Can a poem be paraphrased?

At first sight, it would seem obvious that poems can be paraphrased.  After all, one can show that one understands a poem by giving a description of what it means in other words.  In one sense "paraphrase" just means "interpretation."  And yet some have argued that poetry cannot be paraphrased.  The word "paraphrase" simply means "tell in other words" and the OED defines it as an expression in other words that is "usually fuller and clearer, of the sense of a written or spoken passage or text."  The usual object of paraphrasing is clarification of meaning.  A. C. Bradley famously argued in "Poetry for Poetry's Sake" that one cannot paraphrase good or "pure" poetry.  For instance, of a line in Virgil, he says that "if we insist on asking for the meaning of such a poem, we can only be answered 'It means itself.'"  Some other words cannot express the meaning of the poem.  Peter Kivy in his book Philosophies of Arts (1997) and again in Once-Told Tales (2012) argues against Bradley.  He thinks that, although there are some rare poems that cannot be paraphrased, this is only because they have no meaning or content at all.  The poem by Lewis Carroll that starts "Twas brillig, and the slithy toves" does not make sense and cannot be paraphrased.  Kivy also suggests, strangely, that perhaps haiku poems cannot be paraphrased.  (If haiku poems can have content and yet cannot be paraphrased then doesn't this undercut his entire thesis?) Bradley says that Byron's poem Mazeppa cannot be paraphrased since the line "Bring forth the horse!" The horse was brought; In truth it was a noble steed!" would be paraphrased by replacing the word "steed" with "horse" and the word "horse" with "steed" (since they mean the same thing according to the dictionary), and yet this would lose the meaning of the poem.  Kivy thinks that if the poem is replaced by "Bring forth the steed!" The steed was brought; Indeed it was a noble horse" the meaning has not, contra Bradley, changed, although he admits that the aesthetic character has changed.  (Once, 168).  Kivy insists that a paraphrase is not intended as a substitute for a poem, just as an explication or representation of its meaning. 

Kivy seems right in insisting that most poems can be paraphrased.  But is there no truth or insight in what Bradley was trying to say?  Is the form of a poem sufficiently distinct from its content that one can present the content without the form?  Or are the content and form so closely intertwined that more than just the aesthetic feel is lost when one gives an explanation of a poem's meaning?  Kivy thinks one can separate the meaning of a poem from its form, i.e. the way the meaning is expressed.  I wonder whether this is possible.  Note that Kivy is not saying that a poem's meaning may ever be completely paraphrased.  

Still, the question remains whether there is some insight in Bradley's insistence that the poem expresses what it intends to express in its own words and that the paraphrase is inevitably a distortion of that. Kivy says that the content of a poem is its "propositional content" and he observes that there are three kinds of propositional content:  the content of sentences in poems that describe, the content of sentences in poems that narrate events and actions, and the propositions that the poem might express in the way of philosophical or moral theses, and, on Kivy's view it is self-evident that "if a linguistic instrument of any kind has meaning, which is to say, propositional content, including poetry, it is paraphraseable; its meaning, its propositional content, can be expressed in other words."  (171) 

Why does this make me nervous?  It does seem that Kivy is telling us that poetry has no autonomous domain of meaning, and that it can be pretty much replaced by philosophy or science, as far as significant content goes.  Remember that a paraphrase gives the meaning more clearly.  Isn't this a salvo in the battle between poetry and philosophy going back to Plato, Kivy being on Plato's side in attacking poetry as, for the most part, worthless?  What does poetry give us, on Kivy's view, other than a less clear version of what a good prose paraphrase would give us better?  The decorative element, which Kivy refers to as the aesthetic or "form" dimension, looks to be not only superfluous but also productive only of un-clarity.  Note that it is never even considered that the paraphrase could be in other poetic words!  Propositional content, on this view, is only properly presented if it has the clarity of a piece of analytic philosophy. 

Kivy directs his views against Kant's idea that a poem of genius will gives us not only statable content but also a huge collection of  "aesthetic ideas" which is, as Kivy puts it "the true aesthetic content of the poem."  Kivy has a problem with Kant's idea that the true aesthetic content is "ineffable."  It becomes clear in reading Philosophies of Arts that Kivy wants to bring down what I consider to be Kant's greatest insight into aesthetics.  Kivy correctly reads Kant as stressing that the feeling of ineffability enlivens our "cognitive powers and connects language, which should otherwise be mere letters, with spirit" and further that this material happens in such a way as to connect us with the realm of idea, or what Kant calls the supersensible realm (and what ordinary people would call the realm of God).  Taking way the metaphysics I find implausible (I don't believe in God) this means that ineffability gives us a feeling of transcendence that gives meaning to our existence.  I think this makes Kivy uncomfortable (perhaps he thinks it will bring God in through the back door...there is a lot of nervousness about this amongst analytic philosophers who are atheists) and is the source of his denial of Bradley's thesis that poetry cannot be paraphrased. Kivy thinks that Kant's idea of "aesthetic ideas" leads to Bradley's "conflation" of form and content.  I follow Kivy's reasoning but in a different (the opposite?) direction, i.e. that uncovering the inner truth of Bradley's otherwise implausible thesis brings us back to Kant's discovery of the importance of the ineffability of poetic meaning, which Kivy ultimately wishes to undercut (partly because he is a nervous atheist, and partly because, as an analytic philosopher, he wishes to always assure that distinctions between concepts, as for example, between form and content, are rigid....in general the motive behind this is a worry that without rigid conceptual distinctions humanity will descend into chaos...it is noteworthy in this respect that the worst thing an analytic philosopher can say about someone is that they have conflated two concepts).