I should have used Steele’s definitions. There are different kinds of stress in speech: word, sentence, and contrastive, amongst them. When Jonathan notes that “veritably unimpressive [is] a phrase containing five syllables in a row, all unstressed,” he’s speaking of word stress. And he’s right.
But what I should have said is “you cannot have three consecutive metrically unaccented or accented syllables in English.” And I should have quoted Steele on the meaning of metrical accent:
By metrical accent, I refer to the character of a syllable when it is considered simply according to the offbeat-beat system of scansion. If it receives more emphasis than the other syllable or syllables of the foot in which it appears, it is metrically accented. If it receives less emphasis, it is metrically unaccented.With that finally said, I can say that Jonathan’s “veritably unimpressive” does prove my point in fact. I would scan those two words in a line as: trochee, iamb, trochee, trochee, with metrical accents on the 1st, 4th, 5th, and 7th. If we were to use a four-level system of scansion: 1-weak 2-semi-weak, 3-semi-strong, 4-strong, I’d scan the eight feet as such: 3-1-1-2-3-1-4-2 .
As to "weak positions in the meter should not coincide with stress peaks:” If by this we mean “metrically unaccented positions in the meter should not coincide with sentence stress peaks,” then I’d agree. Kind of. Jonathan thinks that’s elegant; I think it’s slack (I'd change 'should not' to 'cannot') and incomplete (maybe I was looking too hard for what was missing and thus thought it overly technical.) Again, using the 4-level system, it’s possible to scan a rising 1-2-3-4 ("With how sad steps") as two iambic feet. Yes, it’s true that weak positions do not coincide with peaks. But it fails to recognize the slope.