[A version of this article appeared in The Southern Review for Summer, 2000.]
 

     In 1917 T. S. Eliot, who had been in London and Paris for several years and who had observed the opening salvos of Imagist theory with an eye more experienced than most of the campaigners, published "Reflections on Vers Libre" in the New Statesman. "When a theory of art passes," stated Eliot, with premature sagacity, "it is usually found that a groat's worth of art has been bought with a million of advertisement. The theory which sold the wares may be quite false, or it may be confused and incapable of elucidation, or it may never have existed." (Selected Prose 31) Further on, he adds: "If vers libre is a genuine verse-form it will have a positive definition. And I can define it only in negatives: (1) absence of pattern, (2) absence of rhyme, (3) absence of metre. . . . What sort of a line that would be which would not scan at all I cannot say." There follows a discussion of the greatest subtlety and erudition in which he ends by arguing that anything that is verse at all must continue to be an extension of an established meter, quoting for illustration his favorite Jacobean playwrights and mentioning Shakespeare. His argument that Webster's lines in The White Devil "deliberately rupture the bonds of pentameter" is unassailable; but he fails to remark that the same playwright deliberately restores the pentameter. "There is no escape from meter; there is only mastery," said Eliot, in the frequently-quoted passage. Why not, I would ask, a mastery of escape? And yet what Eliot says is so fundamentally sensible that it seems to me to need only one more small step that would admit the possibility of extraline variation, or line substitution--call it what you will--in which the concept of the metrical foot as the fundamental division gives way to the line itself as the atomic unit.
     Eliot's essay remains one of the earliest and best general assessments of what occurred in prosody in this century's second decade; it should be read and reread entire, and not quoted in snippets. Eliot continued to take exception to the idea that vers libre was really possible, and also attacked the prose poem in an article published in the New Statesmen for May 19, 1917, "The Borderline of Prose." Thirty years later, Eliot had softened considerably on these issues, but at that time he insisted on a definite distinction between prose and poetry, and insisted there must be no discussion of either free verse or prose poetry. And only four years later he came around to an admission that long poems could take on qualities of prose and that prose could aspire toward poetry; in the Chapbook: A Monthly Miscellany of April, 1921, he wrote:
 

       Poetic content must be either the sort of thing that is usually, or the same thing that ought to be, expressed in verse. But if you say the latter, the prose poem is ruled out; if you say the former, you have said only that certain things can be said in either prose or verse. I am not disposed to contest either of these conclusions, as they stand, but they do not appear to bring us any nearer to a definition of the prose-poem. I do not assume the identification of poetry with verse; good poetry is obviously something else besides good verse; and good verse may be indifferent poetry.

    Here Eliot has raised an issue that might seem to be entirely separate from the question of free verse. Much as we may enjoy "Casey at the Bat" or "The Cremation of Sam McGee," we do not promote them to equal status with Paradise Lost or with Keats's great odes. Even Robert Service knew better than that, refusing to call himself anything more than a versifier.
 My purpose here is to demonstrate how efforts to construct a more positive and definite system of free-verse scansion (than Eliot's) immediately run into absurdities and self- contradictions that cannot be resolved. The reason is that, unlike Eliot who recognized that free verse required a substrate of regularity, Yvor Winters and William Carlos Williams aimed at providing the new method with its own self-consistent rationale. In the end they differed from Eliot and from each other in radically opposite ways, Winters attempting an exacting and overspecific notation, and Williams simply asserting the existence of his "variable foot" without providing any clear examples of what he meant by it.
     A seldom-mentioned fact is that in avant-garde publications of the period 1913-16, Williams's name appeared almost as often as that of Joyce, Pound, and Eliot. Perhaps because most of his work consisted of short poems, and because his spare and deliberately unliterary style lacked the panache and knowingness of his contemporaries, a display of esoteric learning that T. E. Lawrence found ridiculous in Ezra Pound, Williams appeared at first to be a minor camp-follower of the Imagists. Later on, he barely had time as a full-time physician to write poetry and fiction; a completely articulated theory of prosody was not one of his achievements. Yet theorize he did--sporadically, spontaneously, thoughtlessly, obsessively, and with less self-contradiction than one could allow him, given all the distractions of his busy life. His disjointed utterances have spread and sprouted like the wild thistle.
     Two recurrent obsessions lent Williams's obiter dicta a certain consistency. First, he decided early on that he agreed with what Pound later said in the Cantos, "To break the pentameter, that was the first heave." At any time during the next fifty years he was ready to get out his scalpel and cut away at the diseased tissue of British metrics. In an unpublished letter to Kenneth Burke of 19 July 1955 he wrote: "To take a flier, I am completely through with the concept and the practice of blank verse. The counting of the five regular syllables [sic] makes me grind my teeth." But along with this was the conviction that poetry could not do without order. Structure was essential to everything in the universe. At times Williams seemed willing to learn what he could from earlier prosody; prior to attending a writer's conference, at which he knew Allen Tate would be present, he set himself to studying George Saintsbury's three-volume history of English prosody and admired much that he found there. In 1947 he tried to interest W. H. Auden in assembling a seminar of four or five "master poets" to discuss technical problems in the composition of poetry; he proposed for texts Samson Agonistes, one or more of Pound's Cantos, a poem by André Breton, and--amazingly enough--Eliot's Four Quartets. Williams was willing to treat with the enemy if he could learn anything thereby; of R. P. Blackmur he remarked humorously that he disliked him so much that he was anxious to meet him. When it came to his own metrics, however, he insisted that any orderly new prosody was going to have to emerge from his work rather than serving as a pre-established template, that structure was inherent and not external to the poem. Yet he did not embrace the personalized organicism characteristic of many who have considered themselves followers of Williams. In my view, at least, the amalgamation of poet-city-doctor in Paterson is a construct of the imagination that is almost as detached from Williams himself as Joyce's Dublin and Stephen Daedalus are from that artist.
     The most important thing that Williams took from his early friendship with Pound and his connections with the Imagists was the concept--which fitted well with his medical training--of submitting oneself to the object under investigation and seeing it on its own terms. He felt, along with Ford and Pound, that the metrical practices of nineteenth century had falsified the poetry and that America in particular needed to free itself of these constraints. At the same time, he always distrusted Whitman, sensing a formless momentum in his poetry that might help to sweep away the older traditions but which had nothing constructive of its own to offer. As Williams said of Carl Sandburg in 1948, "There never has been any positive value in the form or lack of form known as free verse into which Sandburg's verse is cast. That drive for new form seemed to be lacking in Sandburg." (Selected Essays of William Carlos Williams 272) As soon as Williams himself had ceased to imitate Keats, and had abandoned rhyme and accentual-syllabic meter, he began to cast about for something to take its place.
     Having rejected meter, his next step was to reject the absence of it; in 1917 he contributed an article, "America, Whitman, and the Art of Poetry," to the Poetry Journal. Here he argued that free verse was a self-contradiction and that all verse had to be subject to some sort of control. This became his settled conviction, and his efforts to hit on a method of control a lasting obsession. In 1954, in "On Measure--Statement for Cid Corman," he repeated: "No verse can be free, it must be governed by some measure, but not by the old measure. There Whitman was right but there, at the same time, his leadership failed him. The time was not ready for it. We have to return to some measure but a measure consonant with our time and not a mode so rotten that it stinks." (Selected Essays 339) That Williams was also free of Ford Madox Ford's imperfectly articulated concept of the personal cadence--which Williams refers to as "rhythm"--is evident in the first paragraph of the same essay:
 

               Verse--we'd better not speak of poetry lest we become confused--verse has always been associated
                in men's minds with "measure," i.e., with mathematics. In scanning any piece of verse, you "count" the
                syllables. Let's not speak either of rhythm, an aimless sort of thing without precise meaning of any sort.
                But measure implies something that can be measured. Today verse has lost all measure.
                (Selected Essays 337)

Paul Mariani, summarizing points from an unpublished essay, "Speech Rhythms," which Williams wrote in 1913, argues that "Williams had already rejected vers libre even as other American and English poets were on the point of discovering it." (William Carlos William: A New World Naked 108) In the essay Williams compared a poem to the sea, which was "an assembly of tides, waves, and ripples." (107) This he set against an imposed metric of iambics that he had previously made himself at home in. Here we see his interest in discovering "natural" rhythms.
     At about the same time that Williams began to try to explain free verse, Amy Lowell was also speculating about it. Lowell's closest collaborator in the science of prosody was a certain Dr. Patterson, of Columbia University, whom she never tired of quoting and whose experiments at measuring metrical intervals with laboratory apparatus she recounted with girlish rapture. In "The Rhythms of Free Verse" (The Dial, January 17, 1918) she said, speaking of Professor Patterson, "The man who could write 'by listening for rhythm in irregular sequences, in the criss-cross lappings of many waves upon the shore, in the syncopating cries of a flock of birds, in the accelerating and retarding quivers of a wind-blown tree, we have found a new form of pleasure,' knows very well what poetry is." (51). Many of Williams's and Lowell's early thoughts about rhythmic possibilities were the same, and it seems likely that he did take certain ideas from her. For many young poets, Lowell's was the only theory available in 1915.
     Like Pound, Williams felt that Whitman had broken open the iambic line and made room for new measures, but that was only the beginning. What was needed was some new unit based on temporal succession rather than syllable-counting or foot-chopping. Williams's misfortune was his failure to realize that English meters had never been anything like as abstract or mechanical as he imagined them to be. His obsessive Americanism--characterized by Allen Tate as a "naive jingoism"--made him certain that any poetry that associated itself with British or European models, myths, or subjects would turn out to have been trammeled by its meters. Even Gerard Manley Hopkins seemed to him "constipated," though he would return to Hopkins more than once in trying to establish the nature of his own "variable foot." (Mariani 598, 681)
 Williams's theorizing about free verse is so impatient, eclectic and at times capricious--and yet in other ways so consistent--and his manner is so insistent, that it is difficult not to be led around by the nose when discussing him. He rejected accentual-syllabic meter--and he rejected free verse. He admired Whitman--but considered him wrong-headed. He detested the idea of a mechanical meter--and he studied the prosody of Robert Bridges, perhaps the most self-conscious metricist ever to achieve distinction. He considered Pound, and especially Eliot, as turncoats who had sold out to European culture--and when he finally hit on his "triadic" or "step-down" poetic line, he could not resist comparing his discovery to Dante's terza rima.
 In his 1948 talk at the University of Washington, "The Poem as a Field of Action," Williams's irritable disaffections are obvious:
 

             I propose sweeping changes from top to bottom of the poetic structure. I said structure. So now you are beginning to get the drift of my theme. I say we are through with the iambic pentameter as presently conceived, at    least for dramatic verse; through with the measured quatrain, the staid concatenations of sound in the usual stanza, the sonnet. More has been done than you think about this though not yet been specifically named for what it is. I believe something can be said. Perhaps all that I can do here is to call attention to it: a revolution in the conception of the poetic foot--pointing out the evidence of something that has been going on for a long time. (Selected Essays 281)

  The one thing that the poet has not wanted to change, the one thing he has clung to in his dream--unwilling to let go--the place where the time-lag is still adamant--is structure. Here we are unmoveable. But here is precisely where we come into contact with reality. Reluctant, we waken from our dreams. And what is reality? The only reality that we can know is MEASURE. (283)

       We have no measure by which to guide ourselves except a purely intuitive one which we feel but do not name. I am not speaking of verse which has long since been frozen into a rigid mold signifying its death, but of verse which shows that it has been touched with some dissatisfaction with its present state. It is all over the page at the mere whim of the man who has composed it. This will not do. Certainly an art which implies a discipline as the poem does, a rule, a measure, will not tolerate it. There is no measure to guide us, no recognizable measure. . . .
 

   Relativity gives us the cue. So, again, mathematics come to the rescue of the arts. Measure, an ancient word in poetry, something we have almost forgotten in its literal significance as something measured, becomes related again with the poetic. We have today to do with the poetic, as always, but a relatively stable foot, not a rigid one. (Selected Essays 339-40)

Although there are hints of his method somewhat earlier, in segments of Paterson, the period 1952-56 found Williams arranging his newly-discovered variable feet in groups of three, which stepped down the page from left to right before returning to the left-hand margin to start over. Here is what may be the best known example of his practice:
 
 

 Of asphodel, that greeny flower, the least,
    that is a simple flower
     like a buttercup upon its

 branching stem, save
    that it's green and wooden
     We've had a long life

 and many things have happened in it.
    There are flowers also
     in hell. So today I've come

 to talk to you about them, among
    other things, of flowers
     that we both love, even

 of this poor colorless
    thing which no one living
     prizes but the dead see

 and ask among themselves,
    What do we remember that was shaped
     as this thing

 is shaped? while their eyes
    fill
     with tears. By which

 and by the weak wash of crimson
    colors it, the rose
     is predicated

         (Collected Poems 238-39)

     Stephen Cushman traces accurately the origins of Williams's notions and the steps by which he settled on them; Cushman also correlates and judges numerous respectable opinions about the variable foot and the triadic, or step-down, line. After dismissing all efforts to discover aural or temporal regularity, he concludes that the effect is mainly visual, a "design that is symmetrically elegant and dignified." (Cushman 92) If I agree with Cushman on the whole, it is not with any intention of denying that rhythm, cadence, the urgencies of the speaking voice are essential to Williams's poetry. But his use of the triadic line and his belief in the variable foot was really no more than a rationalization of a far more complex set of rhythmic events. Williams knew that he had accomplished something new in prosody and he felt uncomfortable at his inability to say just what it was. Confronted with Saintsbury, whose history he owned and even attempted to master, Williams could not help but wish for some explanation of his own principles.
     Williams apparently felt that each unit of each triad contained only one major accent, and that each unit occupied the same amount of time, and that he had therefore achieved his desired measure. Later, he even considered what he had done overelaborate. But if what Williams left us is to be read as he intended, he ought to have specified which syllable in each unit is to receive greatest emphasis, and how many seconds are to be allowed for each line. Since the first line above is longer than any other, perhaps we could take that as a standard for time; the way I read it, it lasts 5.29 seconds. I have no idea which syllable is most important, but I tend to come down hard on "greeny." The line is in fact a passable iambic pentameter. As Williams originally composed it, the poem seems almost to have slipped beyond his control back toward accentual-syllabics; reading the passage from Kenneth Burke's letter of 7 November 1955 to Williams, I wonder if the restless arranging and rearranging that Williams was known for may not have been a way of systematically defeating regular meter as much as it was an effort to "get it right." Allen Tate went one step in this direction in his pentameter poem, "The Mediterranean," by calculatedly making every single line's rhythm different from every other while retaining a governing pentameter. The next step would be conscientiously to make each successive line even more radically unlike the previous one. But let us listen to Burke:
 

   The first time I read the "Asphodel" poem, it seemed so completely dissolving that I actually began to feel faint. All the little nodules of fight had been melted, turned into a succession of breathings-out (each tercet being in effect one such moment). The most disarming kind of utterance one could imagine. Ironically, however, your "titular" moment is iambic tetrameter:
   Of asphodel, that greeny flower,
   I come, my sweet, to sing of you!
  What shall we make of that?
      (East: Williams/Burke Letters)

What indeed! At any rate Williams made sure in his revision that the tetrameter disappeared. Returning to the final version of the poem, and looking at the shortest line, "fill," I find that it takes me about a half-second to get it out, leaving a five-second pause--if we consider the lines isochronic--long enough, I suppose, for the brimming of an eye.
 Recordings of Williams's voice reveal his habit of nervous insistence on certain words; this carries over into his prose writings in his frequent use of italics for emphasis. To the extent that there is any meter to "Asphodel" it probably consists of a loosely-regularized record of his ordinary voice. To say this is tantamount to an admission that there is no meter at all--but it is not to deny Williams a metric; his metric consisted of a continual violation of traditional metrics.
     Williams's true gift was to do the unexpected. His imitators mostly fail because expectations of the unexpected have already been aroused; the same thing happened with English landscape gardening toward the end of the eighteenth century when one architect's surprises were copied by others; after the first "ha-ha" (a wall disguised by a depression in the ground beyond it, so called for the expression of surprise it was meant to elicit), the ha-ha was nothing but a ha-ha. Williams, in his impatient, irritable, and sometimes contradictory denunciations and temporary espousals, found nothing satisfactory in any prosody and succeeded in writing poem after poem that simply was not like any established mode. He was ornery and contrary by nature, though at the same time large-hearted and full of curiosity and enthusiasm. My explanation is not intended to discredit Williams; what he did required the greatest sensitivity to conventional metrics, if not a detailed knowledge of their history. His adolescent compositions in the manner of Keats were not wasted effort; they were the groundwork for his variations. Once he departed from iambic pentameter he was gone forever, unlike, say, Theodore Roethke who made periodic escapes and returns.
     According to Paul Mariani (271) Williams, either in late 1927 or in January, 1928, wrote to Yvor Winters repeating anew what he had concluded long before, that free verse did not exist and that a new measure was needed. Winters, who believed that only finished work was important and that personal correspondence was to be discarded as irrelevant, kept no letters, but Williams mentioned this letter in a journal. The exact dating is important, because the letter coincides with Winters's ceasing to write free verse and his new insistence through doctrine and example that adequate judgment of the subject of a poem was only possible employing shades of meaning afforded by the framework of a fixed meter.
     Winters had been among the first ever to offer intelligent and sympathetic assessments of Williams's work, and was for some years almost an imitator. Their later years were marked by expressions of mutual contempt of the sort that aging generals reserve for one another long after a war; Winters himself told me (in 1963), "Poor old Bill Williams; he didn't have a brain in his head." But in 1927 he might have been more receptive to what Williams said, especially if he already suspected that he himself had been on the wrong track. In the preface to The Early Poems of Yvor Winters (1966) he wrote, "Early in 1928 I abandoned free verse and returned to traditional meters." (13) He asserts that nothing about his new position at Stanford nor any sort of intellectual or religious conversion effected this change, but that he found that he could not hope to emulate the poets he most admired unless he employed regular meters.
     According to Winters, his ideas on the scansion of free verse had been worked out while he was at the University of Idaho at Moscow, which was about 1925-26, after he had already written most of the poems to which the theory applied. About his change of method in 1928 he insists, "My shift from the methods of these early poems to the methods of my later was not a shift from formlessness to form; it was a shift from certain kinds of forms to others." (14) Even at the end of his life, then, Winters remained convinced--or insisted on asserting--that for him free verse had been as orderly as any other and that his poems could be scanned:
 

  When I say that these poems had form, I refer not only to the possibility that the free verse may be scanned by my method, perhaps with difficulty; I refer to the fact that these poems are rhythmical, not merely from line to line, but in total movement from beginning to end, and that the relations between the meanings of the parts is an element in the rhythm, along with the sound. (15)

Winters here recalls the ideas of Amy Lowell, such as her statement, "It is the sense of perfect balance of flow and rhythm." That sentence occurs in the account of free verse that appeared in the 1916 Imagist collection. Similar phrases crept into discussions in the issues of Poetrymagazine that Winters read; as a young man he had been given the freedom of Harriet Monroe's offices in Chicago. Winters makes no reference at all to Lowell in In Defense of Reason, and nowhere else have I been able to find in his work anything more than passing references to her. But he was certainly aware of her as a personage, if nothing else, describing her to classes that he taught as "a great big woman who smoked great big black cigars," and--as I have said--he certainly followed with close attention all the debates over form and formlessness.
     Because Winters was for thirty-five years the most serious defender of conventional metrics during a period (1930-65) that saw the proliferation of whole schools of formless poetry, he is worth attending to. His arguments provide a hidden agenda for the movement now called by many the New Formalism. Among his best-known contemporaries, Allen Tate was almost alone in treating Winters with polite respect, a tribute that Winters did not reciprocate. Nearly everyone else was either a convinced disciple or an opponent--and there were many more of the latter. He made himself the center of a school the mentality of which bears comparison with the monks of the island of Iona who kept the true faith secure in one of the more distant reaches of Britain while the Roman empire collapsed and the barbarians took control of its former provinces.
     Although from his early twenties Winters read and admired traditional poets such as Bridges and Hardy, there was a way in which he backed into English and American poetry. Exiled to the American Southwest by tuberculosis, he took the complete files of Poetry that Harriet Monroe gave him, and subscribed to the Little Review and Others, both packed with the latest avant-garde work. In the isolation of mining towns and provincial state universities he absorbed the most radically experimental poetry of the century; it was as if an architectural student had put himself to school to Gaudi or to Frank Lloyd Wright without a sufficient acquaintance with gothic or Palladian styles--something that I do not doubt has often happened. When he began to write free-verse poems, he was attempting to be radically different from something he did not know enough about to differ from, or at least he seems not at first to have realized that his models owed their success precisely to what they were working against. Also, he had not as yet worked out for himself the practical consequences of certain romantic doctrines, and had not yet observed the effect of those doctrines on the lives of his contemporaries. It is true that his earliest published poems were in regular meters, but he seems at first to have believed that he was embarked on a voyage of discovery when he took up free verse rather than a reaction against those meters. He was not--in his early twenties--as at home in accentual-syllabics as H. D., cummings, or even Williams had been in their formalist juvenilia. Winters abandoned free verse as soon as he understood that he did not know what he was doing--or so I would explain it. A number of the formal poems that he subsequently composed are stiff, stilted, and spoiled by contentious rhetoric; but I am at one with Hayden Carruth in thinking "To the Holy Spirit" one of the high points of modern poetry, along with "The Marriage." Winters probably took the course that was best for him. He  began to perceive some kinds of free verse as evidence of a self-destructive romanticism or spiritual hubris, and this perception coincided with a growing understanding of formal meters and a realization that he had been aiming at variation without having yet possessed a sufficiently well-established norm to vary against. Put another way, in his youth he began to imitate the styles of H. D., Williams, and others without having been immersed in accentual-syllabic meters to the extent that they had, and without feeling the same urgency to escape nineteenth-century metrics as they did. This is not to imply that Winters was deficient in learning; he knew a great deal about the prosodies of the Romance languages, as well as English.
     As I shall show, the radical inconsistency of his "scansion" of free verse is just as irrational as Yeats's explaining human personality in terms of the phases of the moon. The best that can be said about Winters's theorizing is that it demonstrates his early and serious attention to some of Williams's best poems, and provides the occasion for him to say penetrating and memorable things about various others, including his own work. The essence of free verse is that it cannot be scanned into feet; the effect of Winters's method is to prove that free-verse systems of scansion are intrinsically self-defeating. On the level of the individual line of traditional accentual-syllabic meter, a spondee is a spondee because it is not an iamb or a trochee, and the same is true of a pyrrhic. That is, the spondee and the pyrrhic can only be used as substitute feet; they cannot form the basis of a scansion. To talk about the scansion of free verse is as reasonable as it would be to talk about spondaic pentameter, and that is an impossibility. With free verse, the problem is reversed. As Winters says, "the norm is perpetual variation" and the expectation is that each line will be rhythmically different from all the other lines in the poem, or at least from those in the immediate vicinity. At the same time, there will be an implied counterpointing against the possibility of a recurrent rhythm--or, as Eliot put it, the ghost of a meter will advance menacingly or withdraw. The effect can be experienced but is much too complex to be diagramed using anything resembling macron-breve notation. Free verse is free verse precisely because it does not have an identifiable foot and because it does not have a recurrent and nameable line. To put it that way will not, however, deter subsequent projectors whose prosodic speculations have something in common with the occupations of those who inhabited Swift's flying island. There will always be those who have squared the circle and those who are prepared to assert the existence of a free-verse foot; the impossibility of saying exactly what it may be adds to the charm of the idea.
     Winters apparently saw free verse as a new variant of English accentual meter. In "Section II: General Principles of Meter" (from The Influence of Meter on Poetic Convention) he makes an unfortunate connection: " . . . the accentual, or Anglo-Saxon, system, according to which the line possesses a certain number of accents, the remainder of the line not being measured, a system of which free verse is a recent and especially complex subdivision . . . " (In Defense of Reason 106). If what he meant was that free verse is an especially complex form of accentual meter that plays natural accents against the accentual-syllabic compromise, we might be able to see this as a reasonable suggestion. What Winters did not mean to suggest was that Old English alliterative meter was a form of free verse. One has to read Winters with care and note that he says "a certain number" and not "an equal number." The real error, however, is to try to make free verse a newly-established regular metrical system.
     Further on (108), Winters makes a questionable claim for the superiority of accentual meters: "Accent, like quantity, is unlimited in its variations. In practice, the manner of distinguishing between an accented and an unaccented syllable is superior, I believe, to the manner of distinguishing in classical verse between a long syllable and a short." No one knows how the Greeks and Romans read their poetry; Winters takes a schoolmaster's view of quantitative meters as rigidly fixed, "arbitrarily classified by rule." (108)
     To my chagrin, since I would prefer to be in agreement with both of them, I find Winters lining up with Douglas Bush on the question of Milton's prosody in Samson Agonistes, which many readers, including T. S. Eliot, have been willing to see as free verse. Winters is disagreeing with Robert Bridges when he says, " . . . he scans Milton incorrectly, it appears to me, for this reason, and more particularly Milton's later work, which merely represents learned variation to an extreme degree from a perfectly perceptible accentual-syllabic norm, variation expressive of very violent feeling." (109) Milton completely departed from the pentameter line, though the pentameter line made possible that departure. Winters and I may be saying the same thing in different words, and he puts it this way so as to leave himself at liberty to propose a system of free-verse scansion, a system that he would not care to apply to Milton because it would not work.
     The next point that Winters makes, to which I would also take exception, is his statement that the sprung rhythm of Gerard Manley Hopkins and the syncopated effect that is produced by writing syllabic meters in English--without regard to the accent--are the same thing. It is very hard to see how to equate syllabics, which tend to suppress natural stress, with sprung rhythm, which tends to exaggerate it. Nevertheless, as one says when reading Ruskin's denunciations of Michelangelo, it stimulates thought to see such a thing argued.
     I do not agree with Winters's statement that "Wyatt employs the accentual variety of sprung rhythm . . . " (110) Wyatt's roughness--however lovely the poem, "They Flee From Me"--may have resulted from an uncertainty as to how to fuse accent with syllable-count. This was plain to as sensitive a metricist as Auden, among many others. It will not do to compare Wyatt with  poets of later centuries who went counter to the norm. Wyatt had no norm except Chaucer, whom it was difficult to read properly; he was in the process of reinventing the norm, not diverging from it. One might say that Wyatt, by aiming first of all to give us a poem, and only secondarily to achieve perfection in meter, introduced an ingratiating awkwardness to his pentameters--something he never did in tetrameters, where he was completely at home. But it is the charm of actual naiveté, not a stratagem of sophistication: "a sweet disorder in the dress." Pointing out a certain awkwardness in Wyatt's pentameters (as compared, say, with Sidney's) does not diminish the stature of the best English poet of the earlier sixteenth century, any more than it would damage Giotto to be compared with Raphael.
     Winters had a way of inventing metrical criteria to justify his preferences--as we all do. In praising Barnabe Googe's graceful line, "Fair face show friends when riches do abound," he says: "Here the accentual weight of the first and third places is increased to equal approximately the weight of the second and fourth; we might describe the first two feet as spondaic, except that, as there is no compensatory pair of pyrrhics, two extra accents are introduced into the line, with the result that the accentual measure is abandoned and we have no measure left save the purely syllabic." (110-11) Implied in this statement is some strange doctrine that requires that every spondee in a line of iambic pentameter be balanced by a pyrrhic; one would like to stop the conversation and inquire what to make, then, of Byron's line, "Roll on, thou deep and dark blue ocean, roll!" and of Donne's, "For God's sake hold your tongue and let me love." There are moments when Winters's criticism seems as relevant to his practice as a poet as Hart Crane's proclivity for dropping typewriters out of high windows was to his.
 But his discussion of prosody is serious enough. Close attention to placement of accent in conventional meters turns out to be preparation for the arguments advanced in Section III, "The Scansion of Free Verse." Winters opens, as usual, with seeming lucidity: "The foot which I have used consists of one heavily accented syllable, an unlimited number of unaccented syllables, and an unlimited number of syllables of secondary accent. This resembles the accentual meter of Hopkins, except that Hopkins employed rhyme." He ought also have added that, however extravagant each foot may be, there are equal numbers of feet in each line of a Hopkins poem, or in corresponding lines of their stanzas, except at the end of his "curtal" sonnets--those in which he cut the sonnet form short.
     Winters continues by setting out his method of notation. This consists of marking primary stresses with "double points," which look like quotation marks, and secondary stresses with "a single point," or apostrophe. Initially he makes no statement as to whether his own free verse lines will contain one or two feet; the example that follows contains both; or at least some lines have a single primary stress, while others have two. But wait: "Since a line which is complete metrically may for the sake of emphasis be printed as two lines, I shall place a cross-bar (/) at the end of each complete line." So some lines are really not lines. Let that go, however; let us see if--when we allow the "cross-bar" to terminate a "line"--we see anything symmetrical. And so we do--but only for the first five "lines." In most of the poem a "line" includes two "primary stresses," which may appear (typographically) on two successive lines which are really one line. But when we reach "line 6" (the seventh line), we see only one "primary accent" marked.
 

    "Quod Tegit Omnia"
 

       "       "
1   Earth darkens and is beaded/

 "       '          "         '
2   with a sweat of bushes and/

             "           '
3   the bear comes forth:

              "  '
     the mind stored with/

             "      '            "         '
4   magnificence proceeds into/

              "       '          "       '
5   the mystery of Time, now/

        '       '        "
6   certain of its choice of/

         "               '    "      '
7   passion but uncertain of the/
        '          "
8   passion's end.

  [query: why two different scansions for "passion"?]

         "
    When/

          '       "        '                 "
9    Plato temporizes on the nature/

          "          '           "
10   of the plumage of the soul, the/

  "    '    "  '
11   wind hums in the feathers as/

   '         "         "  '
12   across a cord impeccable in/

           "          '         "    '
13   tautness but of no mind:/

            "
14      Time,

            '       "              '
    the sine-pondere, most/

         '       "               "        '
15   imperturbable of elements,/

             "       '       "
16   assumes its own proportions/

        "        '         '     "
17   silently, of its own properties--/

            "         '              "      '
18   an excellence at which one

          "
       sighs./

           "           '
19    Adventurer in

  '  "           '
 living fact, the poet/

           "           '           "
20   mounts into the spring/

  '         "       "
21   upon his tongue the taste of/

          "       '        "       '
22   air becoming body: is/

    "               " '
23   Embedded in this crystalline/

   "     '         "
24   precipitate of Time./

  (In Defense of Reason 113)
The inconsistencies of the method employed here seem obvious to me, though others may choose to remain convinced by it. In advance of setting out the scansion of a second poem, Winters clears up one anomaly : "The imperfect lines (unassimilable half-lines) are marked with a single asterisk." So perfect lines contain two "primary stresses," though the lines may be printed as two lines, each of which only has one "primary stress." But sometimes these lines with only one "primary stress" are really "imperfect." Then we reach the end of the poem, and he reconsiders:
 

   This poem is marked, as I have said, as if it contained two feet to the line. It is possible, however, to regard the poem as having a one-foot line, in which case the lines marked with the single asterisk and those unmarked are regular, and those marked with the double asterisk are irregular. The two-foot hypothesis involves the smaller number of irregular lines, and it would eliminate for this poem a difficulty in the matter of theory; to wit the question of whether a one-foot line is a practical possibility. (115)

      /       /      *      *    *       /  *
  Earth dark | ens and | is beaded
 

This may seem to offer scant likelihood of a regular meter, but it is not that far from the rhythm of opening lines in poems whose regularity no one can question, such as Campion's,

   *   /        /   *       /      /  *
  I care | not for | these ladies
 

or one by Housman--

     /    /        *   /      *     /   *
  Ho, ev | eryone | that thirsteth
 

In Winters's poem, though, the trimeter is promptly obliterated or at least evaded, returning only two or three times, and then only tentatively, such as:

    *      /       *  /   *   /      *
  The mys | tery | of time now
 

But the last line of the poem is--despite his wilful marking of it--a perfect iambic trimeter:

      *  /  *    /     *   /
  preci | pitate | of time
 

To some degree his free verse is contrapuntal to an implied background meter, but sometimes he simply trusts his instincts as a poet and breaks his lines accordingly.
     At about this time he had come to realize what consequences could ensue from a thoroughgoing application of antirational Emersonian principles; both he and Allen Tate attempted to warn Hart Crane to abandon these self-destructive romantic dogmas--with the consequence that both of them have been blamed for Crane's subsequent suicide. (To understand precisely how Winters explained the tragedy of Crane's life, one must read his essay on that poet.) His fear of the dissolution that romantic doctrine had worked on Crane helps to explain his rejection of free verse in which, as he rightly perceived, he was drifting toward an impulsive and irrational metric. Although Winters never did recant his willful and inconsistent attempts to impose an orderly scansion on free verse, the fact that after 1930 he never composed anything except in regular accentual-syllabic meters is a sufficiently convincing rejection of the earlier method. His subsequent achievement is surely on par with, say, the best of the French Parnassians of the nineteenth century. To some this may seem either exaggerated praise or an ironic dismissal, but I mean it in perfect seriousness. Leconte de Lisle's poem about the elephants remains one of my favorites--as it did for Winters, and his own work displays an equal formal dignity.
     The tools for satisfactory analysis of free verse do not exist; even in scansion of accentual-syllabic meters, the idea of "substitution" of a foot is relatively recent, and fruitful new concepts--the implied offbeat and the unrealized beat--are known to few of those who teach poetry or write about it. The limited usefulness of the "foot" concept vanishes in the discussion of free verse; here we have "substitution" of entire lines, and the difficulty of accounting for what happens approaches that of the three-body problem in celestial mechanics. If there is a foot in free verse it is the entire line. Winters may have been verging on that realization.
     Much more successful are his discursive accounts of his practice:
 

   My own free verse was very often balanced on this particular tightrope. [By which he means a tension between iambic pentameter and a more irregular accentual meter that runs against it.] During the period in which I was composing it, I was much interested in the possibility of making the stanza and wherever possible the poem a single rhythmic unit, of which the line was a part not sharply separate. This effect I endeavored to achieve by the use of run-over lines, a device I took over from Dr. Williams, Miss Moore, and Hopkins, and by the extreme use of a continuous iambic undercurrent, so arranged that it could not be written successfully as blank verse and that it would smooth over the gap from one line of free verse to the next. (116)

Amy Lowell and others among the Imagists had called for the whole poem, or at least a strophe thereof, to serve as a self-contained unit; Winters's reading in those early discussions is evident here. We also see in Winters's discussion the idea of the line as the block from which the strophe may be composed, much as the foot in older scansions was a distinguishable part of the line, though subsidiary to the line. The "iambic undercurrent" of which Winters speaks, which never can quite be resolved into regular pentameters, seems to me quite similar to what I mean when I argue that, line by line, a good free-verse poem runs counter to the possibility of some regular meter. James East has suggested the image of a palimpsest, with erasures of visible measures faintly glimpsed behind the bold text that holds our attention. I am not speaking of writings that are the effect of incompetence, but rather those by poets who are so thoroughly at home in various meters that they can extemporize and divagate; good free verse resembles the lost art of eighteenth-century musical improvisation--lost because it could not be recorded. It also resembles some varieties of jazz.
 In the rest of his discussion, Winters applies his concepts and his notation to other free-verse poets whom he admired. His taste seems impeccable, but his method is an unsuccessful attempt at schematizing the remarkable effects achieved in these poems of Williams, H. D., Marianne Moore, and Richard Aldington. "The free-verse foot is very long, or is likely to be," he says. "No two feet composed of different words can ever have exactly the same values either of accent or of quantity." (123) In that case there is no free verse foot; the whole idea of the foot is to identify units that are essentially the same, not ones that essentially different. What we have in free verse is continual variety, continual evasion of the norm. At the end of this section Winters admits as much:
 

  The free-verse poet, however, achieves effects roughly comparable to those of substitution in the old meters in two ways: first by the use of lines of irregular length, a device that he employs much more commonly than does the poet of the old meters and with an effect quite foreign to the effect of too few or of extra feet in the old meters; and, secondly, since the norm is perpetual variation [my italics], by the approximate repetition of a foot or of a series of feet. (129)

This idea, put in a different way, appears to be what Amy Lowell arrived at as a consequence of her collaboration with her scientific friend:
 

  I quite agree with Dr. Patterson that "vers libre is at its best when syncopating experience predominates." In my "Tendencies in Modern American Poetry," I spoke of Richard Aldington's and "H.D."'s practice of vers libre as always following the syncopating experience. These poets arrived at their conclusions quite independently. ("The Rhythms of Free Verse" 53)

    What Winters called the norm of perpetual variation, what Lowell named syncopating experience, and what Eliot referred to as the ghost of some established meter, are all different ways of describing the same thing: variation in which the entire line constantly plays against rhythmical expectations, not merely the expectation of a succession of iambs but also of any consistency in line length. This, I take it, is what Williams meant when he spoke of an "ethereal reversal." Even Whitman's line can be seen as an extravagant exception to the deadening regularity of American iambics, which had always been Johnny-come-lately imitations of some previous English style. Williams was quite correct to feel as he did that it was high time for a prosodic Declaration of Independence.
    Winters, however, concluded that free verse probably could not attain the precision of traditional verse because in regular meters "each variation, no matter how slight, is exactly perceptible and as a result can be given exact meaning as an act of moral perception." At this point one must begin to consider what is morality and what is perception; happily, I am under no obligation to explain these matters.