Someone asked me in the comments whether I really believe everyone should scan a line of verse in the same way. The short answer is ... maybe.

In other languages I'm familiar with scansion is not controversial. In other words, you can actually assign the accents or divide syllables into long or short in a rather mechanical way. The fact that English scansion is controversial at all shows that metrical theorists have screwed us up. I don't believe scansion should be all that controversial.

Everyone knows what a particular metrical pattern is. Everyone also knows how to pronounce words and phrases in his or her native language. Scansion is the attempt to reconcile the everyday pronunciation of a phrase with some underlying metrical pattern. The more difficult this is, the more metrical "tension" there is in the line. Lines with little or no metrical tension are comparatively rare--

Of hand, of foot, of lip, of eye, of brow.

They aren't really a metrical "norm," then, or even an "ideal"--since a poem composed of such lines would not be ideal at all. In a way, they deviate from the norm simply by being too regular, and thus create a different sort of tension. The normal line has an average degree of approximation to the metrical grid. At the other extreme are lines that raise tension to the maximum degree, mostly by placing strong syllables in weak positions. Logically, the most disagreement over scansion will occur when metrical tension is at its highest. A line like "Bright star, would I were steadfast as thou art!" The contrasting pronouns make you want to say "Bright star, would I were steadfast as thou art." That works fine, because "I" is in a strong position--but not so smooth because "thou" shouldst be in a weak position. The first seven syllables of the line are also quite heavy. Not all of them are equally stressed, but none is light, and then the 8th, lighter syllable occupies a strong position! Then two more heavy syllables to conclude! (Weak syllables in heavy positions don't really cause as much of a sense of strain or tension.)

So what would it mean to say there is a difference of scansion in this line between different readers? A difference in pronunciation, or a difference in the interpretation or representation of phonological facts? I would argue that these differences wouldn't really amount to a lot. A verbal performance of the line that tried to make it sound like "Of hand, of foot..." wouldn't be very good, but you'd still feel a lot of "tension" because it would sound so stupid. In other words, we have to acknowledge, no matter what, that the line is heavy with monosyllables and metrically tense: once again Keats starts a sonnet with a metrically problematic and expressive line. The theoretical method by which you want to reconcile (or explain the difference between) the pattern with the actual pronunciation of the line doesn't really affect the "interpretation" of the line. For example, I would probably use the generative metrics of Kiparsky and the theory of English accent of Liberman & Prince. You might try an old style foot-substitution method. We'd both find a great deal of metrical tension.