The fundamental formal distinction between poetry and all other forms of literary art is this: poems are made up of lines. But how long is a line? Lines of metrical verse are subject to measurement just as surely as if they were made of cloth and both poet and reader has a yardstick. In the case of cloth, you measure physical distance by counting in yards. What do you use to measure the length of a line of poetry? The simplest kinds of meters are measured in syllables. There are only so many in a line.
Such meters occur in the poetry of languages all over the world. Much of the poetry in the Romance languages counts syllables (Halle and Keyser 1980). Hebrew poetry of the Old Testament does as well (Halle 1997). So, too, do the Japanese verse forms known as tanka and haiku (Halle 1970). In the five-line tanka, lines are 5-7-5-7-7 syllables long. In the three-line haiku, the lines are 5-7-5 syllables long. As shown by the poems in (1) and (2), syllables with double vowels (i.e., long syllables) are counted as two units; all other syllables count as one.
Haru tateba | 5 | When spring comes |
Kiyuru koori no | 7 | the ice melts away |
Nokori naku | 5 | without trace |
Kimi ga kokoro mo | 7 | your heart |
Ware ni tokenamu | 7 | melts into me |
(Kokinshuu) |
Kaki kueba | 5 | Eating persimmons |
Kanega narunari | 7 | the bell rings |
Hooryuu-ji | 5 | Hooryuuji |
(Masaoka) |
Japanese poets and their readers both know what a syllable is. They also know the difference between a long syllable and a short one. This shared knowledge is what Japanese meter depends upon. Segmenting a word into syllables requires a great deal of sophisticated knowledge about such things as the difference between a vowel, a consonant, a liquid, and a glide. No modern computer can reliably segment a stretch of speech into syllables. Yet speakers of all languages do this constantly and unconsciously.
Much more complex linguistic machinery is involved in verse where line length is measured by counting feet rather than syllables. Foot-counting verse is encountered in a wide range of poetic traditions, among them Homer and much of the poetry of classical Greek and Latin antiquity, the Old Norse bards, English poetry from Chaucer to Frost, German poetry from Hans Sachs to Rilke, and Russian poetry from the eighteenth century to the present.
To illustrate the more complex machinery of foot-counting meters, we look at the so-called syllabo-tonic meters of English and, in particular, iambic pentameter. This meter is made up, as its name suggests, of five feet. An iambic foot is made up of two syllables followed by a boundary. There is a specific procedure that divides the line into iambic feet. Consider the opening line of Gray's "Elegy Written in a Country Churchyard":
The | cur | few | tolls | the | knell | of | par | ting | day. |
* | * | * | * | * | * | * | * | * | * |
We represent each syllable in the verse by an asterisk beneath the line. The procedure is: insert right parenthesis from left to right, starting at the left edge so as to group asterisks into pairs:
The | cur | few | tolls | the | knell | of | par | ting | day. |
)* | * | )* | * | )* | * | )* | * | )* | *) |
Although there are six parentheses in this line, there are only five feet because an iambic foot is defined as a sequence of two syllables followed by a parenthesis. The rule that inserts parentheses is:
Readers familiar with iambic verse know that many lines of iambic pentameter verse are often longer than 10 syllables. These exhibit what the textbooks call "feminine rhymes." The following couplet from Byron's "Don Juan" (Canto 65) is illustrative:
The line-final rhyming pair is show it: know it. Each member of the pair has a so-called extrametrical syllable -- namely, it. Rule (5) automatically accounts for the possibility of such syllables in iambic verse:
Yet | he | was | jea | lous, | though | he | did | not | show | it, |
)* | * | )* | * | )* | * | )* | * | )* | * | )* |
For | jea | lou | sy | dis | likes | the | world | to | know | it. |
)* | * | )* | * | )* | * | )* | * | )* | * | )* |
As before, right parentheses are inserted, beginning at the left. Notice, however, that a right parenthesis cannot be inserted to the right of the final * in (7) because (5) requires that two syllables are skipped before a right parenthesis can be inserted. Here only a single * follows the last parenthesis. Therefore, no parenthesis is inserted line finally and the line is correctly scanned as containing only five iambic feet.
Just as readers familiar with poetry written in iambic pentameter recognize that some lines are longer than 10 syllables, they also know certain lines are not possible iambic pentameter lines, even though they may be composed of 10 syllables. For example, a line like (8) is not a possible iambic pentameter line.
This causes a problem for the account given so far because (5) scans the line without difficulty:
On | cut | ting | Lu | cre | tia | Bor | gia's | bright | hair | |
)* | * | )* | * | )* | * | )* | * | )* | * ) |
(5) must be modified in such a way that it will continue to scan lines like those in (4) and (7) while ruling out lines like (9).
It is well known that in foot-based meters not only does line length play a role but so does the placement of certain marked syllables. In the English iambic meters, these are the syllables that bear the main stress of the word. We mark such metrically important syllables by inserting a bracket before or after them. The marked syllables in stress-based verse are called stress maxima. A stress maximum is the main stressed syllable in a polysyllabic word. In the history of English, poets made use of slightly different definitions, extending the definition in some cases and restricting it in others. For present purposes, the stress maximum is defined as:
The stressed syllables in the following words are instances of stress maxima:
Lucrétia | meándering |
autobiográphic | pellúcid |
Rule (12) accounts for the placement of such syllables in an iambic pentameter line.
Let us see how rules (5) and (12) construct feet within a line like (13).
The square brackets called for by stress maxima are inserted first. They are no different from parentheses, but we use them to help the reader keep track of which rule is responsible for a given boundary:
The | cúr | few | tolls | the | knell | of | pár | ting | day. |
* | * ] | * | * | * | * | * | * ] | * | * |
Next come the parentheses inserted by (5):
The | cúr | few | tolls | the | knell | of | pár | ting | day. |
* | * ]) | * | *) | * | *) | * | *]) | * | *) |
The line is correctly divided into five feet and, significantly, where right parentheses and right brackets occur in the line, they coincide. Now consider how (5) and (12) assign boundaries to the unmetrical (8):
On | cut | ting | Lu | cre | tia | Bor | gia's | bright | hair |
* | *]) | * | *) | *]) | * | *]) | * | *) | * |
Unlike the metrical lines discussed thus far, this line contains two feet that end on consecutive syllables, namely:
-ting | Lu | cre | tia |
* | *) | *] | * |
This is not a well-formed configuration. We exclude it by means of prohibition (18):
Rule (18) applies after the insertion of parentheses by (5) and (12). All scansions that conform to it are well-formed. Scansions that fail to do so are not.
The theory of meter proposed contains two rules, one output constraint, and a definition of stress maximum. The rules insert right foot boundaries, either to the right of stress maxima or else iteratively from left to right. The constraint rules out scansions with unary feet. The grammar is summarized in (19):
Just as two speakers of a language share a body of linguistic knowledge called a grammar, which enables them to speak to one another, so, too, do poets and their readers share a body of knowledge that enables the one to write metrically and the other to scan what has been put into meter. The rules in (19) are an attempt at illustrating what this body of shared metrical knowledge looks like.
Up to now, nothing has been said about the machinery used to scan iambic pentameter lines. That machinery is identical to that needed to account for the way ordinary speakers assign stress to the words of English (see STRESS, LINGUISTIC), and is part of the natural endowment of human beings that enables them to speak a language -- what linguists have come to call Universal Grammar (UG). In other words, poets who write metrically do so using the same theoretical apparatus provided by UG that speakers use to assign stress to the words of their language. This convergence of the machinery of meter with the machinery of stress assignment is the essence of PROSODY.
Halle, M. (1970). What is meter in poetry? In R. Jakobson and S. Hattori, Eds., Sciences of Language: The Journal of the Tokyo Institute for Advanced Studies in Languages, vol. 2. Japan: Tokyo Institute for Advanced Studies of Language, pp. 124-138.
Halle, M. (1997). Metrical verse in the Psalms. In V. van der Meij, Ed., India and Beyond: Aspects of Literature, Meaning, Ritual and Thought. Essays in Honour of Frits Staal. London: Kegan Paul International, in association with the International Institute for Asian Studies, Leiden, pp. 207-225.
Halle, M., and S. J. Keyser. (1980). Metrica. Enciclopedia IX. Torino, Italy: Einaudi.
Halle, M., and S. J. Keyser. (1998). Robert Frost's loose and strict iambics. In E. Iwamoto, M. Muraki, M. Tokunaga, and N. Hasegawa, Eds., Festschrift in Honor of Kazuko Inoue. Tokyo, Japan: Kanda University of International Studies.