Speech is naturally rhythmic. (I'm not talking about prose, but about oral speech.) That is, it tends to fall into intonational patterns, most simply: UUUUP dooooown, UP down... What is the status of metrical rules, or codifications of these intonational patterns? For example, the French Alexandrine (12 syllables line) breaks up into two symmetrical halves, in a call-and-response pattern. But the same intonational pattern would be there if you occasionally had half-lines of 5 or 7 or 8 syllables instead of the required 6. Conversely, you could have metrically correct verses of 12 syllables in which the intonational patterning was relatively weak. What doesn't work is to destroy both the metrical rule and the intonational pattern at the same time. In short lined, heavily enjambed free verse, the line is a spurious unit, in the sense that it obeys neither a metrical nor an intonational imperative:

"The alphabet of the trees // is fading in the song of the leaves"

(I forget exactly where WCW breaks these lines.)

The line could be either a unit of performance, or a purely visual effect. When I read this poem aloud, I observe the line-breaks, introducing a pause but not interrupting the intonational pattern.

As I said to my friend [up] // because I am always talking [level]
John, I sd [up]// which was not his name [level]
The darkness surrounds us [down]// what can we do against it [up]
Or else should we, and why not? [up]// buy a godamn big car, drive [down]
He sd for Christ's sake [up]// look out where you're going [exclamatory]

This is greatly simplified, of course. Your assigment: to translate the Creeley poem above into perfect, rhymed French Alexandrines, then compare the metrical and intonational patterns of the translation and the original text.