Tag Archives: prosody

Poem: (Carrion Comfort)
Poet: Gerard Manley Hopkins

Way back in high school, before I knew anything about poetry (though even then I styled myself a poet, in hindsight a “poet” of the kind you are liable to find on wordpress), I had my first encounter with Gerard Manley Hopkins. Even then, (Carrion Comfort) struck me, though the basic overview of sprung rhythm we received was inadequate to allow me to unlock its rhythm (and, to accept what blame is mine, I lacked any interest then even in learning to scan simpler meters).

Now almost a decade later, I’ve learned and mature enough to take a crack at it. Though a sonnet, (Carrion Comfort) is in hexameter rather than pentameter (we’ve seen this before). In this, I follow Hopkins himself, and disagree with Edward A. Stephenson, which disagreement I note since it was Stephenson who isolated the nine principles of sprung rhythm that make clear its workings. I owe him a great debt, but in this instance he is wrong.

In the scansion below, I have marked primary stresses with an acute accent and secondary stresses with a grave accent (though note that the accent on “bruisèd,” line 7, indicates pronunciation and not stress). Outrides are underlined; all outrides follow Hopkins’ own markings. Afterwards, I discuss some interesting features that emerge.

Nòt, Í’ll nòt, cárrion cómfort, Despáir, nòt féast on thée;
Nòt untwíst — sláck they may bé — thése làst stránds of mán
In me ór, mòst wéary, crý I cán no móre. I cán;
Càn sómethìng, hópe, wísh dày cóme, nòt chóose nòt to bé.
But áh, but Ó thou térrible, why wóuldst thou rúde on mé
Thy wríng-wòrld ríght fòot róck? lày a líonlìmb agáinst me? scán
With dárksome devóuring éyes my brúisèd bónes? and fán,
O in túrns of témpest, me héaped there; me frántic to avóid thèe and flée?

….Whý? That my cháff might flý; my gráin lìe, shéer and cléar.
Nay in áll that tóil, that cóil, since (séems) I kíssed the ród,
Hànd ráther, my héart lò! lápped strèngth, stóle jòy, wóuld làugh, chéer.
Chèer whóm though? the héro whose héaven-handling flúng me, fóot tród
Me? or mé that fóught him? O whích one? is it éach òne? That níght, that yéar
Of nów dòne dárknèss I wrétch lày wréstling wìth (my Gód!) my Gód.


  1. Hopkins in this poem takes full advantage of the freedom that sprung rhythm offers as a logaoedic meter. The base rhythm in sprung rhythm is a mix of trochees and dactyls, with monosyllables and first paeons (one stress followed by three slack syllables) as rarer exceptions. The danger of such a meter is that it becomes simply unprincipled, haphazard. Here that does not happen. The dominant rhythm is trochaic and often claustrophobic (see my second comment below), with dactyls used to create breathing space. To give just one example, in the first line we have three trochees (I’ll not; -pair not; feast on) and, ignoring for now the rove-over syllables, a monosyllable (thee). There is also the line-opening anacrusis (Not), which in my view takes a secondary stress. Two of the trochees have secondary stresses (I’ll not; -pair not), making the line very heavy. Yet in the middle we have two dactyls (carrion; comfort, de-). After the inward-looking insistence of “Not, I’ll not,” Hopkins turns outward to his oppressor. The rhythm captures this outward turning beautifully.

  2. Hopkins also uses dipodic feet to excellent effect in modulating the rhythm, even where the pattern of primary stresses and slack syllables (counting secondarily stressed syllables as slack) is identical. Consider the endings of line 3 (cry I can no more. I can) and line 11 (lapped strength, stole joy, would laugh, cheer). Both have the following stress pattern: / u / u / u /. In line 3, however, every slack syllable truly lacks a stress, whereas in line 11 every slack syllable takes a secondary stress. Both are incredibly heightened moments of the poem (though what isn’t, in Hopkins?), yet they achieve very different effects. In line 3, the speaker’s cry comes through clearly and cleanly, is uncluttered and musical. By contrast, in line 11, the effect is that of a slow build, two steps forward (the primary stresses), one step back (the secondary stresses), finally culminating in that crucial word, “cheer,” which brings us to the central tension of the poem, “Cheer whom though?”, about which more in the next comment.

  3. It is an interesting feature of the poem that the word “not,” which appears six times in the first four lines, always takes a secondary stress and not a primary stress. This runs contrary to my first inclination on reading the poem, but in the context of the whole it is actually quite crucial that “not” never takes a primary stress. Notice what syllables are forced to take a primary stress because of this: I’ll; (des)pair; feast; (un)twist; choose. All concern either the speaker (I’ll) or the speaker’s actions (the rest). The stressing thus points to the self-centeredness of the speaker, setting up the later question, central to the entire poem: “Cheer whom though?” And notice also that once he asks this question, once he asks whether it is himself or his oppressor whom he cheers, how the word “me” takes or fails to take a stress. When considering his oppressor (heaven-handling flung me, foot trod / me), it does not. But when he considers himself (or me that fought him), it does. Hopkins’ genius shows itself in the littlest things.

I mentioned in my first attempt at a proper scanning of Gerard Manley Hopkins’ “No worst, there is none…” that I did not really understand the purpose of outrides, and that I used them in line six out of felt necessity. This morning, in Edward Stephenson’s essay “Hopkins’ ‘Sprung Rhythm’ and the Rhythm of Beowulf,” I found an example of an outride used to include an extrametrical syllable, and this confirms that I used them incorrectly in my previous attempt. Here is a revised scansion correcting the error, with changes from the previous attempt marked in red. Commentary to follow:


I have changed the way I scan seven of the feet: one foot in line 2, two feet in line 6, two feet in line 10, and two feet in lines 12-13. I will explain each change.

Line 2. The change here does not reflect a change in the rhythm or the distribution of stresses, but rather my new understanding of outrides. Previously, I had scanned “More pangs will” as a dipodic foot with a primary stress on “More,” a secondary stress on “will,” and no stress on “pangs.” I assumed “pangs” did not take a stress because it has to be muted relative to the stressed syllables that flank it. In fact, however, it is a quite heavy word, and something is lost in marking it simply as unstressed. What Stephenson’s essay taught me is that this is precisely the context in which a syllable should be treated as an extrametrical outride (his example is “shake’s heel sweeps” from The Windhover, with “heel” as the outride).

Line 6. The recognition that “pangs” is an outride in line 2 has as its corollary the recognition that “on an” in line 6 is not a pair of outrides. But this creates a problem. They cannot be attached to the prior foot (“world-sorrow”) for two reasons. One, it would violate isochronicity. “World-sorrow” is already long enough, and even though “on an” passes quickly, the semicolon enforces a pause that adds to the length. Two, it would make the foot contain five-syllables, which while technically allowable, is a special case in sprung rhythm, and nothing in the content of the poem here seems to warrant such an exception. At the same time, “on an” cannot be attached to the next foot (“age-old” in my original attempt), because that would require giving a primary stress to “on” and a secondary stress to “age,” a manifest violation of the way in which the line can actually be read. For this same reason, “on an” cannot be treated as a foot of its own, because, again, this would require giving “on” a primary stress it plainly does not demand (not to mention that it would violate isochronicity pretty badly, since “on an” skates by rather rapidly in this line).

The only option left is to introduce a rest, such that “on an” becomes the tail end of a foot whose primary stress is taken up by a rest beat. This respects that “on” here takes a light stress. It eases the pressure on “world-sorrow” by removing the pause after the semicolon from that already lengthy foot. So it is an attractive option for several reasons; it is not merely a last resort. But it does introduce a problem, for it seems to make the line contain six beats (world; rest; age; an[vil]; wince; sing). In fact, however, this allows me to treat “age-old anvil” as a single dipodic foot, which I think improves the isochronicity of the line: both “age-old” and “anvil” on their own are shorter than “world-sorrow” even without the pause. Introducing a rest in line 6 thus seems to be the correct way of scanning the line.

Line 10. Here my change has no connection to outrides. Rather, I simply realized that “no-man-fathomed” reads nicely as a dipodic foot, while “sheer” can stand on its own as a monosyllable (especially since it is followed by a comma that gives a pause). Indeed, treating “sheer, no-man” as a single foot is quite awkward. This change requires promoting the stress on “no” and demoting the stress on “fath[omed],” but this is a natural change, for “no” begins a new adjective and so can quite properly take a stress. This change thus allows me to treat each of the three descriptions of the “cliffs of fall” equally, which respects the meaning of the line. Honestly, with hindsight I am not sure why I ever scanned it the other way.

Lines 12-13. Previously I scanned “deep. Here! creep, / Wretch” as four consecutive stresses. I was worried about this from the start, since it seemed a bit too much even for such a heightened moment in the poem. Another reason to be worried is that it gives the first foot of line 13 (“Wretch, under a”) a somewhat awkward scanning pattern (/ \ u u). While technically an allowable foot, it is to my ear the least elegant of the ten basic feet sprung rhythm can include. This particular instance of it is made more awkward by the comma between “Wretch” and “under.” Demoting “Wretch” to a secondary stress (and thus making it the completion of the foot begun at the end of the previous line: “creep, / Wretch”) and elevating the first syllable of “under” to a primary stress (yielding the foot: “under a”) resolves both problems. It also, to my ear at least, reads naturally.

I am now fairly confident that I have the proper scansion of this poem, though if history is any guide you can expect another revision tomorrow.

Poem: [No worst, there is none. Pitched past pitch of grief]
Poet: Gerard Manley Hopkins

A few days ago I published a reading of this poem, trying to show how Hopkins deftly modulated the rhythms of the poem both to accommodate his heavy use of alliteration and to enhance the power of his imagery. The only difficulty was: the poem is in sprung rhythm, and at that time I had no real idea how sprung rhythm works. I then read a fine paper by Jeanne Levasseur, from which I came away with a sense of how to scan Hopkins. This post is my attempt at a scansion of “No worst, there is none…” I should say at the outset that I am aware that the precise nature of sprung rhythm is controversial, and I have no doubt that some scholars will take exception to Levasseur’s interpretation. I am not competent to judge whether she is faithful to Hopkins’ vision. What I can say is that, after applying what I learned from her to this poem, I feel I have a better sense of how to read it.

A brief introduction to sprung rhythm, as Levasseur presents it, is in order. She discusses nine principles of sprung rhythm, derived (with modification?) from an earlier paper by Edward Stephenson. This latter I have not read, though I look forward to doing so. In any event, the nine principles are:

  1. Sprung rhythm is always a falling rhythm, meaning that each metrical foot begins with a stressed syllable.
  2. The feet in sprung rhythm are of variable length. There are four basic feet: the (stressed) monosyllable, the trochee, the dactyl, and the first paeon (a stressed syllable followed by three unstressed syllables). The trochee and dactyl are carried over from standard metrical theory.
  3. Each foot is approximately isochronous (i.e. of equal duration), regardless of the number of syllables in the foot. This effect is to be achieved by stressing (which can naturally lengthen a syllable), as well as by pauses.
  4. Sprung rhythm makes heavy use of dipodic rhythm. Dipodic rhythm emerges when a foot contains two accented syllables, one of which contains a primary stress, the other a secondary stress.
  5. Sprung rhythm makes use of clashing accents, in which two or more stressed syllables occur in a row (i.e. with no intervening unstressed syllables).
  6. Sprung rhythm makes occasional use of rove-over lines, in which the last metrical foot in a line spills over onto the next line. (This will occur wherever a line begins with unstressed syllables.)
  7. Certain syllables, called outrides, may be treated as extrametrical and are not counted in the scansion. (I confess I am not sure when a syllable counts as an outride, though I make use of them in one place in the scansion below.)
  8. Alliteration, assonance, and rhyme provide clues to the proper emphasis of the line.
  9. In some cases, rests may substitute for the primary stress of a foot.

In scanning “No worst, there is none…”, I will make use of the following symbols:

/ = primary stress
\ = secondary stress in a dipodic foot
u = unstressed syllable
o = outriding syllable

With these preliminaries out of the way, here is my attempt at scanning the poem:


With the exception of the ninth, all Stephenson/Levasseur’s principles can be seen at work in this poem. Some examples:

  • Variable length: here we can find monosyllables (grief; line 1), trochees (pitch of; line 1), dactyls (comforter; line 3), and first paeons (mother of us; line 4).
  • Dipodic rhythm: at least one dipodic foot can be found in each line of the poem, with the exception of lines 3, 4, 8, and 12. The first foot of the poem (No worst) is dipodic.
  • Clashing accents: there are three instances of clashing accents in the poem, in line 3 (where, where), line 8 (fell: force), and, most extravagantly, lines 12-13 (deep. Here! creep, / Wretch,). Note that in each case punctuation enforces a pause between them.
  • Rove-over lines: found in lines 4-5 (-lief / My), lines 6-7 (sing — / Then), lines 7-8 (ling- / ering), etc. Interestingly, more than once a dipodic foot roves over, as in lines 5-6 (chief / Woe) and lines 11-12 (small / Durance).
  • Outrides: I could not find any way to scan line 6 that (a) contained only five primary stresses and (b) preserved isochronicity without treating “on an” as outrides. I am not sure this is correct, but I gain confidence when I consider that they follow a semicolon. That forces a pause that allows me to treat “on an” as a brief interlude before the next foot begins with “age.”

Ultimately, I think my main rhythmic insights from my previous post survive this exercise, though they need to be expressed differently. To give just one instance, in that post I scanned line 11 as two trochees (durance deal with) followed by three iambs (that steep or deep. Here! creep). That scansion fails on two fronts. First, it massively violates isochronicity, because it forces me to tree “small” in line 10 as a monosyllable. In fact, however, the rhythm of the poem forces me to proceed directly from “small” to “Durance” without a pause, leaving “small” (if treated as a monosyllable) markedly shorter in duration than any other foot in the poem. Moreover, such a scansion forces me to understress “Here!”, despite contextual clues that it takes a heavy stress (the exclamation point being the most obvious such clue).

Numerous other errors in how I was reading the poem found correction in this exercise, not least because scanning it as iambic pentameter forces it into a rising rhythm. Thus, for instance, I had earlier treated “No worst” as an iamb almost by default, when in fact it sounds much more natural as a dipodic trochee. In general, I can say without question that, having scanned the poem in this way, I have a much better feel for how it must be read.

Poem: Pindaric 3
Poet: Geoffrey Hill
Link: Google Books

Alexander Pope, in his “Essay on Criticism,” frequently mimicked in meter what he described in words. His dazzling technical proficiency in doing so is one of the poem’s greatest attractions. Here is my favorite instance:

A needless alexandrine ends the song
That like a wounded snake, drags its slow length along.

An alexandrine is a line of iambic hexameter. Pope is here satirizing poets who used forms that involved concluding a stanza otherwise entirely in iambic pentameter with a single alexandrine. The trouble was that the alexandrine threatened to slow down the flow of the poem too much, which Pope illustrates with a brilliant alexandrine of his own. The cluster of stresses in the middle—wounded snake, drags its slow length—brings the line nearly to a halt, as if it itself were the wounded snake described. (Note for fellow pedants: in scanning the line, “drags its slow length” would most likely be read as a trochaic substitution—drags its—followed by a normal iamb—slow length—but this just illustrates how impoverished scansion is when it comes to capturing the full rhythm of a line.)

Geoffrey Hill, in the third poem in his sequence of “Pindarics,” draws from this Papish well. Here are the poem’s final four lines:

Power’s not every place that virtue is,
and anarchy by files deploys to order
as if through modes of conduct or of weight:
dactyls advancing against a contrived rest.

It is the last line I want to focus on. Here is how it scans:

Dactyls advancing against a contrived rest.

The rest of the poem is in blank verse (though with Hill’s usual interspersing of the occasional clipped line), so most lines have five beats and are in the vicinity of ten syllables. This line, however, opens in dactylic meter. Because dactyls (stress, unstress, unstress) have three syllables, while iambs (unstress, stress) have only two, this creates a tension between the two measures of line length: is this line going to be a five-beat line (thus stretching out to an unwieldy fifteen syllables), or will it stick to ten syllables or thereabouts (at the cost of falling short of the full five beats)?

Neither is an ideal solution. Hill solves the problem with, as the line tells us, a “contrived rest.” After three perfect dactyls, the “-trived” in “contrived” should be the start of the fourth. But instead of continuing on, Hill grinds to a halt on the stressed “rest,” thus bringing the line up to five beats (in only eleven syllables). This rest is doubly contrived precisely because it follows the word ‘contrived,’ a disyllabic word with the stress on the second syllable. While the “rules” of meter in some circumstances permit two stressed syllables to appear back-to-back (as in the mid-line trochaic substitution in Pope’s alexandrine), they generally forbid it when the first stress falls on the second syllable of a disyllabic word. Read Hill’s line aloud and you will hear why: the line gets caught up there. “Contrived” really wants to be followed by an unstressed syllable. Hill denies it this satisfaction.

Hill’s line thus consists in a line of dactyls advancing against a contrived rest.

At this point, having described the metrical perfection of the line, I would like to go on to say something insightful about how it enhances the meaning of the poem as a whole. Unfortunately, I find the poem as a whole basically incomprehensible right now. So I shall have to stop here.

Poem: Improvisations for Jimi Hendrix
Poet: Geoffrey Hill

While I have struggled with many of the poems in Hill’s Without Title, they have all offered me at least reason to want to delve into them further, to unpack their gnarled thickets of words. “Improvisations for Jimi Hendrix” is the first poem I do not care to pursue further. Even without a clear grasp of its sense, some of the lines I find inherently uninviting, irredeemable. For one:

Enlarge the lionized
apparatus of fucking.

Here one may suspect I am a prude when it comes to poetry, and I can’t exactly deny the charge. It is not that I have any general objection to fucking in poetry, but I do find that I have specific objections to nearly every instance, and this is not the rare exception. It does not help that the next line is “Wacko falsetto of stuck pig.”

I do not much like being negative, so I’ll run through only one further line before concluding:

There is no good ending admits fade-out.

I find this line frustrating on two counts. First, semantically, it is so condensed that it will not resolve to any meaning. It is ambiguous between meaning something either “no fade-out can be a good ending” and “when there is no good ending, a fade-out is admitted.” (The poem opens with a quotation of lyrics from “The Wind Cries Mary,” which I presume ends with a fadeout. I could not find a studio version of the song online to check.) I do not see how the poem gains from being undecidable between these opposites.

Perhaps it is unfair to call the sense undecidable. The poem is generally appreciative, so it is likely that the latter is meant. But this still leaves the second issue, which is that the line is hideously unmusical. It starts well enough, but “admits fade-out” allows for three possible readings (bold indicates stress placement):

admits fade-out
admits fade-out
admits fade-out

The first is how it would be read in normal English, but the clog of stresses just sounds… bad. The second reading perverts the normal stressing of ‘admits’ in a way that preserves a perfect trochaic meter for the line, but it is too unnatural to be be believed, and forces the line to be read in a sing-song way that overemphasizes the meter—only that will force the stress into line. The last reading is perhaps best: in other contexts both ‘fade’ and ‘out’ might take a stress, so it less of a deformation to switch the stress around. Moreover, because on this reading the line switches to iambs for the final two feet (after three trochaic feet to begin), it is not so unnatural as the second reading. So it is almost tolerable, but still awkward.

Perhaps with time I will come to appreciate this poem, but for now I can only see it as a misstep.