Serendipitous Poetry from the New York Times

Yesterday Times Haiku launched on tumblr, posting finds from a computer algorithm written to find unintentional haiku in the New York Times.  Like other "found" poetry algorithms I've posted about before, it starts with a simple premise of scanning sentences for syllable counts.  (Poesytron also starts with syllable counts, and in its simplest form, ends there.)  But the creator, Jacob Harris, Senior Software Architect and self-described "news hacker" for the Times, has added a few bells and whistles to the program that give the end result real beauty and elegance.

First is the added visual element that's incorporated.  I've spent a lot of time thinking about how I present Poesytron's haiku visually, and have opted to make decisions on punctuation, syntax, and layout myself (given that these are very difficult things to make a computer program do).  Harris has instead opted to present every haiku in a left-justified, five-seven-five layout, but has added a visual element that can be determined by a program:

"On every image, you’ll notice a seemingly random background pattern of colored lines. The different orientations of those lines are computer-generated according to the meter of the first line of the poem."

Second is linking each haiku back to the Times article it came from.  Found poetry has, by its very nature, an  incredibly ephemeral quality to it: the act of discovering unintentional poems is happenstance, serendipitous, and dependent on where you happen to be looking at that very moment.  Assigning a computer program to do the looking for you (as opposed to finding haiku in newspapers with a human eye) does not erase those qualities from found poetry.  The Times Haiku algorithm checks the New York Times homepage for newly published articles several times a day.  So the haiku you see on the front page of the tumblr are inextricably linked to what's in the news in the past few days, and gradually an archive of daily-news ephemera will build up on the site.  And yet, the links make it possible to instantly put the haiku back into their original context, which is not something I've seen with found poetry before.

I'm definitely looking forward to watching the Times Haiku project unfold.

Winter Remembering

Leaving a candle,

only winter remembering

in night shadow eyes.

This is the first of Poesytron's poems to literally give me chills.  I post a lot of nonsense here, and that's even after sorting out the really bad output that makes absolutely no sense, no matter how hard I try.  (In case you're wondering, of all the output that I go through, I consider slightly less than half of it worthy of posting.)

But the question I've been asking all along is, if I give a computer some poems, and some simple rules for selecting words out of those poems, can it create meaning?  And if it can, can it create unique meaning--and say something new with the words it rearranges?

And this haiku gives me a resounding YES to that question.  One haiku in the input database Poesytron draws on uses both "candle" and "winter," and another haiku entirely uses "winter" and "remembering."  There is not a single haiku that uses both "candle" and "remembering"--and yet those two ideas work well thematically, linked by a common thread.  And the same can be said for any three words in this haiku.

This doesn't always happen.  I've talked before of Poesytron getting stuck on one haiku in the database, and if that were the case here, I wouldn't be so excited about this result.

Sure, there's some luck involved.  This haiku somehow came together syntactically as well as thematically.  Most of Poesytron's haiku have at least one "a," "of," or "the," stuck in just the right spot to break the syntax and jar the reader out of any potential flow to the poem.  But my goal isn't to give Poesytron an exhaustive list of rules and parameters to try and make it as complexly human as possible.  To paraphrase the great economist E.F. Schumacher, "Any fool can make something complex, but it takes a touch of genius to move in the opposite direction."  My goal is to find out how few rules you need before you generate something more or less "human"--and then, of course, to explore what that means for us as readers as we encounter poems like the one above.  What happens to the obvious "meaning" of the poem when we know Poesytron had no intention behind it?

I don't even begin to have an answer for that yet.  All I know is I find a poignancy in this haiku, and that isn't something that even I expected Poesytron to be able to achieve yet.

Questions

Going through Poesytron's most recently generated haiku, I was struck by this one, and how it very nearly forms a narrative:

Point.  What's deaf?  What's she's?

What's inside?  Dad says, where the

first snow, a autumn.

Snow, my stiff onions

This week, I'll be posting some wintry haiku generated by Poesytron.

Snow, my stiff onions.

Upper-- With boy go to the

snowfalls. Power on.

What I find fascinating in this one is how 'snow' shows up in both the first and last lines.  The Poesytron code attempts to pair consecutive words that get used together, so similar words often end up near each other.  Consider, for example:

ants! she the net the--

as by the snow birthday can't--

ninetieth blow blows

Here we get "the net the" and "blow blows."  In the second case, because Poesytron's database is so small (drawing on 500 haiku written by real people), I can actually track down how the similar words became paired together.  A bit of trawling through let me find these two haiku:

90th birthday
flickering candles
he can't blow out

            - m. harper


second birthday —
he blows out all
the dandelion seeds
            - collin barber

You can tell from comparing Harper's haiku to Poesytron's above the problem that comes from having such a limited database.  (That's something I'm working on.)  Once Poesytron picked the word birthday, it selected Harper's haiku (out of all haiku with 'birthday' in them in the database--a whopping three) for the next word, and it got stuck there, choosing can't, then ninetieth, then blow from the same haiku.  Then, assuming Poesytron picked birthday again, it then couldn't use it due to the syllable count.  So it went looking for another haiku that used birthday, settled on Barber's haiku, and chose blows for the next word.

This is exactly what I want to happen--words that often get used with the same words get used together.  There are more elegant ways I'd like to code this process, of course, but for now even my crude method is working.

Anyway, back to my original point--somehow, in the first haiku I shared, Poesytron went from snow to snowfall, just as it did from blow to blows, but yet it wound its way through several other words first--stiff, onions, boy....  In this case, though, it's not because Poesytron is 'smart' or that there is some cool linguistic chain that can be drawn through snow-stiff-onions-boy-snowfall, but because there are only so many words in the database (1,815, to be exact), so the laws of probability mean that there's a pretty good chance that two words which "go" together will end up in any given haiku.  (And if you're not convinced, read this.)

Haiku for May 23rd: Ice Hard Sharp

Slips sun, ice hard sharp.

Blunt thoughts slices sour thoughts-- is

gone.  Ice.  The flash pines.

This is one of those where the decisions I make on formatting can really change the reading of the haiku.  Is it "Slips sun, ice-hard sharp," or rather "Slips sun-ice, hard sharp"?

Some other formatting options could be:

Slips sun ice hard, sharp

blunt thoughts. Slices sour thoughts is--

Gone ice. The flash. Pines.

Or maybe:

 

Slips. Sun ice hard. Sharp.

Blunt. Thoughts slices sour. Thoughts is

gone. Ice the flash pines.

Poesytron has no "intention" when it creates.  In formatting its haiku for you, the reader, I am another author--or a translator, maybe, an intermediary force that shapes the poem and influences the meaning ascribed by your mind.

I am sure this intermediary step could also be automated, but I am in no way ready to tackle that computing challenge.

Unintentional Poetry: Poetry "Found" on Twitter, Facebook, etc.

One of the inherent themes in computer generated poetry is the unexpectedness of what comes out.  The control over the composition of the poem lies not with a human "author," but partly within some controlled yet random process.

Jackson Mac Low was the poet who first grabbed onto this idea and ran with it.  He used several "diastic" methods, or methods that incorporated algorithms and some random chance.  Mac Low said that by using the methods, the poet "is neither the dictator nor (when he participates in the ensemble) the primary soloist.  He is willing to risk moments in performances that he will not perceive as beautiful.... That is a risk I am deliberately taking."

This idea of deliberate stochasticity is one that, I think, really stretches the boundaries of modern poetry.  It transforms what the poet is able to say within a poem, and transforms the intention of communication within a poem.

When I come across other poetry projects that also push at this boundary, I'm going to share them.  Poesytron 575 is not the only program out there that can do these things--and using a computer program is not the only way to incorporate these concepts into the composition of poetry, either.

One site that I'm very happy to see get started is Internet Poetry.  The mission of the site is to share "poetry* being spread with guerrilla tactics on the internet."  I imagine most of what they post is not going to be poetry that was intentionally written and spread, but rather poetry that is found within the structure of the internet, picked up by a discerning, poetic eye.  This is chance, controlled by the ability to pick out what is poetry (loosely defined) among the textual clutter of the internet.  And I'm very excited to see what beautiful, poignant, stark meaning might emerge.

*loosely defined


Interestingly, there is a very long tradition of haiku embedded within prose.  Matsuo Bashō, one of the first (if not the first) masters of haiku in 17th century Japan, also wrote haibun, which combines prose and haiku.  Jack Kerouac once wrote in a letter, "There are a million haikus in one good prose work," and this has led at least one scholar to look for embedded haiku within Kerouac's own prose.  The logical extension of this idea in the technological age is to look for unintentional haiku posted to Twitter.