Joan Houlihan declares that
Like all other forms of writing, poetry is a communication. The evidence is in its release from the poet's brain onto a medium designed to be read. The fact that it was written down, made readable, makes it a communication even if its only reader turns out to be its creator at a later time. Furthermore, whatever one feels about the role of the reader, or author-as-reader, there's no dispute that there is a role—a poem without a reader is not a poem, but just an artifact of the imagination. (The Boston Comment)
Neither poetry nor fiction is "a communication" in the way Houlihan clearly intends the term to be understood here. Poets and novelists do not "communicate" information or messages or ideas or propositions or wisdom or anything else by writing poetry or fiction. If these forms of writing are to be considered methods of communication, they are very poor ones indeed, since in most accomplished poems or novels the best that can be said is that their messages or "points" are communicated in a very roundabout way, a strategy that would seem merely self-defeating if the goal of writing them is to satisfy readers looking for the points being made or the message communicated. Most of the great works of literature would surely by now have been judged failures by the communication test: if the value of those works from the past we still read were to be found in their clearly signaled meanings, their unambiguously announced "themes," we probably would not still be reading them. We would have gotten the point long ago.
That poetry is written down, "designed to be read," doesn't in itself demonstrate it's to be taken as communication, although most of us do admittedly have a harder time separating the medium in which the literary arts are created from the artistic effects of which that medium might be capable than we do with painting or sculpture or music. Since we do use language to directly communicate, we assume all language must be used for that purpose—or that all uses of language can't escape its origins in communication or discourse. We are much more willing to grant that music, say, (the scores of which are also "written down") is something other than communication, in most cases, in fact, would resist the idea that behind the music we like is primarily an effort to communicate ideas or messages. But why is it not possible simply to grant that when poets or novelists set to work they are using language for some purposes that can't be reduced to "communication?" A poem or novel is an artificial construction of words. You may not like what has been constructed in a specific instance, but it hardly seems useful to say that it didn't communicate with you.
All of this is just confirmed if we further consider Houlihan's own contention that a poem communicates "even if its only reader turns out to be its creator at a later time." This seems frankly bizarre. If a poet at some future date "reads" a poem she has written, is she really "communicating" with herself? Would this poet even be aware of what's being communicated? Wouldn't she be looking at the poem's formal qualities, the aptness of its word choice, etc.? Did Emily Dickinson consider herself finally a failure because the vast majority of her poems didn't "communicate" with anyone? Might she have been satisfied simply that she had created hundreds of well-made poems?
Even more bizarre, at least to me, is the claim that "a poem without a reader is not a poem, but just an artifact of the imagination." I will agree that ultimately most writers want readers, but what's wrong with those readers considering a given work as "an artifact of the imagination"? Isn't this the very way to define all works of art? Perhaps the problem Houlihan sees here is not that poems are products of the human imagination, but that they might be regarded as "artifacts," something that has been made by "artificial" means. Presumably poetry ought to be "natural," indeed an effort at communication. This distinction is probably at the heart of most complaints against works of art and literature that go too far in their brazen use of artifice or that are pronounced "obscure." But anyone taking up the writing of poetry and fiction is committed to an endeavor that is inescapably artificial. Poetry is an inherently unnatural disruption of our ordinary sense of what language is for (just ask all those freshmen struggling through intro to lit), but if you really resent writers playing this kind of game with words (game-playing, however, being just as integral to human nature as the need to communicate), you probably shouldn't be reading (or writing) poetry in the first place.
I looked for evidence in Houlihan's essay that I was myself misreading her message, misconstruing her point, but was only reinforced in my analysis by her conclusion, in which she writes of "poets who betray what talent they may have for the approbation of peers, who engage in the worst self-delusion: that they have something to say that can only be said in a poem." Once again we are dealing with the assumption that literature is a forum for "saying something," even if it is something "that can only be said in a poem." I've never been clear exactly what things "can only be said in a poem." If you can reformulate the poem into what it "says," then obviously you have said it in another way. If you can't put what it says into words, then just as obviously it's not saying anything. The only other alternative is that a poem just is what it is, in most ways precisely avoiding saying anything in particular. If what we have in such a piece of writing is a failure to communicate, this failure is the poem's greatest success.