Edit, April 2014:  The below post refers to 46,656 Poems, something I put on my own website because I had no idea what to do with it.  But I revised it, gave it direction, and it was published as More than 25 Million Poems about the Midwest in the final issue of Verse Wisconsin.  I’m pleased with how the project ultimately cohered.  The below describes its genesis and refinement.

* * *

That’s a lot of poems, so if you’re eager to start, you can do so right here.

Recently, I’ve been exploring Twine, “a tool for creating interactive stories.”  In its most basic form, Twine is an intuitive, extremely easy-to-use program to build hypertext stories–à la Choose Your Own Adventure books.  But there’s also some possibility of including basic code under the hood, which made it perfect for resurrecting this project.

I originally wrote 46,656 Poems back in grad school.  I had done some research in the library (remember those?) about Oulipo.  Oulipo generates literature through constraints and aleatory practices, and their techniques were something that spoke very strongly to my actuarial background.  Specifically, I was fascinated by Raymond Queneau’s Cent Mille Milliards de Poèmes (A Hundred Thousand Billion Poems)–a collection of ten sonnets whose lines (and rhymes) were all interchangeable, resulting in 10^14 possible poems.  (An interactive version is here.)

46,656 Poems is a much more modest example, obviously, and dispenses with both meter and rhyme.  There are only six lines, with six choices each, resulting in 6^6 possible combinations.  Since it’s impossible to account for all of the iterations, I mostly focused on making sure each possible first line would feed into each possible second, then each second into each third, etc.  Of course, some work better than others–due to the nature of the concept, the poems swing dramatically from specific imagery to broad assertions.  And just because there are more than forty thousand poems doesn’t mean they’re all good; in fact, it’s pretty easy to come up with a clunker.  But there is something surprising, if not magical, by reading lines that come together in unforeseen ways.

Originally, I had envisioned the project to require minimal user interaction:  just click a button and it will generate a random poem.  I have a spreadsheet that does just that.  Twine could do that too, but I realized that its default structure of asking the reader to make choices was more beneficial, more natural.  As you click through the links, you’re undertaking a process similar to writing a “real” poem:  only half-remembering what came before your chosen line, and with no idea of what might come next.  This project also starts to destabilize some of the terms we frequently use, presuming we know what we mean by them.  Is 46,656 Poems a poem?  Given there are only thirty-six lines total, are there genuinely 46,656 poems?  Are some of them poems, while others aren’t?  There is also the question of authorship:  Who wrote Poem #16,904–me, or the person who clicked through the links to generate that particular poem?  Can one write a poem that requires a reader to finish, to exist?

So, have at it:  46,656 Poems.  You’re only six choices away from your first one.