Quantitative Verse for English


        My primary intention in this essay is to convey a purely technical point: that quantitative verse is possible in English. By "quantitative verse" I mean a system of meter and scansion, like that of the Greeks and Romans, that is based on the length of syllables as they are pronounced, without regard to stress or accent. I would have begun simply by launching into the details of this system, taking for granted that anyone who would bother to read an essay on English verse technique shares with me the basic assumption that technical matters are interesting for their own sake. I am advised this is not so. A few words, then, on how I’ve come to quantitative verse and why I think it might be ofwider interest.
        For me, one of the pleasures of poetry has always been the technical triumph ofmaking the difficult seem easy - making words go in a pattern not natural to ordinary speech, while, at the same time, making such pattern seem natural and effortless, and with the hope that the pattern heightens rather than impedes the meaning of the words themselves. I don’t think that sort of pleasure is confined either to poetry or to me - think of the line in A River Runs Through It: “to my father, all good things - trout as well as eternal salvation - come by grace, and grace comes by art, and art does not come easy.” You can no doubt think of plenty more examples in other disciplines (they are called “disciplines”), and in poetry itself. Pretty much anything in Horace’s Ars poetica, or Frost’s “writing poetry without rhyme is like playing tennis with the net down.”
        But then Yeats to the effect of “every poet eventually gets tired of rhyming mountain and fountain,” and so to the first part ofmy point: I felt that the existing forms of English poetry weren’t difficult enough. This isn’t to say that writing good poetry is easy - it obviously isn’t - or that there isn’t room for almost unbounded technical accomplishment within existing forms - there is. The bar to writing good verse is as high as ever; the bar to writing adequate verse is too low.
        I wanted a more difficult form for several reasons. Firstly, I was exhausted with trying to determine anew, for each poem, what form might be appropriate. I envied the 18th century poets who could just choose heroic couplets or iambic tetrameter or whatnot and go to town. Secondly, I wanted a difficult form so that any coherent statement that met its demands could be considered an achievement. In other words, I wanted sometimes to write for the sake of satisfying technical demands rather than for the sake of saying something intrinsically interesting or new - which, as with the first point, wasn’t and isn’t always (or maybe ever) within my capability. You’ll see, then, that the desire for technical difficulty comes partly from the negative desire to avoid the difficulty of self-determination, which is inherent in prevailing notions of organic form. On the positive side, I thought and think that overcoming technical difficulty is beautiful in itself, as I’ve described. I also think that difficulty overcome might open new possibilities ofmeaning in even the substantive matter of the verse. The example here might be Bishop’s “One Art,” or, indeed, any of the well - handled villanelles and sestinas in our language.
        The above arguments apply to technical difficulty in general. Insofar as quantitative verse is difficult, they support the attempt to write it. But, on the one hand, we could invent arbitrarily difficult verse forms and support them on the same grounds, or, on the other hand, we could perhaps practice quantitative verse until it became as trivially easy as heroic couplets. The second way to justify the project, then, relies on the merits unique to quantitative verse, without regard for its difficulty or novelty. As a historical matter, most of the few preceding attempts to write quantitative verse in English have been motivated by the desire to more closely imitate the classics. That’s probably worth something in itself if you believe in the value of poetic tradition, historical consciousness, and so forth. But imitation of the classics isn’t worthwhile just because they’re classics but because they’re good (hence classic), which brings you back to the same ground.
        The unique merit of quantitative verse lies in its use of duration as an organizing principle. In theory, you can organize verse by any feature, phonetic, visual, or otherwise - by first letter (alliterative verse), by pitch and syllable count (accentual-syllabic), by syllable count only (syllabic), by extrinsic qualities (acrostics, e.g.). These features, as a basis for organization, range from reasonable to entirely arbitrary. Duration is unique amongst them because it is a temporal quality. Quantitative verse, therefore, has a necessary relationship to time. To borrow a musical analogy, I think we perceive rhythm as more fundamental than pitch or timbre or any other feature of a sound. It makes sense to me, then, to base verse patterns on duration, the fundamental feature, and let the remaining features vary above it. Even if you could do the opposite - specify the pattern of a non-essential feature and let the fundamental feature vary below it - it would be counterintuitive: perhaps rather like Schoenberg’s tone rows when compared to the conventional use of time-signature, tempo, and rhythm. 
        As a final point, I’d say that using or learning to use quantitative verse gives us, at minimum, the capacity to hear and use new sound-features in poetry. That’s not an argument for quantitative verse per se - once you can hear both quantity and accent, you can hear them equally whichever feature you use to organize your verse - but it is a nice side effect that hopefully holds some attraction even to those who aren’t convinced by the other arguments.
        A word on my predecessors: the project of writing quantitative verse in English is not a new one. It has proceeded on and off - mostly off - since the Renaissance, and it has attracted all sorts of cranks and crackpots through the centuries. Though the project has repeatedly failed to gain mainstream attention or acceptance, it hasn’t been useless. I am convinced that two (I believe exactly two, but there may be someone I’ve overlooked) people have independently solved the technical problem of quantitative verse in English. First was Thomas Campion of the English Renaissance, whose much-misread Observations on the Art of English Poesie seems to me to establish an essentially correct view of quantity in English. Campion was part of a vogue for quantitative verse that included Spenser, Sidney, and others, but their work had no lasting effect. Then, second, much later, Charles Johnson Stone wrote a pamphlet On the Use of ClassicalMeters in English, which corrects the many errors of previous experiments and sets out a somewhat fragmentary but ultimately usable system of quantity in English. Stone died young. His work was reprinted, annotated, and adapted by his friend Robert Bridges, but, once again, the project lapsed. I rely on Campion and Stone for their examples and rules (the ones you see below are synthesized from their observations), but the most important thing they give is a clear-sighted separation of the quantitative problem from the errors and misconceptions that surround it. I hope this essay lays out simply and with great clarity the comprehensive solution to quantitative verse that they were working towards.


        Consider music. Any given sound has the three attributes of duration, pitch, and volume. These qualities are separate, and vary separately. You can think, for example, of a trumpet playing quarter notes up and down a scale at mezzo-piano. This is variation of pitch with a constant duration and volume. A crescendo on successive B-flat quarter notes would be variation of volume with a constant pitch and duration. A one-note samba at mezzo-piano would be variation of duration with a constant pitch and volume. Hopefully it's easy to see, in the musical context, how the three attributes are all present in a given sound and yet are all independent of each other.
        Now consider speech. Speech is sound, and so speech sounds, too, have the three attributes of duration, pitch, and volume. Mapping those musical terms onto the terms of prosody, we say that duration is called "quantity" or "length"; pitch is called "accent"; and volume is called "stress." I believe that mapping to be accurate to fact, though I realize it goes against almost every received notion about English prosody and the nature of the language. In existing discussions of prosody, the terms "stress" and "accent" are often used interchangeably, and the terms "long" and "short" are sometimes used to refer to "accented" and "unaccented" (or even "stressed" and "unstressed"). Such terminological confusion must be avoided - "quantity," "accent," and "stress" are three separate attributes and should not be conflated. In practice, this means insisting on differentiating between "quantitative scansion," which is described in terms of "long" and "short," and "accentual scansion," which is described in terms of "accented" and "unaccented."
        Before we can use a sound-quality as a formal aspect of poetry, we need to learn to hear it. Because the vast majority of English poetry is written in an accentual or accentual-syllabic meter, we can hear and identify accent effortlessly; scanning or writing in iambic pentameter, for instance, is a trivial exercise. (That recognition, of course, is learned, not innate. New poets have to be taught to hear accent. The French, whose poetry is syllabic, deny that their language has any accent at all, though of course an English speaker hears it readily. The same thing happens, I imagine, with the "tones" in Mandarin. Or, on a smaller scale, a language learner needs to distinguish minimal pairs of sounds not present in the native language: French speakers and "bit-beat" or English speakers and "boue-bu.") As a musician would say, we need ear training. Our goal is to distinguish quantity, accent, and stress in the context of English speech. This requires careful listening.
        We'll deal first with the proposition that musical pitch maps onto "accent," or in other words, that the English accent is primarily a pitch accent. This was the hardest proposition for me to believe when I first encountered it, though, as I've said, I now believe it to be correct. There might be an admixture of volume/stress in what we hear as accent - which the general interchangeability of the terms "accent" and "stress" in ordinary English prosody would go some way to support. Incidentally, it's not a proposition absolutely central to the main argument about quantitative verse - which is to say you could disbelieve it and still accept the rest. You need, at minimum, to grant that accent can vary separately from quantity (i.e. duration), even if you don't want to say that accent is purely or mostly pitch.
        I take for granted that you can hear and identify English accent where it occurs - our goal is to clarify the nature of this accent. The best way to do this is to listen carefully to ordinary words with an ear for accent and pitch. Take a word and say it very slowly, listening to pitch. "Ordinary." Then say the same word with sung nonsense syllables. "YaNyaNyaNya." Slow it down and sing it again. Say it again; sing it again - going back and forth between the speech sound and the music sound, with a careful ear on the element of pitch, as independent of the word as you ordinarily process it. Do you agree that "ordinary," which scans AccNunaccNaccNunacc, is spoken with a higher pitch on "ord-" than "-i-" or  "-y"? There's more here to explore, but as it's tangential to the main argument, we can bracket the question and move on.
        The mapping ofmusical volume onto "stress" means only to suggest what happens when you emphasize a particular word in a sentence, as though to put it in italics. "Not that one but *that* one." Even then volume stress is not used alone but in combination with pitch and duration. Otherwise stress, properly so called, is unimportant in ordinary speech.
        Finally, and most importantly, we're going to train the ear to listen for quantity, which, as stated above, is equivalent to musical duration. I think it's easier to hear quantity than pitch/accent, because here we're starting with fresh ears. English verse doesn't deal with quantity, so we don't have the same preconceived notions about syllable length that we have about pitch and accent. Again, though, the method is to go back and forth between music and speech until we can isolate the attribute of quantity and consider it separately from the other attributes. To borrow Stone's example: "Quantity." "Quiddity." "Quantity. Quiddity." "Da-da-da. Da-da-da." Go back and forth between saying "quantity" and saying nonsense syllables "da-da-da" as if you were saying "quantity." Then do the same with "quiddity." Then compare the two. Listen for the length of time it takes to say each word - then each syllable. Do you hear the contrast between the length of the first syllable of each word? Try saying both words as fast as you can: which takes longer to say? Both words are accented strongly on the first syllable - if you find this distracting, chant the word on a single pitch - both as the word and as nonsense. Focus on duration, not pitch or volume. Hopefully you'll agree that the first syllable of "quantity" is much longer than the first syllable of "quiddity." ("Quantity," in fact, scans long-short-short and "quiddity scans "short-short-short." This is discussed later.) It helps to have someone work with you in person, but this is a strong example. Most people I've talked to have been able to pick up the difference pretty readily. With open ears and careful attention to how the words actually sound, you should be able to hear the difference.
        Once you've got some idea of quantity in that "minimal pair" context, it's time to listen to other words and syllables with the same ear. You're just practicing hearing quantity in the same way you once, in high school English, maybe, practiced scanning Shakespeare sonnets to hear iambic accents. In reading and writing verse with the new appreciation for quantity and accent as two separate attributes of sound, it might be useful to use two different sets of scansion marks, and use both in parallel so that you can see the independence of each. I use accent grave-dot for accentual scansion of accented-unaccented, and reserve the traditional em dash-breve for quantitative scansion of long-short. I will be using that distinction for the remainder of the essay.


        As you practice, you'll discover that there's a whole range of syllable lengths, from very short to very long, and everything in between. This shouldn't surprise you - accents, too, vary from weak to strong, and we can hear that in practice, even ifwe don't draw formal technical distinctions between them. (I.e. we only say "accented" and "unaccented." There's no technical vocabulary to distinguish more degrees of accent than that.) In quantitative practice, every syllable will be described as either "long" or "short" - two categories - though, as with accent, there is lots of underlying variability to the true length of the sound. The rules for classifying "long" and "short", once you've learned to hear quantity, should be pretty easy to learn and apply - all they do is confirm what you're already hearing and help to differentiate edge cases. Every rule applies at the syllable level; all rules apply to the syllable as pronounced, not as spelled. The rules are as follows:

  1. Some pure vowels, and all diphthongs, are long "by nature." If you remember your elementary school phonics, these are the vowels that say their name, as in the sequence: bait, beet, bite, boat, boot. Then also: bout, butte, boil.
  2. Remaining vowels are short by nature. Bat, bet, bit, bot, but.
  3. Short vowels are made long "by position" when they precede two consonant sounds: "bet" (short) becomes "best" (long), "bot" (short) becomes "botch" (long), etc. Consonants count across word breaks, so "but" (short) becomes "but go" (long, long).
  4. Consonant sounds are counted phonetically. Spelling is treacherous in this regard. "th" is spelled with two consonants, but is only one consonant sound. So too "sh," "wh," "ph," etc. "Ng" is one sound when soft as in "singer" but two when the "g" is sharply pronounced as in "finger." A good rule, where the case is doubtful, is to think about how the mouth moves to make the sound. If you have to reposition your mouth to pronounce the sound, it's two consonants.
  5. Some little grammatical words, aka proclitics, can be scanned short to reflect their pronunciation in speech, even though they are properly regarded as long. "To," "a," and "the" are examples that come to mind. There's room for discretion based on your ear, but be careful not to confuse accent and quantity in this context. "To," unaccented, becomes "tuh" in speech - it's short not simply because it's unaccented but because the lack of accent leads to the vowel changing its sound. "Me," "we," and so forth aren't shortened in this way - their vowel stays pure no matter how unaccented they are.
  6. Elision is generally not available in English. Two adjacent vowels are always separated, in speech, by a consonantal "w" or "y" sound. "To any" is pronounced "toowenny," etc.

        The above rules are general. Once you've learned to hear quantity, you should be able to tell for yourselfwhether a given syllable is definitely long, definitely short, or somewhere in between. It's the between cases that require reference to the rules - which, of course, exist for the sake of consistency. The idea is that any given syllable, as pronounced, is always used the same way. Alternately, though, you could imagine a practice in which syllables of ambiguous length can be used as either long or short in the meter. Campion seems to suggest this; he has an intermediate length category he calls “common.” In practice, there are a few open questions in the systematization of English quantity. The ones I've encountered most often are:

  1. Does the same consonant twice across a word break still make long "by position"? Consider the difference between "but a moment" - "but you" - "but time." (Or “been away” - “been too long” - “been nearly a year”; this applies to all consonants, not just “t”) Using the rules above, the first example keeps the "but" short, as it is by nature. The second example makes the "but" long by position. What of the third example? The mouth doesn't have to move between the "t" sounds, so it's possible to pronounce them as one, though properly speaking you should enunciate both. I think theoretically I prefer to enunciate both and count it as a double consonant making long by position; in practice I think it's useful to be able to keep the syllable short without having to start the following word with a vowel.
  2. Is the sound "ee" long, short, or variable? In the basic rules, it's stated as long. But you can hear a difference between the length of the sound in "green" and in "any" or "goodly." It almost seems there are two sounds, the full long vowel in "green" and the short, unaccented suffix "-y" or "-ly". In practice I've mostly used that distinction.
  3. What does "r" do? Is it a consonant sound or is it a vowel modifier? Where "r" precedes another consonant it's a non-issue: the syllable ends up long either way. But the "-er" or "-or" ending might cause problems. On the one hand, "riverine" has a much shorter "er" than "river-bank," which suggests that "r" here is a normal consonant subject to the positional rule. But in "if ever a woman," the "er" sounds distinctly longer than the other short syllables. Does the short "e" at the heart of the "er" sound govern? Or does the "r" lengthen the vowel's nature?


( - )
( - )
( - )


( u u )
( - u )
( u - )
( - - )


( u u )
( - u )
( u - )
( - - )


        The above description has hopefully given you a basic understanding of how quantity appears in English. Quantity, like any other attribute of the language (phonetic or visual), can be used as a formal principle to organize verse. Quantitative verse is verse that uses quantitative meters: set patterns of long and short syllables that determine each line. The accent is left to vary above the quantitative base. You'll see that this is simply the reverse of the usual accentual-syllabic meters, in which the pattern of accents is set and the quantity is left to vary. In both cases, there's no need for the quantity and accent to align - indeed, there's great opportunity for poetic effect in playing the accent off the quantity. It's this possibility of counterpoint that makes quantitative verse so promising in English. Because we're already sensitive to accent, learning to hear and use quantity gives us a second layer of sound-pattern to pay attention to at the same time we hear the first. This might be as close as speech comes to "harmony." 
        In the following section, I’ve scanned some well-known lines according to both accentual-syllabic and quantitative principles. You can look at these examples for further practice in hearing quantity. You can also listen for the effects created by the interplay of quantity and accent. The “combative accent” when a quantitatively short syllable carries accentual stress I think is especially beautiful. (It's interesting to consider to what extent we already have some unacknowledged sensitivity to quantity in English. We already recognize "whose woods these are I think I know" - all long syllables - as very smooth. If I were a PhD candidate, I think a promising thesis topic would be to reexamine English verse, especially Tudor and Elizabethan, to see whether or not quantity was an obstacle to the early regularization of the iambic line. Might the "roughness" of some of Wyatt be explicable this way? Alternately, you could extend the field and look at great English poetry more generally and see whether a patterning of quantity doesn't underlie some notably beautiful passages.)


        It should be clear by now that we've moved well beyond the usual heresy of suggesting that English writers must take "long" to mean "accented" and "short" to mean "unaccented." There are many attempts, starting with Southey, to write in classical, quantitative meters despite having no understanding of quantity. These writers simply write accentual-syllabic verse on a pattern suggested by the classical meter. That would be a valid project as far as it goes - the simple exercise of inventing a new accented-unaccented pattern to govern a line - except for the confusion it has wrought in the whole field. The adoption of terms from classical quantitative prosody has made this infinitely worse - so that now we must rigorously distinguish between an "accentual iamb" and a "quantitative iamb," and so forth. You’ll likely find it useful to scan using two different sets of scansion marks, as I’ve been doing throughout this essay.
        Most of the quantitative verse written in English to this point has simply borrowed classical meters. Given that classical imitation has historically been the primary motivation of the quantitative verse project, that makes sense. It's very convenient to have the meters ready-made. Besides, the Greek and Roman poets presumably settled on the meters they did because those particular patterns of long and short sound good. There's room for an English poet to develop new quantitative meters that are tailored to English - Campion seems to have attempted it - or for each poem to determine its own pattern of long and short syllables. It's a new field; you can see how the potential for development mirrors the historical development that has in fact taken place in the accentual-syllabic meters. One could even imagine a hybrid accentual-syllabic-quantitative system in which every line had a specific pattern of accents and a specific pattern of quantity. Of course that would be incredibly demanding. Quantitative verse is hard enough to write as it is, not because it's inherently more complex than accentual-syllabic verse (after all, both systems only specify the patterning of a single feature on a given syllable count), but because we lack the lifelong practice assessing and deploying quantity that we have in the case of accent.
        I've put together a few example pieces of quantitative verse, each of which uses a different classical meter. The first is a dedication in hendecasyllables, after Catullus I. The second, an imitation ofHorace's Ode I.19, uses the meter of its model, alternating "glyconic" and "asclepiad" lines. The third, in Sapphics, is a response to Sappho's "Anactoria Poem." The fourth is another asclepiad-based meter taken from Horace. Those meters are written out below for your convenience; the first three poems are almost strictly metrical, but the fourth takes some liberties, especially in the second half of the lines.