Alumni Voices with Ellen Kazimer: The Evolution of Literature from Monkeys to Artificial Intelligence

Even with a Hamline MFAC degree in hand, a writer can find validation to be elusive. Published or not, everyone is a critic, and well-intentioned friends or relatives can say something that intensifies our insecurity. They might even repeat that tired old theory that if you put monkeys in a room with typewriters, those monkeys would eventually write Shakespeare.  

Hopefully you’ve read Jonathan Gottschall’s The Storytelling Animal: How Stories Make Us Human and learned that the monkey theory is simply not true. Someone actually tested the hypothesis in 2003. They locked macaw monkeys in a room with typewriters and discovered the monkeys did not produce any literary works. Monkeys did demonstrate a predilection for the letter “s”. I’ve always thought this theory insulted both writers and monkeys. Isn’t it presumptuous to assume monkeys would write Hamlet? Wouldn’t they have their own unique viewpoint and story to tell?

It appears, however, that writers have evolved from monkeys to computers. The Washington Post recently featured the following blurb above a book review:

‘Computers will be making interesting and meaningful contributions to literature within the decade’—Artificial-intelligence expert Malcolm Ryan, describing progress being made designing computer programs that can create original stories. Ryan supervised development of a program that writes moral-laden fables with characters expressing a variety of emotions.

Could artificial intelligent computers replace writers? For several days, I scanned the Washington Post to see if anyone had a comment. I found nothing, so I did a little research.

Professor Ryan is from the University of New South Wales’s Computer Science and Engineering Department and the Director of the UNSW Game Design Lab. One of his students, Margaret Sarlej, has devised a computer program called the Moral Storytelling System or MOSS for short. MOSS generates stories based on morals found in Aesop’s fables. The Guardian ran an article on Ryan and Sarlej with two examples of these MOSS fables.

Sarlej believes that morals and messages are one of the key purposes of storytelling throughout the ages. While this maybe true to a certain extent, I couldn’t help thinking that many of our favorite children’s books laced with satire. How would MOSS emulate the work of Jon Sciezka or the late Maurice Sendak?

Both Sarlej and Ryan found that when they considered all the possible events, outcomes and characters reactions storytelling was an “extremely complex business.” Adding plot to the mix of events, characters and outcomes is no monkey business either. In The Guardian article, Dr. Ryan says, “Computers need everything to be defined logically, but it is very difficult to specify hard and fast rules for plot.”

Indeed. Immediately, I thought of Jane Resh Thomas’s double helix of plot and Jackie Briggs Martin assertion that plot is simply what happens, but heart is what your story is about. Where is the heart of the story when the storyteller is a computer? If we are writing about what haunts us, what haunts a computer? Any thoughts, HAL 9000?

What about fictional time and flashbacks? And dialogue? Computer-generated characters in role-playing games can talk, but do they use subtext? Kelly Easton would say that which remains unspoken remains the why of the story.

Truthfully, combining my limited understanding of computer languages and gaming with children’s literature and writing, I am intrigued by Ryan and Sarlej’s work. Dr. Ryan’s blog sometimes expresses ideas similar to our own as writers. He discusses his disappointment in role-playing games that fail to demonstrate motivation for the main quest. In our parlance, the game designer has failed to establish what the protagonist wants and what he is willing to do to get it. Worth reading was Ryan’s entry on The Tale of Peter Rabbit and how difficult it is to teach a children’s story to a computer. Perhaps he should have started with adult genre fiction that has a set format. After all, children’s books appear deceptively simple, but they are extremely complex.

As it turns out, Ryan believes that artificial intelligence programmers must to work with experts in the field—writers and narrative theorists. He envisions MOSS not as a replacement for writers, but as new way of telling stories. Ryan wants a cross section of gamers, writers and artists to create things never envisioned. As he says in his UNSW online resume, he is interested in “code-based art in which the artist is the programmer.” Picture the mischief Mo Williams or Jon Klassen might create if they wrote computer code.

It is gratifying to know our ability to write stories is not inconsequential or easily replicated by macaw monkeys or artificial intelligence-based computers. Writing is hard, but surely we will welcome new ways of telling stories. Just imagine the stories those macaw monkeys will tell someday with artificial intelligence-based computers.

*

Ellen Kazimer is a January 2014 graduate of the MFAC program. She lives and writes in Virginia.