Where do we draw the line between what is poetry and what isn't poetry? Or, more specifically, what makes a poem a poem? Ask a poet like Howard Nemerov, and you'll get a beautiful answer, in the form of a poem called "Because You Asked about the Line between Prose and Poetry":
Sparrows were feeding in a freezing drizzle
That while you watched turned into pieces of snow
Riding a gradient invisible
From silver aslant to random, white, and slow.
There came a moment that you couldn't tell.
And then they clearly flew instead of fell.
It's nice (especially since it compliments poetry, making it fly, while prose merely falls), but in the end Nemerov's answer about the nature of poetry is evasive, offering little more than the "I'll know it when I see it" argument that people used to invoke in debates about pornography.
I don't really have a better way to answer the question, except to say that the only real way to answer anything is to quit looking for trans-historical, absolute truths and start rooting around in contexts, in the history of how a question has been answered, and the reasons those old answers made sense at the time. If we do this, I think we can say that something changed in the way we answered the question "how is poetry different from prose" right around the beginning of the nineteenth century, and that, with some small modification, the new answer poets came up with at that time is still with us. This is Romanticism, people, nor are we out of it.
One of my touchstones, when I want to think about pre-Romantic poetry in French or in English is a statement by Roland Barthes in his great, early study Writing Degree Zero. Here, he argues that for poets of the pre-Romantic period, there was a fundamental similarity between good poetry and good prose. “Poetry is always different from prose,” writes Barthes, but in the classical period “this difference is not one of essence, it is one of quantity. It does not, therefore, jeopardize the unity of language, which is an article of classical dogma.” In other words, the classical or Augustan writer doesn’t grant poetry one of the rights that we moderns and postmoderns grant it: the right to follow rules significantly different from those governing discursive prose. It is more or less the same stuff, but with the added use of particular literary devices such as rhyme and meter. Think about it: Alexander Pope's poem Essay on Man is more or less what it sounds like — a discursive, explanatory essay — it just happens to add versification. Even the title indicates a fundamental continuity between prose and poetry. Prose and poetry are more or less up to the same sort of things: explaining, talking, arguing, narrating.
And why did people write poems? Many reasons, of course: but they tended to be the same sorts of reasons one writes prose: utilitarian ones. Often poetry was a courtly game, but all games have some purpose. In this case, it was to show one's wit and cleverness — sort of a peacock's tail thing, in that verse may have seemed useless, but it was a sign of health and desirability, therefore useful in securing a mate or some sort of payoff or privilege. Not that there weren't critics of the rewarding of such displays, like Lord Burleigh, who objected to Queen Elizabeth giving Edmund Spenser a hundred pounds in recognition of a poem, shouting "All for a song?!?" Sometimes there were market motives (remember Johnson's comment that no one but a blockhead ever wrote for any reason other than money?). Sometimes there were more elevated, socially concerned ideas — think of Sir Philip Sidney's argument that poetry presents us with ideal worlds and characters we can try to emulate, since the poet "goeth hand in hand with nature, not enclosed within the narrow warrant of her gifts, but freely ranging only within the zodiac of his own wit. Nature never set forth the earth in so rich tapestry as divers poets have done... Her world is brazen, the poets only deliver a golden." The old John Donne wrote to save souls ("batter my heart, three-personed God!"), the young John Donne wrote to get laid ("For God's sake hold your tongue, and let me love!"). So: personal advancement, money, getting the girl, social improvement, didacticism: there were a host of good, utilitarian, reasons to write a poem, just as there were a host of good, utilitarian reasons to write a piece of prose. Prose and poetry were close cousins formally (one just had extra verse elements), and close cousins in terms of their telos or purpose.
Then things got weird.
They got weird right around the time the Romantics hit the stage (not that this is a bad thing: indeed, it's the beginning of everything I most love in poetry). Suddenly, poetry wasn't more-or-less continuous with prose, or prose-plus-special-effects: there was what Barthes calls a "discontinuity of language." Poetry was reborn as something very different from prose, and different from verse, too.
Consider Coleridge's famous definition of poetry in the Biographia Literaria. Here, we learn that poetry, unlike prose, generates its own rules. We learn that the poem, unlike prose, looks inward upon itself, seeking a co-ordination of all parts to the whole. And we learn that the statement or use-value or telos of the poem is secondary to its formal composition. Poetry is strangely autonomous, being all about its formal wholeness, rather than any specific external purpose. Poetry, says Coleridge, is a kind of communication “opposed to works of science, by proposing for its immediate object pleasure, not truth; and from all other species… it is discriminated by proposing to itself such delight from the whole, as is compatible with a distinct gratification from each component part." The organic harmony of the “legitimate poem,” Coleridge continues, necessitates that the parts
… support and explain each other; all in their proportion harmonizing with, and supporting the purpose and known influences of metrical arrangement. The philosophic critics of all ages coincide with the ultimate judgment of all countries, in equally denying the praises of a just poem, on the one hand, to a series of striking lines or distiches, each of which, absorbing the whole attention of the reader to itself, becomes disjoined from its context, and forms a separate whole, instead of a harmonizing part; and on the other hand, to an unsustained composition, from which the reader collects rapidly the general result unattracted by the component parts.
It is the fusing of all elements that is essential to the poetic work— so much so that Coleridge comes to include within the category of poetry all works, including those in prose, that reconcile all their diverse elements.
It almost doesn’t matter to Coleridge how such fusing is accomplished. Or, rather, it does matter, in that a particular form of the fusing must not be prescribed: “could a rule be given from without, poetry would cease to be poetry” says Coleridge. This does not mean the work is arbitrary, though, merely that it must generate its own rules from within. Coleridge’s most elegant formulation of this principle comes in the essay “Shakespeare’s Judgment Equal to His Genius,” where we read that “No work of true genius dares want its appropriate form, neither is there any danger of this,” because “genius cannot be lawless; for it is even this that constitutes genius — the power of acting creatively under laws of its own creation." Such self-generated rules are, in fact, Coleridge’s famous “organic form,” arising from within each true aesthetic or poetic work.
This business of the poem’s autonomy, its inner, organic wholeness, and its fusing of disparate elements is familiar enough stuff to students of Romanticism. But it’s worth rehearsing in the present context because of what it implies for the meaning of poetry. Because of its intense organic fusing-together of all of its parts, the poem has what Coleridge, in his essay on Shakespeare’s genius, calls “untranslatableness.” In a true poem, the meaning cannot be paraphrased: “it would be scarcely more difficult to push a stone out from the pyramids with the bare hand,” says Coleridge, “than to alter a word, or the position of a word, in Milton or Shakespeare… without making the author say something else, or something worse." Indeed, the meaning, in the sense we usually reserve for that word as something comprehensible and cognizable, is irrelevant to the work’s status as a poem: the poem’s deepest meaning lies in its activity of bringing together apparently disparate elements.
One refraction of all this is the idea of the symbol. “Symbol” is a term that Coleridge uses inconsistently over the course of his career, but the important sense for the present context comes in a passage of The Statesman’s Manual where Coleridge tells us that, unlike the allegorical figure, the symbol is characterized “above all by the translucence of the eternal through and in the temporal." Unlike the allegory, with its one-to-one relation of sign and meaning, the symbol offers polyvalence and endless suggestive possibilities. Interpreting the symbolic work is an ongoing coming-toward and escaping-of full apprehension (think of Coleridge's own "Kubla Khan" here: it is pregnant with meaning, but you can't pin it down to any particular meaning). The symbolic work is certainly at a far remove from the simple, discursive statement of a particular meaning. As Coleridge states in an appendix to The Statesman’s Manual, “discursive understanding, which forms for itself general notions and terms of classification” can only give us “Clearness without Depth." Poetry, when it is symbolic, gives us depth without clarity. We can never quite say what it means, or what it proposes to accomplish in the world.
So poetry becomes a bit of an embarasment in a utilitarian society (and nineteenth century society was getting more and more utilitarian by the day). Poets were meant to chase mysterious symbols, and create organic wholes. The old laureate job or writing odes to the monarch became something of a joke, out of step with the new notion of what poems were all about. The poet isn't supposed to answer to any patron, or, for that matter, to try to make things happen (there were exceptions — consider Shelley's political rabble-rousing poems, but the general trend was toward organic form, symbol, autonomy of imagination, and all the rest). The poet became, in some sense, a figure outside of society. The poet's work was either a quiet withdrawal from bourgeois utilitarian society, or an angry rejection of it. Take your pick. Either way, poetry was different from prose in two respects: it played by different rules, and it was written for different reasons. All of this was new with Romanticism.
While these things may have started with Romanticism, they continued after the Romantic movement in new permutations and combinations. The French Symboliste movement (Mallarmé and company) was in some ways an intensification of the idea that the poem was beyond paraphrase (it should be like music), beyond utility, beyond the logic of means-and-ends. And, despite some of their anti-Romantic rhetoric, the American New Critics in the 20th century made huge intellectual investments in Coleridge's ideas: "The Heresy of Paraphrase" is a rehashing of Coleridge, and the New Critical ideas of irony and balance are just Coleridge's organic form revisited. The poem was about itself, about form, and not subordinate to anything else. (If the Romantics turned to the autonomy of art because they were critical of utilitarian values, the New Critics were a bit different: there's a real story to be told about how they saw the autonomous poem as an essential component of the autonomy of English as an academic discipline: if poems were essentially different from prose statements, they needed a special academic department. If they were like prose, they could be absorbed into history or sociology or whatever — this actually comes up in their correspondence with one another).
But we're past all that now, right? Well, no. In many ways, I think some of the most thoughtful poets of the last few decades have been practicing a kind of modified or inverted version of Romanticism. Think about elliptical poetry: so much of it is all about the lack of formal coherence that you think it'd be the farthest thing from Coleridge's organic form or its New Critical offshoot, the well-wrought urn. But the deliberate incoherence of elliptical poetry is really out to accomplish the same sorts of things Coleridge outlined. First of all, elliptical techniques are all about differentiating poetry from prose, about upholding what Barthes called "the discontinuity of lanuage." Poetry is different, we see, because it doesn't try for prose coherence. And in the deep ambiguities and incoherencies of elliptical verse, we're looking at effects similar to those Coleridge saw as belonging to the symbol: we avoid paraphrasable meaning, we escape the utilitatian logic of means-and-ends. Some see this as purely a matter of beauty, some see it as a critique of a society that relies on a logic of language to support its logic of power. Again, all of this is vey much in line with the general trend of thinking that runs from Coleridge through Mallarmé, and even through the New Critics.
So: where do we draw the line between poetry and not-poetry? Well, it seems that we've been doing it by insisting on poetry's autonomy, its freedom from use-value, its freedom from specific meaning. Poetry is poetry because it is unlike prose: it is more free, and stands somehow (we like to think) outside of the utilitarian world. That seems to be the best answer we've come up with in the two centuries since the poets left off praising popes and princes in exchange for patronage. I wonder: how long it will last? For now the old answer flies, but someday it, too, will fall.