Rather than go about this the usual way, I think I will discuss Morrison's The Devious Way by looking at the requirements for a verse novel or, indeed, narrative verse generally.
The reason we talk in prose rather than verse is that prose is extremely tolerant of approximation and looseness. When you read a paragraph of prose, most of the language simply 'goes by'; it is not put front and center but pressed into the surface of a stream of thought that is always going to, or coming from, somewhere else. Verse, however, is different in precisely this way, that it is not tolerant of approximation and looseness. In verse, language is never allowed simply to 'go by', because it is always put out front. Stream of thought is carried not by the purely functional way in which you say this only in order to get to that; it is carried out by turn and return, i.e., verse in the quite literal sense of a turning. This function of turn and return could be carried by rhyme, or by metrical pattern, or by repetition, or by alliteration, or any number of other things, but precisely the thing that differentiates the coherence of verse from the coherence of prose is that in verse poetry verses must call out or respond to other verses in some way. This, incidentally, is why free verse, which is easy to write, is so extraordinarily difficult to write well: it requires no discipline to write it, because turn and return, call and response, are not formalized at all, but free verse, in order to be verse and not bad prose, must still have the turn and return of verses, the call and response, in some way, and given what free verse is, your means of doing this must be far more subtle than it is in other forms of verse.
When looked at this way, however, the notion of narrative verse presents something of a problem. Prose is very narrative-friendly, but good narrative verse is quite difficult because narrative verse poses problems narrative prose does not; we might say that narrative is itself more prose-friendly than verse friendly. Decent epic verse is a greater achievement than brilliant lyric verse because there are two additional temptations to avoid when writing the former. The first temptation arises from the fact that the most natural thing to do with verse is to present snapshots: things frozen in time. If you think about most great, striking, or memorable verse, this is precisely what you will tend to remember: some very vivid detail, some remarkable description, a perfect allusion, an abstract or figurative characterization of action, some movement or process somehow caught in a single moment of stillness. In verse everything said is front and center at some point, and this makes it easy to bring narrative to standstill. To the extent you do so, however, your narrative powers have failed: narrative must move. The other temptation, of course, is to lose a grip on your verse as verse, so that the only thing that distinguishes it from prose is that you scatter it on the page less efficiently. In bad narrative verse the poet is repeatedly defeated by both temptations: the narrative thought stops and starts spasmodically, the 'verse' is really just absurdly out-of-shape prose -- bombastic, maudlin, or flabby.
This, then, is the problem of narrative verse. As we know from Homer, Virgil, and Milton, though, it is a soluble problem. The narrative in the Iliad or the Aeneid, or Beowulf, or Paradise Lost, moves, and while it sometimes slows down, it does so deliberately without any choking of the engine. Everything moves forward. Even the great narrative poets struggle with the problem, however: hearing or reading the narrative is never as easy in the Iliad as it is in even merely competent prose. The narrative has to be carried, not by the flow of language, but by how verses in their turn and return, their call and response, are woven together. All good narrative verse, then, requires two things: a discernible order contributing to the direction of the narrative and a tightness of language suitable to verse rather than prose.
The Devious Way is very inconsistent in how it meets these criteria and therefore does not succeed as a solution to the problem of narrative verse. It does not help that Morrison is writing a free verse novel, which compounds the problem to an extraordinary degree. Fortunately, he is actually a capable poet, and so recognizes that free verse is consistent with occasional and irregular use of traditional poetic technique, which he sometimes has to good effect. But 'free verse novel' is a label for what is perhaps the most difficult poetic problem known to the human race: how to write a long series of discrete verses carrying a continuous narrative suitable for prose, while not using most of the poetic tools for verse creation yet not falling into prose in the process. Morrison's book is not a total failure to solve this problem; it does many things right. Morrison chose an excellent narrative for the task, a modernized Troilus and Cressida tale. Chaucer had already showed the potential of this kind of narrative for verse, and although Morrison's 'modern' (i.e., non-epic, since most of what we mean by 'modern' in literature is really just 'inconsistent with epic' -- epic deals with things too timeless, or perhaps handled in too timeless a way, to be distinguishable as modern) characters are necessarily smaller and less interesting than Chaucer's epic characters, this is not a problem of any sort, just a difference in the kind of work they are producing. I went into the novel expecting the narrative to be one of the things I would not like about the book, but in fact it is a good narrative for a book of this kind. What is more, Morrison is quite sensible about the narrative; he keeps it both simple and reasonably interesting as a novel. So Morrison had the "narrative suitable for prose that could also be told in verse" part of the problem resolved quite well: it's a novel-ish narrative with verse-ish potential.
It's the rest of the problem that this verse novel fails to solve. Descriptions in particular seem to pose a problem for Morrison. He has some quite good ones, like this one (p. 46):
Stretched in a canvas chair, her lap in shade,
Her legs in sun, Christina leafed a book.
Her thought ran wavering up and down the page,
And what she tried to read was David's thought.
This is a good description of a recognizable phenomenon: Christian is reading a book David has thrust on her, and she cannot follow it both because it is not the kind of book she would usually read -- she is only reading it to find David's thought -- and because she is really more interested in thinking about David. The narrative function of these lines is clear, in four lines it perfectly captures a rather complicated scene, and while it's possible it could be improved, it is still a quite capable attempt to do what a poet should attempt to do in verse and poetic prose alike: say things exactly as they should be said. What is more, it's good as verse, too: each line presents us a reasonable stage for language to play its parts, and the lines are connected by turn and return. "Stretched in a canvas chair" calls and "Christina leafed a book" responds; "her lap in shade" calls and "Her legs in sun" responds; "her lap," "Her legs," and "her thought" gives us a more complicated structure of the same kind; the "Her thought..." line calls and the "...David's thought" line responds. It's true that it's not too far from what you would get in a perfect prose description of the same scene, but that's simply because perfect verse descriptions of scenes like this are going to have many similarities to their counterparts in prose; it's an ordinary scene with simple needs. This is solid free verse, suitable to narrative.
We get recurrences of this kind of description, so it's not an accident, but a sign of real potential. Most of the book is not like this, however; it reads like flabby, overstrained prose. People do not merely yawn; they yawn cavernously. People do not merely get angry; their faces knot with anger. People do not merely answer telephones; they answer telephones with fierce urgency. They do not just grin; they grin so that their grins show like cracked rinds. This is a book that needed a Jack Sprat editor. For instance, we get these two lines (p. 23):
Talk was catharsis. David sighed again,
Still unrelieved; the purge was incomplete.
Besides the unfortunate undertones of constipation (a misfortune only exceeded by a later line's description of someone's restraint by saying that they have 'the continence of a running spigot" (p. 110)), this is twice as many lines as actually necessary. "David sighed again, his catharsis incomplete" would do just as well and would involve less distraction from the narrative as we concentrate on David's purging.
Another example (p. 83):
Sighed as if heaving upward with her breast
Some weight of stone that had a lodgment there.
Do we really need the "as if"? Is there going to be any confusion about whether Christina's aunt actually has a heavy rock in her chest? Metaphor is more natural than simile, especially in verse; you should only similize when you get something out of it. Likewise, Morrison (and this is a recurring fault) can't say the mundane "a heavy stone that had lodged there"; instead of a heavy stone in the chest we get a weight that has a lodgment. There are, to be sure, conceivable poetic situations in which it would be better for something to have a lodgment than to lodge or be lodged, but they are surely quite rare, and it seems unlikely that Christina's aunt's heavy sighs are deserving of the honor.
We also regularly find verse shifting into something that is (1) very like prose but (2) only is distinguished from typical prose by the fact that it would usually be said better in prose. Part of the difficulty, I think, is that Morrison's Troilus and Cressida tale faces a serious problem that Chaucer nicely avoided. An epic Troilus, or Cressida, or Diomede, will naturally speak in verse. A modern-day David, or Christina, or Reuel who speaks in verse is very difficult to imagine. People do not converse epically, ever. They do not even usually converse lyrically. But a novel-ish narrative involving a love triangle and a Pandarus figure (Morrison actually has two, to split the functions of Chaucer's) will inevitably involve conversation. Morrison handles the conversation pretty well on its own, but it is very prose-like conversation (as it would have to be) embedded in descriptions that, if they were prose, would be purple or flowery. Now, I am on record as defending purple and flowery prose; but the contrast between conversations described in ways very similar to what you would actually hear and the conversational environment described in ways that would not be obvious to one actual experiencing them is jarring. And sounding like purple prose or flowery prose is not adequate for making something genuine verse; it's still just prose, however weirdly laid out.
So what can we say about The Devious Way? Verse novels can fail in three possible ways: they can fail both as verse and as novel; they can succeed as verse but fail as novel; or they can fail as verse but succeed as novel. Morrison's work falls into this third category. This is a pretty decent novel (or rather novella, as it would be in prose): plausible characters in a good story, and Morrison's modifications of the standard Troilus narrative, by making it less tragic (it is not a tragedy but a story about the devious ways life travels) make it more readable, and are done well. Ironically, given that Morrison is a poet, it fails as verse. Some of the verse is very good; most of it, however, is inconsistent, and some of it is really just bad prose. Because it succeeds as novel, and because Morrison does have poetic skill, it is probably better than most verse novels. I am very much in favor of the idea of more verse novels; but The Devious Way, I think, shows that we still have a long way to go before we get any that are quite adequate.