khan’s” metrical irregularity takes the form of feminine endings, truncated openings, inversions, and a few spondaic or pyrrhic substitutions—none of these variations novel or surprising in the Coleridge canon. Rather it is the real amount of restraint Coleridge has exercised which has perhaps made several critics feel the poem to be a whole.7
Purves is quite right to argue that “Kubla Khan” is a highly structured poem, yet not quite in the conventional manner that the above description suggests. The claim that Coleridge “discard[s] the loco-descriptive pentameter couplets, in which he had painstakingly detailed the natural world,” so as to conclude his poem with “truncated tetrameters … a favorite form for mysterious and chant-like verse,” unduly straightjackets the poem’s metrical range, which frequently encompasses lines of odd rather than even numbers of syllables, sevens and elevenses rather than sixes and tens.8 So too does it suggest a reductive conjunction of form and function (pentameters are for describing, tetrameters for chanting) that “Kubla Khan” does much to unsettle. Such occlusions stem in large part from Purves’s restriction of his analysis to the unit of the individual line, which abridges the ways in which Coleridge’s poem continually combines itself (or threatens or flirts to combine itself) into larger metrical wholes, which also are generic wholes. In short, Coleridge’s poem is indeed organized, but both more and less than Purves implies.
More recent, broadly formalist accounts of Coleridge’s verse do not significantly advance our specific understanding of “Kubla Khan.” J. C. C. Mays’s Coleridge’s Experimental Poetics (2013) discusses the poem briefly, declaring that its three strophes “exactly mimic the turn, counter-turn and stand of the Pindaric Ode”9—where “mimic” leaves suggestively open whether or not the poem is Pindaric. I myself am susceptible to the same charge, in that my Coleridge and the Philosophy of Poetic Form failed to say anything about the poem, despite its title.10 But as a final instance of just [End Page 245] how difficult it is to account for the sonic organization of “Kubla Khan” (before I lay my own cards on the table), I want to consider an approach that differs greatly from the sober guise of monograph or journal entry: Herbert Tucker’s interactive prosodic website, For Better for Verse. “4B4V,” as it has become familiarly known, enables students to conduct traditional foot-based scansions in electronic form, by marking stressed or unstressed syllables with a click of the mouse.11 Selected poems can be organized alphabetically or by perceived level of difficulty, with the specimens divided into “warming up,” “moving along” or “special difficulty.” Perhaps unsurprisingly, “Kubla Khan” is a case of “special difficulty.” Yet despite that disclaimer, Tucker chooses to read the poem in a similar fashion to Purves’s earlier account: as a generally iambic construction that resolves itself into even numbers of syllables. Individual lines push this principle to the breaking point. Take line 25, “Five miles meandering with a mazy motion,” where despite Tucker dragging out “meandering” with the diligence of an elocutionist, so as to make an even-syllabled line, the stress pattern deviates wildly from his stated norm:
Five miles meandering with a mazy motion
4B4V is configured in such a way that only when you scan a line “correctly,” according to Tucker’s precepts, can you access his description of that line’s effects. Many are the times that I have found myself dutifully clicking through a scansion to which I do not wholeheartedly subscribe, in order to get another dose of Tucker’s incomparable prose style. The line above is a case in point. Tucker glosses:
Stress “with” here at mid-line if you must. But there’s more fun to be had, legal too, in exploiting the capital afforded by the spondee in foot 1, and shooting the rapids of four consecutive slacks mid-line. (Best practice, in this case, would make feet 3 and 4 a pyrrhic and an anapest, respectively; alas, the coding behind 4B4V won’t permit both that and the safer anapest-iamb combo for those feet, so grit your teeth and scan the middle foot as 3 slacks, a super-pyrrhic). The line just above this one is nearly as impressive in the same vein: straight iambic scansion is okay there but rather feeble from a hydraulic standpoint. (Try it again?) There’s plenty of recovery time downstream in the quite regular lines that are coming up.
It feels somewhat evasive to hide behind what the code “permits,” given that Tucker himself designed that code.12 In any case, the prosodic variability [End Page 246] that the above gloss concedes is far from a momentary overflow that settles down into “quite regular lines … coming up.” Throughout 4B4V’s scansion of “Kubla Khan,” Tucker routinely stresses prepositions and conjunctions in an increasingly rearguard attempt to salvage the magical iambic patterning. He advocates accenting “to” in “to a lifeless ocean” (28), the first “of” in “The shadow of the dome of pleasure” (yet not the second), “on” in “midway on the waves” (32), and, in a strange trochaic inversion, “of” in “Singing of Mount Abora” (41)—but not “Mount”! The same holds for conjunctions such as “as” (“And here were forests ancient as the hills” ), “and” (“A savage place! as holy and enchanted ), and “for” (“By woman wailing for her demon-lover!” ). Polysyllabic words come in for comparable treatment: we are told to stress “-mer” in “dulcimer,” to make it iambic, and “ly” in momently, not once but twice (19, 24).
It is perfectly acceptable to promote the stress of occasional monosyllables, so as to uphold an established rhythmic pattern; yet the continued recourse to such linguistic wrenching as we see above necessarily comes at the cost of forcing a particular kind of declamation, rigid and stagey. That fact suggests that the tendency for duple stress-patterns to capsize into pyrrhics, anapests, or other such variations, is far greater than Tucker, Purves, or Saintsbury wish to allow.13 A similar story emerges at the level of elision. Here Tucker’s protocol is not to elide when it procures an even number of syllables (hence the drawn-out “me-an-der-ing”); and to elide when it procures an even number of syllables (hence “towers” in “With walls and towers were girdled round” ). The poem itself struggles against a framework that has already been predetermined, whose operative constraints are already encoded within the system.
In short, it has proven difficult to account for the sonic organization of “Kubla Khan.” Given such difficulty, it is perhaps unsurprising that so many critics should drop all pretense to analytical precision, and either move on to thematic concerns, or simply designate the form more vaguely as “lyric,” a term whose uncritical pervasiveness Virginia Jackson and Yopie Prins, among others, have repeatedly explored.14 (Coleridge’s own little story of the poem’s inspired and truncated composition certainly does [End Page 247] encourage several lyric shibboleths.) In what follows I will endeavor to give a more specific account of the poem’s sonic organization. Yet I do so in the belief that such an account holds consequences that extend far beyond exactitude for its own sake (whatever that might mean). My reading will ground two related claims: first, that the singular rhythmical variety of “Kubla Khan” offers an opportunity to reconsider the relationship between conventional scansion, on the one hand, and more quantitatively exact analytic methods, on the other. Second, I claim that this rhythmical variability necessarily alters what we take to be the political resonance of Coleridge’s poem.
My reading voluntarily confines itself to the same, conventional foot-based analysis that Purves and Tucker adopt, for several reasons: so as to offer a direct alternative to their findings; so as to demonstrate how even such relatively crude measures can still model rhythmical diversity; so as to indicate the recurrent moments at which Coleridge’s poem engages critically with historical conventions, regardless of the ontological claims that we may or may not be willing to accord to iambs and all the rest; and so as to significantly contrast with other prosodic modes that I will explore in due course. It is anchored in two prosodic or generic forms that my preamble has already evoked: the endecasillabo, the principal meter in Italian poetry, whose oddness contrasts with the more standard, even-numbered tetrameter, pentameter, or trimeter; and a form of sonic organization that extends beyond the individual line, namely, ballad meter. Those two forms, which bear an extremely contrastive set of historical associations and expressive possibilities, serve early and emphatic notice of the unusual generic hybridity of “Kubla Khan.”
Of all the micro-modifications permissible within ballad meter, Coleridge is particularly fond of the stanza that features eight, eight, eight, and six syllables. Both “Love” and the “Ballad of the Dark Ladie” follow such a pattern. So as to have its sound in our mind, here is the opening to the former:
All thoughts, all passions, all delights,Whatever stirs this mortal frame,All are but ministers of Love, And feed his sacred flame.15
This work—once one of Coleridge’s most highly-prized works, now fallen into disregard—clearly differs greatly from the mood and tone of “Kubla Khan.” Yet the latter work opens in accordance with very similar metrical [End Page 248] principles, albeit with an additional tetrameter line. It is curious, then, that ballad (in comparison to the more capacious designation of ode or lyric) has seldom been used to comprehend any of “Kubla Khan”’s successive modifications. Saintsbury’s identification of “broken and cuttit” verse is in fact true but not in the manner he intended, insofar as the form indicates the ballad rather than any “freely” irregular ode:
In Xanadu did Kubla KhanA stately pleasure-dome decree:Where Alph, the sacred river, ranThrough caverns measureless to man Down to a sunless sea.(1–5)16
Thomas Gray’s scaldic odes are perhaps the nearest thing to a literary precursor, in their partial integration of bardic tradition. Yet in those poems one generic tradition typically wins out: “The Fatal Sisters: An Ode” modifies the Horatian rhyme-scheme and divides itself into quatrains, with the result that ballad clearly sounds the dominant note. By contrast, “The Bard: a Pindaric Ode” (which Coleridge had much admired only to tire of it in later years) moves freely and rapidly across accentual-syllabic variations, despite both the work’s title and Gray’s erroneous claims to have replicated metrical devices specific to oral Welsh verse.17 Where in the first case ballad wins out, Pindaric triumphs in the second.
“Kubla Khan,” by contrast, summons the ghost of ballad only then to hold it at an uncertain distance. In these opening five lines the mode is no sooner present than it is subject to disfigurement, at first of a relatively mild kind, with the trochaic inversion of the postponed trimeter that makes us stress “Down” hard, a vocal thrust quite unlike the pitter-pattering iambs of “Love.” But the poem then heads off on a very different tack. We know from the so-called Crewe manuscript, discovered in 1934, that Coleridge altered the very next line, which had read “So twice six miles of fertile ground,” to “So twice five miles of fertile ground.” What substantive difference might underlie the apparently trivial shift from six to five? The latter certainly gives Coleridge another longer-drawn vowel sound, to complement that lodged within “twice.” But another plausible answer, I want to suggest, is that Coleridge here indulges in the sort of self-reflexive metrical pun to which he was so often prone, where “twice five” would constitute the hallowed iambic pentameter, which at separate historical moments [End Page 249] came to supersede other metrical repertoires that had preceded it. The surmise is made more than merely speculative by the manner in which, at this very juncture, “Kubla Khan” begins to lengthen its metrical stride beyond ballad:
So twice five miles of fertile groundWith walls and towers were girdled round;And there were gardens bright with sinuous rils,Where blossomed many an incense-bearing tree;And here were forests ancient as the hills,Enfolding sunny spots of greenery.(6–11)
This opening strophe concludes with one foot firmly planted in the iambic pentameter form, which tempts us to naturalize it as akin to those “ancient” forests and hills. Yet the ragged right-hand margin, which swells out prior to retracting, already suggests that the passage tests such constraints—a supposition that closer attention confirms. Lines 8 and 9 can at a push be made into regular iambic pentameters, but only by eliding a syllable: in the first case, “sinuous,” the vocal contraction is hardly unusual; but in the following, we have to take rather more drastic measures, quickly blurting “many,” perhaps, so as to avoid the excess of the endecasillabo.
The second strophe proceeds to subject those norms to still greater pressure. Its opening line (“But oh! that deep romantic chasm which slanted” ) features another word, in “chasm,” that we might choose to elide or not. (Tucker does.) But this time, even if we do contract (as both standard pronunciation and the accelerating pace of the poem certainly encourage), it exceeds the iambic pentameter that critical consensus holds to be present, stretching either to eleven or twelve syllables. But why should such syllabic excess constitute a threat? In and of itself, such metrical ambiguities are neither uncommon nor noteworthy. English is a language notoriously full of words bearing “floating” stress, while the great tradition of iambic pentameter (Shakespeare, Milton) itself features numerous hypercatalectic lines.
Yet “Kubla Khan” radicalizes this general linguistic and literary ambiguity, first, through the determined frequency with which it employs such syllabically ambiguous words that, despite Tucker settling each case de facto, produce a residue of metrical uncertainty; and second, through the manner in which the poem’s earlier prosodic shifts have already sensitized us to the difference of a syllable here, a syllable there. If the tension between ballad meter and iambic pentameter throws up a self-evident abundance of broader social associations, the emergence of the endecasillabo brings another kind of cultural capital: that of classical quantity, the Italianate tradition, exotic specimens to trespass upon sturdy, English, pentametrical soil. We [End Page 250] know that such matters were far from incidental to Coleridge, who at the point of composition was obsessively experimenting with traditional quantitative meter: depending on whether we date “Kubla Khan” to 1797 or 1799, it falls immediately before or after a translation that would later be published under the title “Catullian Hendecasyllabics” (1834), a work that, whether by intention or design, cast the classical form in a looser accentual guise.18 Tennyson and Swinburne would similarly employ “Englished” versions of the hendecasyllable so as to probe the relation between quantitative meter and the contemporary vernacular. Part of my argument is that Coleridge’s poem represents a largely unheralded anticipation of what Ben Glaser—taking a more positive cue from Saintsbury than I was earlier able to muster—calls the “polymetrical” poem.19
I here choose to refer to the eleven-syllable line as an endecasillabo, rather than the classical hendecasyllable, due to my entire agreement with Martin J. Duffell’s claim “that quantitative verse can be no more than an intellectual exercise in English because of the language’s strong dynamic accent and tendency towards stress-timing. ”20 While the classical tradition undeniably did intrigue Coleridge, Swinburne, and Tennyson (among several others), their Englished hendecasyllables in practice generally amounted to accentual renderings of what had been long and short quantities. The endecasillabo, by contrast, proves relevant in two particular ways. On the one hand, the Italinate form featured an alternating stress-pattern with an unstressed terminal syllable, whose pertinence to the ensuing form of “Kubla Khan” will become evident. On the other, the same form was involved in a no-less-significant historical process of subsumption into the English tradition, through which Chaucer modified his continental source-texts into the iambic pentameter proper. By revealing the residual trace of that supplanted form within the very structure that superceded it, “Kubla Khan” thereby exposes (even if it cannot reverse) that passage of cultural appropriation and transfiguration.
Where the first strophe teeters between iambic pentameter and the endecasillabo, the latter subsequently emerges as the operative mode of [End Page 251] “Kubla Khan”’s second strophe, eclipsing the regular decasyllable that had previously eclipsed ballad. Of the nineteen lines that it contains (and allowing for ambiguities of the sort that I have described above), I take an appropriate eleven of them to contain eleven syllables, a consistency that again distinguishes “Kubla Khan” from any irregular ode of which I am aware. Of course one might object that such syllabic variations suggest that we would be better off counting not syllables, but rather beats. Yet for all that Coleridge’s interest in accentual meter is well-documented, the poem at hand offers a very different specimen from “Christabel,” and never more so than in its studied deployment of the eleventh syllable, all the more significant for its not bearing stress. In every single hendecasyllable line within the second strophe, the extra syllable permits a terminal feminine rhyme (I mark the eleven in italics):
But oh! that deep romantic chasm which slantedDown the green hill athwart a cedarn cover!A savage place! as holy and enchantedAs e’er beneath a waning moon was hauntedBy woman wailing for her demon-lover!And from this chasm, with ceaseless turmoil seething,As if this earth in fast thick pants were breathing,A mighty fountain momently was forced:Amid whose swift half-intermitted burstHuge fragments vaulted like rebounding hail,Or chaffy grain beneath the thresher’s flail:And mid these dancing rocks at once and everIt flung up momently the sacred river.Five miles meandering with a mazy motionThrough wood and dale the sacred river ran,Then reached the caverns measureless to man,And sank in tumult to a lifeless ocean;And ’mid this tumult Kubla heard from farAncestral voices prophesying war!(12–30)
The cumulative effect of these extended bursts of feminine rhymes is to force the voice in two directions at once, as it struggles both to downplay and to accentuate the force of the word concerned. Each endecasillabo in itself stages this rhythmical tension; but so too does it further extend that variability through a contrapuntal relationship with those iambic pentameter lines that are present (that former note not being definitively silenced, but living on as a trace), each of which brings its heavy masculine rhyme. After a run of seven hendecasyllables with a significantly varied rhyme [End Page 252] scheme (12–18), we reach four couplets, whose heaviness we cannot but tie to the emphatic content that they communicate: the forcing of the fountain, the rebounding “hail,” the “thresher’s flail.” In a still more insistent register, the second strophe itself concludes with another heavily accentuated rhyme, “war!” Here the iambic patterning upon which Purves and Tucker were half-right to insist exists only in tension with an endecasillabo that liquidizes masculine rhyme into the meandering river or ocean. Just as “Kubla Khan” had earlier situated ballad and iambic pentameter in the expressive distance that spanned each form, so it now discovers a further historical repertoire (here the Augustan couplet tradition, albeit that Pope would have shuddered at such run-ons) that emerges at the very moment that it is estranged.
From at least as early as T. S. Eliot’s “Prufrock” we have become familiar with the idea that a poem might doll itself up in metrical fancy-dress, plucked off history’s shelf. Perhaps the critic John Beer was correct to describe “Kubla Khan” as “a poem about poetry,” not because it offers a stock description of creative inspiration, but rather because it is the first poem to raid the historical stockroom in such a sustained and comprehensive manner.21 When Coleridge’s poem speaks of a “mingled measure,” or of being “measureless to man,” it therefore offers a more literal truth than we customarily suppose.
The recent formalist resurgence betrays a curiously uneasy relation to more exact and quantitative modes of analysis. This phenomenon has deep roots, and helps to explain the common misconception of I. A. Richards’s “practical” criticism either as subjective immersion in the text at hand, or as a form of knowledge that would hold parallels to scientific method—when for Richards it was both. Today, “new” formalist approaches generally work in opposition to or disregard the various quantitative methodologies that hold currency, be they Marina Tarlinskaja’s statistical researches into English verse, Franco Morretti’s distant reading (itself conceived in opposition to “close” approaches), or the various computational scansion devices available (Scandroid, OpenMARY, Prosodia).22 Tucker’s For Better for Verse, for all its pedagogical utility (to which I can wholeheartedly attest), is essentially a straight translation of conventional reading practices into a computational setting. [End Page 253]
Insofar as the account of “Kubla Khan” that I have offered above is part of that current formalist agenda (and I am rather uneasy about that affiliation), I want now to explore some parallels and contrasts with another approach to quantitative analysis. That approach is Reuven Tsur’s “cognitive poetics”—a methodology that he has spent the best part of the past half-century developing. Cognitive poetics seeks to bridge the gap recently described, between the reader’s subjective experience of a poem, on the one hand, and more verifiable, generalizable metrics, on the other, in two related ways. First, Tsur argues consistently that individual subjects interpret (or embody) verse as a complex rhythmical gestalt; second, he seeks to reveal the commonalities and disparities between such gestalts, through a dedicated phonological analysis that goes far beyond the conventional foot-based metrics that I have applied above, measuring fine-grained vocal variations in pitch and tone.
The most systematic account of Tsur’s methodology arrives in his Toward a Theory of Cognitive Poetics (2008).23 For reasons of self-evident relevance, however, I wish in this section to concentrate rather on a comparatively lesser-known work, Kubla Khan—Poetic Structure, Hypnotic Quality and Cognitive Style (2006). That book is something of a patchwork affair, comprising four sections: the first two, which chart the critical reception and formal structure of Coleridge’s poem, were first printed by Israel Science Publishers in 1987; the third, which conducts empirical studies on a range of performances of Edward FitzGerald’s Rubáiyát, appeared separately in 1990–91; while the fourth part, which analyzes sound recordings of various recitations of “Kubla Khan,” was appended to the whole and published by Tel Aviv University Press in 2006. It is on this fourth and final section that my analysis focuses, in an attempt to correct in however small a way the neglect that literary criticism has on the whole shown for Tsur’s project, though it has gained greater traction in the fields of linguistics and cognitive psychology.24 In the specific field of Coleridge scholarship, David S. Miall is one of the few critics to apply something like Tsur’s approach in any sustained manner.25 [End Page 254]
Cognitive poetics deserves such a correction for several reasons. First, and most obviously, Tsur’s attention both to the micro-textures of “Kubla Khan” and to the changing whole that they cumulatively form—an attention characteristic of his work as a whole—serves as a welcome respite to the more general tendency, surveyed above, to reduce Coleridge’s poem to a constituent unit that would then serve as the basis for hasty generalization. Whereas Robert Fleissner, for example, hyperbolically interprets “measureless to man” as a formulation that “phonologically [suggests] the notion of infinity,” Tsur draws a more measured contrast with “five miles meandering,” which despite repeating the same consonant achieves a very different effect, on the basis that its successive //m// sounds fall alternately on stressed and unstressed syllables.26
Second, Tsur’s empirical case studies suggest the extent to which individual performances differ even within such rhythmical constraints. The two spectrograms reproduced below measure the pitch, duration, and amplitude in the intonation contours of two recitations of the third line of Coleridge’s poem, “Where Alph, the sacred river, ran.”27 Taken together, they suggest that the generic and metrical tension that my earlier analysis uncovered (between ballad and pentameter, between pentameter and hendecasyllable) emerges also at the level of individual recitation. The first (fig. 1) is delivered by the actor Michael Sheen, in a manner that, as Tsur summarizes, “is very consistent in suppressing versification in favor of syntax when conflicting intonation contours are demanded.”28 In the second (fig. 2), which charts the performance of the actor Ralph Richardson, even those unversed in waveforms can immediately note significant differences: most obviously, the comparatively large increases in pitch, and the tendency for that pitch to level out into a “hat,” which suggests the emphasis of particular syllables. Tsur takes such data to offer an objective correlative to what we intuitively hear in Richardson’s declamation: that, far from [End Page 255] sacrificing the demands of versification to the demands of syntax, he rather adopts an incantatory tone. The exceptionally sustained length and pitch of “ran,” Tsur states, “displays conflicting cues at this point—both of continuity and discontinuity, generating a sense of firm closure at the line ending, and an impetuous forward drive beginning a run-on sentence.”29
Such variations lend some empirical ballast to my earlier claim that “Kubla Khan” is a poem that at separate moments falls into and resists the emphatic declamation that we associate with more rigid forms of ballad. Tsur, indeed, speculates that Coleridge’s poem lies midway between the metrical consistency of Pope’s Essay on Man (though in the present context we might want to think rather of Blake’s “The Tyger,” or the nursery-rhymes of A. A. Milne), and the odes of Keats and blank verse of Shakespeare, both of which permit extremely liberal promotions and demotions of stress. Such a placement raises the intriguing suggestion, fully realized neither in this work nor in Tsur’s output as a whole, that we might measure English verse in general not just despite variation, but rather, through it. Taking a speculative cue from the analysis of stochastic events, that is to say, we could compare poems according to their differing levels of turbulence, contrasting the comparatively wide range of expressive possibilities that a poem such as “Kubla Khan” affords (without pronouncing any individual performance “correct”), in comparison to more rigid productions. Positivistic analysis, in short, need not prove crudely reductive. Whereas several strands of generative metrics continue to uphold the notion of correct performance, any divergence from which is thereby “unmetrical,” [End Page 256] Tsur’s more nuanced claim that “a rhythmical performance is a perceptual solution to a perceptual problem posed by the conflicting patterns of language and versification” opens the question as to which kinds of text (and which kinds of reader) produce significant rhythmical diversity—where diversity itself, so far from being a sop to “undecideability,” can be formalized.30
Despite the great utility and suggestiveness of Tsur’s work, however, several of its consequences remain undeveloped, while certain of its premises are more generally open to debate. One indication of the latter appears toward the start of Kubla Khan—Poetic Structure, Hypnotic Quality and Cognitive Style, where Tsur sets out the experimental procedure behind his third part, which analyzes a far wider sample set of performances than the professional recordings of “Kubla Khan”:
In our empirical study we used a relatively unknown Hebrew poem rather than “Kubla Khan” as our paradigmatic hypnotic poem, because the majority of our experimental subjects were Israeli students, who would have no significant or reliable intuitions concerning an English poem. Had we used native speakers of English they would very likely have been biased by the reputation of this well-known poem.31
The rationale is clear enough; but it misses the obvious point that readers are invariably “biased” in a manner that transcends their familiarity, or otherwise, with the specific text at hand. If poetical guinea-pigs or professional [End Page 257] actors declaim poems differently, it is not only a question of character “type,” a fairly rigid notion of which emerges throughout Tsur’s analysis, but also (and perhaps more powerfully) a reflection upon the historically mediated norms for performance. Tsur grants the fact in passing, glossing that “If Sheen’s tone [in reciting “Kubla Khan”] can be located somewhere between the tale-teller and the ordinary speaker, Richardson’s strikes at once an unnatural, ‘bardic,’ incantatory tone.” This latter tone, which Tsur strongly prefers, nonetheless “may sound offensive to some present-day listeners.”32 Yet that thought goes undeveloped: if I happen to prefer a plainer rather than a more emphatic form of recitation, that preference is never merely what Immaunel Kant called an expression of taste; it rather reflects also upon the historical severance of verse and song that leaves ballad and “the bardic,” whether for better or for worse, out of joint with the present.
Such problems are in part intrinsic to Tsur’s methodology. If the various computational scansion machines available or in development all necessarily curtail the variability of English verse, through their binary solution of a language whose linguistic and prosodic basis is famously insecure, Reuven Tsur’s spectrograms risk the opposite failing. So far from effacing the variety of individual recitations, these exact measures might well give us too much vocal granularity, tempting us to refer significant differences to the irreducible singular. From there it is a short step to total and therefore useless hermeneutic license, from which quantitative analysis had been supposed to rescue us in the first instance. Tsur’s spectrograms therefore need to be supplemented by the consideration of a force too subtle to register upon them—the force of generic and prosodic precedent. In the midst of numerous references to “bardic” recitation, the absence of ballad from Tsur’s analysis is telling. The absence of that marker staunches the political and social consequences that might otherwise flow from his unusually rigorous analysis of verse.
My earlier reading of “Kubla Khan” voluntarily refrained from saying anything at all about Coleridge’s third, concluding strophe. By turning to it now 1 wish to introduce another obscured element, namely, the political or social resonance of Coleridge’s poem, without which my earlier analysis would amount merely to a species of virtuosic but hermetically self-involved metrical play. In order to lend some contrastive precision to that reading, I contrast it with a justly influential earlier account: Marjorie Levinson’s consideration of “Kubla Khan,” which formed one chapter of her The Romantic Fragment Poem.33 Levinson’s analysis focuses upon the [End Page 258] headnote that Coleridge added to the poem’s 1816 publication, which she reads in light of the significant alterations in Coleridge’s political outlook in the years since the poem’s initial composition.
That headnote includes not only Coleridge’s tendentious explanation of the poem’s truncation, but also a fragment of another earlier poem, “The Picture, or the Lover’s Resolution,” whose content (“Then all the charm / Is broken—all that phantom world so fair / Vanishes”) offers a cautionary preparation for “Kubla Khan” itself. That Coleridge cannot bring himself to articulate explicitly the parallel between preface and succeeding poem, Levinson reads as a “personal failure.” She continues,
The poem, however, does for itself what Coleridge cannot do for it. It criticizes itself mercilessly, condemning its own compositional method even as it thereby unfolds. “Kubla Khan,” read as a completed fragment, exposes its organism, its spontaneous expressivity, and its escapism as a terrible tyranny.34
On this reading, the narcissism of the gazing youth (“The pool becomes a mirror”) resembles the concluding coda so nearly that we cannot entertain damsels with dulcimers and the rest without a painful twinge of self-deception. This late realization retrospectively alters the political gravity of the poem. While Kubla’s dictatorial regulation may once have appeared life-denying ordinances, the hard-won realism of the headnote now, according to Levinson, recasts them as the least bad option—a direction of political travel all-too-familiar from Arab Spring, to name but a recent iteration. “The 1816 ‘Kubla Khan,’” she summarizes, “represents Coleridge’s effort to reclaim for a sober or a sadder maturity a document born of his errant youth.”35
Levinson is quite justified in the claim that “Kubla Khan” casts doubt upon organicism, spontaneous expressivity, and other such cognates. Yet my account generates a significantly different political moral, in two related ways. On the one hand, Levinson’s rich account of Coleridge’s compositional revisions (and the political recantations to which they are tied) produces a curiously thin account of the reading process. Is it really the case that a reader is so dominated by a superadded headnote that the concluding coda now simply confirms its dangling suspicions? (Levinson’s approach brings to mind Alan C. Purves’s complaint that criticism “has centered on the preface and not on the poem, particularly not on the poem’s structure.”) The question extends far beyond Coleridge. Do we read Matthew Arnold’s notorious Preface to his 1853 Collected Poems with such compliance that we are then inoculated against the romanticism that Arnold [End Page 259] sought to deny in himself? The consequence appears to be a reader that has already decided what a work can give (or what it cannot) in advance— what Eve Kosofsky Sedgwick saw as the “anticipatory” strand of paranoid reading.36 (It also requires that we take on faith the parallel with “The Picture, or The Lover’s Resolution,” despite the ease with which we doubt Coleridge’s other retrospective explanations of “Kubla Khan,” Persons from Porlock et al.)
The second motivating reason for my dissent is more substantive, bearing as it does directly upon the poem’s sonic organization. On that score Levinson’s analysis remains silent, although she does speak generally of the concluding coda as setting up an “organic and negatively capable form [that] requires a complete interpenetration of form and content.”37 (This, of course, is precisely the doctrine that the headnote annuls.) It is hard to know what is meant by “form” here, except in the broadest sense of a naïve idealism that would totalize; and indeed Levinson’s earlier conflation of “formal intention—an idealist interest” suggests just that.38 But if we consider more fully the “formal” dimensions of the concluding coda, a far less monolithic picture emerges. For the earlier version of Coleridge’s poem (that of the Crewe manuscript) already disappointed many of the poem’s earlier claims, without needing the intercession of the headnote. If it did not need the headnote, however, the nature of the poem’s disappointment necessarily differs. In contrast to Levinson’s account, I do not believe that such a process unfolds through the reader’s attentive, self-reflexive comparison of two moments (one textual, one paratextual), held together primarily through “poetical imagery” such as “watery reflection”—an image that I must confess to being unable to locate anywhere in the coda.39 Rather, it more fundamentally occurs through a series of sonorous adjustments, which are only the more significant for existing at a less vigilant and self-policing level of awareness.
The third and final strophe is more properly heard not with the distant headnote in mind, but the several echoes of the poem’s earlier modifications:
A damsel with a dulcimerIn a vision once I saw.It was an Abyssinian maid [End Page 260] And on her dulcimer she played,Singing of Mount Abora.Could I revive within meHer symphony and song,(37–43)
Those echoes cast doubt on the claim that “[i]n the eighteen-line coda, the poet solves the equation he has set up.”40 For whereas the stand or epode generally resolves a Pindaric Ode through its introduction of a new meter, the third strophe of Coleridge’s poem rather signals a retreat to something like the iambic tetrameter, with its ghost of ballad, that we witnessed in the poem’s opening. I say “something like,” because the feel of this passage, despite its recognizable alternation between rough iambic tetrameter and trimeter, in other respects diverges greatly from precedent. Inseparable from the speaker’s painfully subjunctive register is a sudden metrical stability that has no precedent in the poem. Where the early strophe enjambed freely and expanded into iambic pentameter, and the second strophe staged a continuing tussle between decasyllabic and endecasillabo lines, here the verse is consistently (often emphatically) end-stopped and syllabically regular, its rhymes unvaryingly masculine. Those variations that are present extend no further than trochaic catalectic lines (38, 41, 51)—hardly rare departures from such metrical scaffolding. The effect is a more direct performance of ballad that, for that reason, has a greater sense of forced artifice than any of the mongrelized meters that came before—just as the speaker’s exclamatory force (“Beware! Beware!”) almost, but not quite, persuades us that song is being revived rather than hypothesized. (Tsur notes that only at this late point does Michael Sheen’s more nuanced performance becomes truly “incantatory”—paradoxically a sign of the comparatively minimal expressive possibilities that this late passage allows.) That broader cry (“And all should cry, Beware! Beware!”) betrays the rigid iambic patterning that dominates these final eight lines, with the exception of a small trochaic inversion (51); only at this late stage does the regular duple structure, which Purves and Tucker identified with varying degrees of confidence, materialize.
Phrased in such a manner, I make the concluding strophe sound a rather unsatisfying affair; but this non-satisfaction negatively confirms the truth of what had preceded it, whose ghost thereby lives on the more powerfully. This newly stable ballad form represents a failed attempt to force the incantation that its speaker so desires. But that failure only registers by virtue of the comparison with a formal hybridity that the poem, if not reader, has forgotten. “Kubla Khan” never gave us a totalizing romantic organicism [End Page 261] that could only be disabused by extra-textual forces (the slow years of political self-education, the addition of a headnote). For it already contained within itself a complex reflection upon the relation between liberty and constraint. Kubla’s role is indeed recast in light of the poem’s conclusion: yet he represents less a political expediency to which we would retreat, faute de mieux, than an embodiment of the manner in which the most apparently rigid and reified forms of measure contain unsuspected pluralities.
Such an analysis as I am attempting might well come under attack from both sides of the critical (which is also to say political) spectrum. On the one hand, it might be alleged that such an emphasis upon prosody necessarily restricts the poem to those who can perceive and commend Coleridge’s virtuosic metrical performance. On the other, those still loyal to the venerable idea of romantic inspiration might well regard the historical mediation of prosodic gestures and generic forms as a travesty of the sui generis. But if my analysis seems to have indicated an unbridgeable gap between the unconscious inspiration that Coleridge’s prefatory note claims, and a virtuosic poetic technique, perhaps this is only to say that we need to revise our understanding both of unconscious inspiration and of technique. For when Steven Pinker claims that the imagery of “Kubla Khan” represents a kind of “mentalese” prior to language use proper, insofar as scientists and artists “alike think in images,”41 he is quite correct—except that sound, rather than image, would have better suited the case at hand. Just as your unconscious is never only your own property, so sonic micro-adjustments, a syllable here or there, involve us in larger organizations regardless of our awareness of the fact or possession of the proper terminology. That truth inheres in vocal variation, for all that it shows up on no spectrogram.
Ewan James Jones is a Lecturer in nineteenth-century literature at the University of Cambridge, and a fellow of Downing College. He is the author of Coleridge and the Philosophy of Poetic Form (Cambridge, 2014), and of several articles on prosody and aesthetics. He is currently completing a second monograph on the history of the concept of rhythm in the nineteenth century.
1. I am grateful to Reuven Tsur and to the two anonymous readers of an earlier version of this essay.
3. For an impassioned defense, see Michael Hurley, “George Saintsbury’s History of English Prosody,” Essays in Criticism 60, no. 4 (2010): 336–60.
5. Montgomerie’s poem is thought to date to the early 1590s. For a later printing, see Alexander Montgomerie, The cherry and the slae: with other poems (Glasgow: printed and sold by Robert and Andrew Foulis, 1751).
12. For Tucker’s reflection upon the difficulties and aftordances of such a resource, see his “Poetic Data and the News from Poems: A For Better for Verse Memoir,” in Victorian Poetry 49, no. 2 (Summer, 2011): 267–81.
13. At some points the general tendency to stress iambically comes into tension with the attempt (through elision) to keep the syllable-counts even; see 8, for example, where Tucker gives ‘sinuous’ two (iambic) syllables, thus leaving the line with eleven syllables.
18. Despite this title, the poem is actually a rendering of a German original in lines of twelve syllables, a fact signalled by their publication as “English Duodecasyllables, Adapted from Matthisson,” in Poetical Works, ed. J. C. C. Mays, 1:530–31. Edwin Guest discussed the experiments as part of his discussion of Sapphic metre in English, asking “[w]ould some of Coleridge’s lines be very much injured by thus lopping them of a syllable?” (A History of English Rhythms [London: W. Pickering, 1838], 271). Clearly, an extra syllable here or there did count.
22. Charles Hartman’s Scandroid is available at http://oak.conncoll.edu/cohar/Programs.htm; OpenMARY at https://github.com/waywardgeek/openmary-4.3.o/tree/master/doc/javadoc/marytts/language/en/prosody, both accessed 20 July 2016. The Stanford Literary Lab has been developing Prosodia for some time, although at the time of writing it is not available.
24. Although the recent edition of The Princeton Encyclopaedia of Poetry and Poetics (Princeton: Princeton University Press, 2014) does now feature an entry on “cognitive poetics” (272–73).
25. While it occupies a comparatively small proportion of his analysis, Miall’s analysis of phonemic patterning in “Frost at Midnight” suggests how linguistic features might usefully supplement more standard forms of scansion, and in turn be referred to broader (and quantifiable) linguistic habits (“‘Too soon transplanted’: Coleridge and the Forms of Dislocation,” in The Quality of Literature: Studies in Literary Evaluation, ed. Willie van Peers [Amsterdam: John Benjamins Publishing, 2008], 109–112).
26. Fleissner reads the full line as follows: “It suggests that something deep, remote, and sunken is in the caves—even as the sound effect BOUM describes the echo in Forster’s Marabar Caves. That sound-effect suggests not only what is incidentally non-human, but it happens to connote the Hindu word for God, OM, even as measureless to man conveys a similar anagogic effect. Phonologically, both expressions point to the notion of infinity” (Xanadu Re-Routed: A Study in the Ways of Romantic Variety [New York: Edwin Mellen Press, 2000], 14–15, quoted in Tsur, Kubla Khan—Poetic Structure, Hypnotic Quality and Cognitive Style [Tel Aviv: Tel Aviv University Press, 2006], 46). “All this,” Tsur comments with critical relish, “is ad hoc and the result of unbridled free association, nothing principled.”
27. All the audio recordings upon which Tsur bases his argument are available at http://www.tau.ac.il/~tsurxx/KublaEmpirical_folder/KublaEmpirical.html, accessed 20 July 2016.
36. Sedgwick, “Paranoid Reading and Reparative Reading, or, You’re so Paranoid, You Probably Think this Essay Is About You,” in Touching Feeling: Affect, Pedagogy, Performativity (Durham: Duke University Press, 2003), 130.
41. Pinker, The Language Instinct: How the Mind Creates Language (New York: Harper Perennial, 1995), 61. The brief discussion of “Kubla Khan” comes in a broader treatment of the manner in which scientists and artists alike “think in images.”