IF THE TYPE IS TOO SMALL,
PRESS CTRL-+ TO ENLARGE
A Look at
I learned about meter and scansion in the early 1980's, primarily from the books of Judson Jerome. (Actually, I first learned about them in grade school, but promptly forgot.) The method that Jerome taught is one that I have seen mentioned in many books and articles, and I came to accept it as the standard method of scansion, and I still believe that it is the standard method. I didn't consider scansion to be an area of controversy until I started this site and became acquainted with various formalist poets.
Jerome, who made poetry his life-long study and career, estimated that 35%-40% or more of the iambic poetry in the canon of great literature is composed of variant feet, primarily trochees (DUM da), anapests (da da DUM), pyrrhics (da da), and spondees (DUM DUM). (For those who don't know, around 90% of all metered English poetry is written in iambic meters; an iamb is "da DUM".) However, I quickly learned that many of today's formalist poets consider anything more than about 20% of variant feet to be excessive. These issues became the source of many arguments on Eratosphere, the poetry bulletin board at the Able Muse site, which I used to frequent. It quickly became apparent that not everyone was scanning the same way.
The standard method of scanning is really quite simple. Words that receive a strong or moderate stress when spoken are stressed in scansion, whereas words that receive no stress or a slight stress are not stressed in scansion. Thus, this line by Browning, the opening line of "My Last Duchess", would be scanned as follows:
That's my last Duchess painted on the wall,
THAT'S my / LAST DUCH / ess PAINT / ed on / the WALL
trochee / spondee / iamb / pyrrhic / iamb
(Note: There are other ways of representing stress than using capital letters, such as using bolded text or accent symbols, but I have chosen to use capital letters because they are more visually compelling.)
That line has 60% variant feet (other lines in the poem are more regular). But here is where the controversy begins. In the fourth foot, the word "on" can take what Jerome called a "theoretical stress" – that's because it can be given a slight stress without sounding awkward, and some readers may indeed read it that way. However, elevating a slightly stressed syllable to a stressed position suggests to anyone who is reading the scansion that the syllable receives more stress than it actually does.
That line contains yet another foot which is a source of controversy, the second foot. The word "last" can be spoken with slightly less stress than "duch" and therefore, some would say, should be represented as an unstressed syllable (the idea being that only one syllable in a foot should receive a stress). Yet, to me, "last" receives far more stress than any of the unstressed syllables in the line and should therefore be scanned as a stressed syllable.
With those two feet "smoothed out" to conform with the meter, the line then has only one foot which is variant, for a total of 20% variant feet. Here is how that line would scan:
THAT'S my / last DUCH / ess PAINT / ed ON / the WALL
To me, that is an artificial scansion that doesn't reflect the reality of the line. To actually read the line like that would ruin it. Removing the stress from "last" removes the impact of those opening words, and putting a stress on "on" sounds amateurish and unnatural. The resulting line lacks drama and sounds monotonous. In fact, I doubt that any person could actually read it that way without making a concerted effort to do so. So if the line can't reasonably be read that way, what is the point of scanning it that way? More specifically, what exactly are we using scansion for?
Scansion has at least two uses: We use it to study the poetry of other poets, and we also mentally scan our own poems as we write them. Perhaps as a reaction to a century of formless free verse, I believe that some poets are over-reaching in their effort to re-establish the metrical tradition. In such an atmosphere, variant feet would certainly be suspect, and a conservative poet would seek to minimize them. But they really can't be minimized. English being what it is, trochees, pyrrhics and spondees (less so anapests) can't be eliminated. If they can't be eliminated, then at least a poet can pretend to eliminate them – and that is my best guess as to the mindset that has resulted in this type of "meter-based" scansion, as I will call it from now on (i.e., scansion which reflects the base meter as much as possible).
The shortcoming of meter-based scansion is that it doesn't reveal much about the poem being scanned. That's because it does little more than reveal the underlying meter, which presumably the person doing the scanning already knows. If one's purpose is to determine how closely a poem adheres to the meter, then meter-based scansion is the best method; but most people scan a poem to find out how and why it works or doesn't work. It is by contrasting the stresses of a poem with the base meter that those things are revealed, and the actual stresses of a poem can only be seen in the traditional scansion method.
Here is the first stanza of William Butler Yeats's "The Second Coming":
Turning and turning in the widening gyre
TURN ing / and TURN / ing in / the WID / en ing GYRE
Out of 40 feet, I count 18 variant feet, or 45% – a little higher than the average. You may quibble with some of them. "Mere", which starts the fourth line, is sandwiched between two stressed syllables, and thus there is a tendency to take some stress off it (or to pause before it). In that same line, "PON" can be spoken with more emphasis, thereby giving less emphasis to "WORLD". In the sixth line, that stretch of four unstressed syllables is the most that you will ever find, consecutively, in metered poetry. You can elevate "mon" to a stressed position, but that would change the pronunciation of "ceremony", which takes a heavy stress on the first syllable. Likewise, you can elevate "cence" to a stressed syllable, but that would change the natural pronunciation of "innocence". "While", in the seventh line, can take a little emphasis.
Some formalist poets would look at the above scansion and consider it heretical. They would promptly smooth out about 50% of the variant feet to conform with the meter (including all the syllables just mentioned), but then they would have a scansion that would reveal little about the poem. Even more absurdly, some of them might even recite the poem with stresses on those syllables that take only a theoretical stress, resulting in a sing-song rendition. Yeats uses several techniques to provide emphasis in the poem. One is the spondee, and another is the use of many unstressed syllables before a stressed syllable, thereby giving the stressed syllable extra weight. The sixth line of the poem is so powerful precisely for that reason. It ends in a bang with the word "drowned", but that word would have less force if it were not preceded by three unstressed syllables. As seen in the scansion above, that line of iambic pentameter has only three stresses, and each stress is like a small explosion. In contrast, here is the "meter-based" scansion of that line:
the CER / e MON / y of IN / no CENSE / is DROWNED
The impact is gone. You can see that the line has an anapest for the third foot, but that's about it.
One of the ingenious qualities of this poem is that it has a pattern of pyrrhics and iambs (da da da DUM) overlying the basic iambic meter, yet that overlying pattern can't be seen in meter-based scansion. But when the scansion reflects the way the poem is actually spoken, the overlying pattern can be seen. That allows you to compare the overlying pattern with the base meter and thereby gain insight into why the poem works.
Many unstressed syllables in a row can also contribute to a sense of melancholy, and that can be seen in the final line of the stanza. Whereas most of the other lines end with a stress, that line ends with unstressed syllables which give the line a falling rhythm indicative of futility. However, if you scan the line according to the meter, the falling rhythm can't be seen:
are FULL / of PAS / sion ATE / in TENS / i TY
What does that scansion tell us about the line? It tells us absolutely nothing, except that the line is iambic pentameter (which we already know). Indeed, it misleads us into believing that the line has a rising rhythm, which it doesn't. And as with other lines scanned in this way, it doesn't sound natural when spoken aloud. In fact, one of the tests of a good scansion is how natural it sounds when spoken.
Another technique Yeats uses is to occasionally throw in a metrically regular line amidst the more chaotic lines; yet if all the lines are scanned as regular, how can we tell which is which? The point is, a scansion which only shows an imaginary metrical perfection is of no value.
(Note: A reader wrote saying that "widening" in the first line should be pronounced as two syllables – WIDE/ning – in order to conform to the meter and keep the line to ten syllables. However, the sixth line also has eleven syllables unless "ceremony" is pronounced in three syllables instead of four, something I would never do. I think it's fair to say that Yeats wasn't so rigid that he wouldn't slip in an extra syllable once in a while, rather than contort his lines or use old-fashioned ellipses ("wid'ning"). However, the "e" in "widening" is so easily passed over that most readers may indeed speak the word as two syllables. Personally, I like to pronounce the "e" because it makes the word itself wider, which relates directly to its meaning, and in that way contributes to the atmosphere of the poem.)
I created something of a stir on Eratosphere when I posted the first line of one of Shakespeare's most famous sonnets:
Let me not to the marriage of true minds
LET ME / NOT to / the MAR / riage of /
(Although I count the first foot as a spondee, it has a slight trochaic quality, so the first foot could be counted as a trochee.)
That's 80% variant feet. Some posters said that since "of" can take a little more stress than "riage", and since "true" can take a little less stress than "minds", the last two feet are actually two iambs:
-riage OF / true MINDS
But when spoken aloud, that sounds unnatural. One of the posters even claimed that he could pronounce the first two feet as iambs, and I contemplated calling him on the phone to hear him do so! Some argued that the English language was pronounced differently in Shakespeare's time, which is true to some extent; but such a pronunciation as "let ME / not TO" would have sounded as absurd then as it does now. Some said that they pronounced the first four syllables as "let ME / NOT to"; those readers, I believe, were pronouncing "me" with a higher pitch, and confusing the higher pitch with greater stress. Grammatically, the first sentence of the poem has an imperative structure – that is, it is a command, although the speaker is directing the command to himself. Commands that begin with one-syllable words usually take a stress on the first word ("COME to the window"), and I am sure that was true in Shakespeare's time as well.
In all the arguments that I heard on Eratosphere, no one recognized that it is the variants in that line that make it so powerful. One of the techniques of good poets is to open their poems with a metrical bang. Here, Shakespeare does that in several ways: first, by opening the line with a command, which puts more force on the line's opening syllable; second, by clustering stressed syllables together for emphasis ("LET ME NOT ..."); and third, by ending the line with a pyrrhic/spondee combination – da da / DUM DUM. That pyrrhic/spondee combination is so common that it has been given its own name, the ionic. Thus, Shakespeare has provided emphasis both at the beginning and the end of the line, resulting in one of the grandest, most dramatic opening lines in English poetry.
This fear of variants isn't limited to amateur poets on Eratosphere. Tim Steele, a well-known poet and author, scans the fifth line of Robert Browning's "Transcendentalism" like this:
true THOUGHTS / good THOUGHTS / thoughts FIT / to TREAS / ure UP
But I doubt that's the way most people would say it. The emphasis in that line is on the kind of thoughts being mentioned, and thus "true" and "good" would take more stress than either of the first two "thoughts". I read that line like this:
TRUE thoughts / GOOD thoughts, / thoughts FIT / to TREAS / ure UP
It's not my intention to cast aspersions at Mr. Steele, whose scansion method I will discuss shortly. Perhaps he really does read that line as five iambs. There are formalist poets who would break into a cold sweat at the thought of two trochees in a row, and who would alter their pronunciation to suit their prejudices rather than speak it naturally. Only Mr. Steele knows if he is one of them. Such an approach doesn't serve our art.
The third foot of my scansion bears some examination. I pronounce "thoughts" in that foot with medium stress, approximately half way between the highest and lowest stress in the line. I could just as well have written it as a spondee – "THOUGHTS FIT" – but I chose to make it an iamb to conform with the meter. In this I agree with the meter-based scanners: when there is significant doubt, the base meter should prevail.
Here are the lines leading up to that line:
Stop playing, poet! may a brother
There is a number-based scansion method in use by some poets which is similar to the meter-based scansion that I have been discussing. Tim Steele is the primary proponent of this method, and he discusses it at length in his book All the Fun's in How You Say a Thing. It is a four-tiered system of stresses, as follows:
1 = weak
Using that first line of "My Last Duchess" as an example, here is how that line would be scanned:
Steele divides stresses into "speech stresses" and "metrical stresses". The top row of the scansion is a four-tiered treatment giving the "speech stresses" (i.e., how the line is actually spoken), and the second row is a two-tiered treatment giving the "metrical stresses" (i.e., the meter as extrapolated from the four-tiered treatment). The problem is that you can't take a four-tiered system and translate it into a two-tiered system without distortions occurring. If Steele were careful to always include the numbers with the exes and slashes, I would have less objection to this method, but he frequently drops the numbers and gives the exes and slashes by themselves as the correct scansion of a line.
As in meter-based scansion, syllables are demoted or promoted, wherever possible, to conform with the meter. Thus, "last" is demoted to an unstressed syllable because the other syllable in that foot takes a bit more stress; and "on" is promoted to a stressed syllable because the other syllable in that foot takes a bit less stress (resulting, illogically, in "on" receiving a stress whereas "last" does not, though "last" is clearly spoken with more emphasis). For the people who use this method, the all-important question is what relationship the syllables have within their respective feet. They seem to forget, however, that feet are an artificial subdivision, not an actual reality. Since it isn't possible during the recitation or hearing of a poem to determine which syllables come together to form which feet, it makes little sense to use feet as the determining factor in deciding which syllables to stress in scansion.
The ultimate result of this method is that it eliminates pyrrhics and spondees from virtually everything that is scanned. That is a good thing, according to Steele, his argument being that true pyrrhics and spondees (in which both syllables have an equal stress) are rare. I submit, however, that that is irrelevant. The main purpose of scansion is to reveal the primary stresses of a line (i.e., the main ups and downs, or what Steele calls the "speech stresses"), not the secondary stresses (i.e., the nuances within those ups and downs). To focus on the secondary stresses is like focussing on the garnish instead of the roast beef – the meal is not in the garnish.
In his book, Steele gives various examples of spondees and pyrrhics that, in his view, are really iambs. They aren't the best examples, as I actually agree with some of them, so I am not going to discuss them here. The real issue is whether two substantially stressed syllables should be scanned as a spondee even though their stress levels are slightly different (I'll speak about pyrrhics in a moment). I contend that the fact that they are both heavily stressed is more important than that one is stressed a bit more than the other, and I have already given some good examples of why that is so in my analysis of "The Second Coming".
Although I would agree that true spondees are not that common, I don't feel that way about pyrrhics. I have no problem speaking three unstressed syllables in a row. In his book, Steele gives this example, by Wordsworth:
The second "a" in that line has virtually no meaning, and as such it can't take more than a whisper of stress and still sound natural. Using Steele's system, I tend to speak "sleep and a forget-" as 4 1 1 1 4, or possibly even as 4 2 1 2 4, as if the words described a valley with "a" at the nadir. However, even if I speak those syllables as 41214, I still say that's good enough to call the fourth foot a pyrrhic. This line and the line from "Transcendentalism" (discussed above), along with other examples from his book, make me wonder if Steele has developed the unconscious habit of promoting and demoting syllables to conform with the meter. If so, that's certainly his prerogative, but not everyone does that.
Before I continue, I feel compelled to respond to a statement in Steele's book. He says, "Despite the rarity of true pyrrhics and spondees in English verse, many authorities introduce them almost willy-nilly into accounts of poems in conventional iambic measure." I hope that my explanation here of the value of pyrrhics and spondees dispels the notion that there is anything "willy-nilly" about it.
Alternating stress is the "bread and butter" of English poetry, and a four-tiered system doesn't reflect that reality. As I said above, about 90% of all metered poetry is written in iambic meters, and that means alternating stresses. Even in anapestic verse, the alternating-stress quality of the language is preserved, as the two unstressed syllables, taken together, provide a balance to the stressed syllable. Because of the alternating-stress nature of our language, it is important that we focus on the primary up/down stresses. A four-tiered system is simply overkill to represent a two-tiered linguistic reality.
Let me be a little blunt here, and perhaps a little crude: Alternating stress is like the rhythm of sex: IN/OUT, IN/OUT, IN/OUT. What is important is not whether you are going in and out with more or less force, but whether you are going in and out at all. See?
I also object to Steele's system because it isn't very readable. Here is the line by Browning again, without the exes and slashes:
In order to read that, you must look back and forth between the numbers and the text, while at the same time interpreting the numbers into four levels of stress. That's not easy to do. Here again is the regular scansion:
THAT'S my / LAST DUCH / ess PAINT / ed on / the WALL
To read that, you need only look directly at the words. The stresses, represented by the upper- and lowercase letters, are easy to see. Dividing the stresses into strong and weak is much simpler than dividing them into four gradations. Whatever secondary stresses the line contains will be inserted by the reader naturally. It seems to me that a scansion method which cannot be easily read and recited defeats its own purpose.
That line of poetry, incidentally, is full of drama. It is easy to visualize the speaker, accompanied by his guest, raising his arm in the direction of a great painting on the wall. The dramatic quality of the language is visible in the regular scansion, but it is harder to see in the more complicated number-based method.
One of the problems with Steele's method is that it lacks concision. He not only uses two different notational methods – symbols (exes and slashes) and numbers (1-4) – he sometimes reverts to the traditional method to show the "speech stresses". He does this in his own examination of "The Last Duchess"; here are the lines he gives:
She rode with round the terrace—all and each
Which Claus of Innsbruck cast in bronze for me
The depth and passion of its earnest glance
The dropping of the daylight in the West
Quite clear to such an one, and say, "Just this
Then all smiles stopped together. There she stands
With the exception of the vertical lines that denote the feet, Steele has unwittingly given us the traditional scansion of those lines (and I agree with all the stresses). It is precisely because Mr. Steele's numbers and symbols are not readable that he reverts to the more traditional method. One wonders why he doesn't just use the traditional method all the time, since it reveals so much more about a poem.
I should point out that the traditional method doesn't reveal everything about a poem's meter. Specifically, it doesn't reveal the base meter; but I don't see that as a problem. The base meter is easily and quickly determined by counting the stresses and syllables of several lines. Once the base meter is determined, it is unnecessary to indicate it in the scansion. It is the primary stresses, along with other elements such as caesurae, enjambments and word length, that most need to be revealed.
Talking about those symbols – the exes and slashes – they themselves can be a source of confusion. They don't fully represent the base meter, yet they don't fully represent the speech stresses either. Rather, they are an extrapolation of the stresses resulting from the numerical scansion. In iambic verse, trochees are represented by "/ x" and anapests are represented by "x x /", yet the rest of the time, the exes and slashes represent the base meter. One starts to wonder what the symbols really represent. They appear to be Steele's idealized version of a poem's meter. That is perhaps my greatest objection to this method: the symbols don't represent anything real. The idealized meter they represent is a theoretical extrapolation, and as such they contribute nothing to an analysis of the poem.
In closing, I am not saying that Steele's method doesn't have its uses. I myself use the numbers when trying to explain to others exactly how I enunciate a line. I also think that the numbers could be useful as a teaching aid. But for a true analysis of a poem's meter, regular scansion, with its pyrrhics and spondees, is much more productive. I also worry that Steele's method sends the wrong message to beginning poets: it suggests that variants which cannot be "smoothed out" to conform with the meter are somehow wrong and should be minimized. It also suggests that the right way to read poetry is to emphasize the base meter. Neither of those things is true.
What Is a Good Reader?
Although it is a little outside the scope of this article, I'd like to talk about the responsibilities of the reader. Just as there are good and poor poets, there are good and poor readers of poetry. First and foremost, a reader must concentrate on what he is reading. Beyond that, a good reader is one who has an understanding of poetic devices but who is careful to approach each poem with an open mind. The reader needn't be technically proficient but should at least understand such basic things as: the end of a line does not force an intentional pause (probably the most common error). Yet the inability to see a poet's vision because of one's biases is just as serious a mistake. I include in this the inability to comprehend the metrical sounds of a poem because they don't conform to one's notion of what is metrically acceptable.
I have been told that a good reader should discern the base meter and elevate or demote syllables accordingly, but it is possible to do that only when a poem is extremely regular. If more than about 20% of a poem's feet are variants, how can a reader be expected to know when the base meter should prevail? The responsibility rests with the poet, not the reader. If a poem is well written, the reader will have no choice but to elevate or demote syllables where appropriate (i.e., in accordance with the meaning). If the poet leaves room for the reader to apply his or her natural diction, the reader can't be blamed for doing so, even if the result is awkward. Take, for example, the second line of "The Second Coming":
The falcon cannot hear the falconer;
the FAL / con CAN / not HEAR / the FAL / con er
If the reader pronounces "cannot" as "CAN not", the base meter is preserved; but if the reader pronounces "cannot" as "can NOT", resulting in two variant feet, so be it. In my own speech, I usually pronounce "cannot" with the stress on the second syllable, and that tendency doesn't change just because I am reading a poem. Now, let's look at the third line of that stanza:
Things fall apart; the center cannot hold
THINGS FALL / a PART; / the CEN / ter CAN / not HOLD
That line also contains the word "cannot", yet the line is written in such a way that the reader is compelled to apply stress to the first syllable. "can / NOT" is not an option here (well, maybe it is, but most readers will yield to the poet's intentions).
(Shortly before my mother died at the age of 90, I told her that my favorite dessert was "PEE-can PIE", to which she corrected me by saying the correct pronunciation was "pi-KAHN PIE". This is a good example of how the pronunciation of words can fall prey to their context, since the former pronunciation suits my sense of rhythm better than the latter pronunciation.)
In summary, the reader's responsibility, though it exists, is minimal; the poet is responsible for the success of the poem.
The "meter-based" and "number-based" scansion methods examined here are good examples of over-complicating a simple thing. Regular scansion, with its pyrrhics and spondees, is a time-tested technique, and I see little reason to change it. Personally, I think these new methods represent a narrow mindset. Poets who use them tend to write very regular poetry, with fewer grand and dramatic metrical gestures than I see in the classics. There is no harm in writing such poetry, although you run the risk of sounding monotonous, strident or humorous. For the most part, it is the metrical variations that make a poem sonically interesting, although caesurae (pauses in the line), enjambments (line endings that don't coincide with a natural pause), short and long vowels, and other factors all play a part. The base meter of a poem provides the unifying structure as well as the rhythmic impetus which carries the reader forward, and the variants add impact and emotional flavor. Most importantly, the variants work against the base meter to provide tension, which heightens the reader's interest.
I recommend that you scan poetry as you would speak it, and then examine the scansion to see how it varies from the base meter and how those variations help or hinder the sound and the meaning. To learn about variants and how they are commonly used, scan poetry which you know to be great. I specifically suggest studying poetry from the 19th century and later, as such poetry will have a more modern diction. Of course, one should not overlook the great Bard, or other great poets of earlier eras such as Donne and Jonson, but keep in mind that pronunciation in Shakespeare's day was sufficiently different to change the scansion of some lines – for example, the common suffix "-tion" was often pronounced as two syllables. If you find yourself scanning poetry which doesn't seem to fall into any meter, it may be accentual verse, such as Hopkins wrote, or free verse. Some poets, such as W.H. Auden, wrote poetry with more than the usual number of variants. Generally speaking, the more variants a poem has, the more rhythmically charged it is, although some "quieter" emotions are often expressed in regular meters. I should point out, however, that if the variants are excessive, the meter can be lost and the poem may lose cohesion. But I implore you not to fall into the trap of believing that good poetry must be metrically regular. Nothing could be further from the truth.