Oedipus Hex

Anil Menon‘s short fiction has appeared in magazines such as Albedo One, Apex Digest, Chiaroscuro, Interzone, LCRW, and Strange Horizons. His most recent story, The Poincare Sutra, can be found in the utterly gorgeous Sybil’s Garage No. 7. His YA/SF novel The Beast With Nine Billion Feet (Zubaan, India) was released in November 2009. It’s been shortlisted for the 2010 Vodafone-Crossword Children’s Fiction Prize. He blogs at Round Dice and can be reached at [email protected] He doesn’t typically refer to himself in third person.
This is a story. This is a true story. This is the story by a boy soldier in Sierra Leone. This is an unpublished story by Kafka, found in the coat pocket of a dead Jew. Paul Verlaine scribbled this story on a napkin at the Le Chat Noir the night before he shot Rimbaud with a 7mm. This story is like a chess game played with a machine. This story was told by your mother the night you were conceived. This is the tale Shahryar told Scheherazade on their wedding night. This is a story never read before. This is a 2,500-year old creation myth extracted from an Egyptian sarcophagus. This story was found in Sigmund Freud’s pocket the night he died.

These stories are not part of the story I will shortly tell. In fact, to avoid testing your patience, here is the story:

In A Time Of War

Now, I lost my father to a time of war. In the place where I live, people wear soft leather bottomed shoes to glide over soil. Our feet do not make a sound. We show respect to those who fell during the wars of our times.

The fog of war buries everything. Something inside told me never to venture farther than the fog allowed because I could never smell danger like my father could.

But the stink of time grew too strong to bear. One sweltering afternoon day, I opened the trunk my mother forbade me to open. Inside were a pair of leather bottomed shoes, a ring, and an empty canvas bag.

The bag had once held heads. For there were men of the earth who hungered for my people’s flesh. If I did not provide them with a sacrifice to abate their sorrows, they would take my body and walk amongst my people like one of the undead. They would find ways to sip their lives into their own empty souls.

Now, I fill the bag with faith and set off to find my father. The fog grows strong, and at these heights my lungs beg for air. As I closed my eyes I could hear my father’s voice guide me along the hidden pathways of the mountain unbeknownst to boys who sit and watch the sun rise and fall in their beds.

Then, as I was about to reach the fog-shrouded top of the mountain, a giant form appeared before me in the shape of soil. As I came closer to it I realized that it was not a giant form but composed of hundreds of small people from the earth. The mountain had come alive, and it did not want me to pass its presence.

I heard a voice say: “If you are my son then where are your father’s leather bottomed shoes and bag and ring?”

How does one claim a mountain as a father? Without hesitance, I lifted my pant legs and began to dance in father’s leather bottomed shoes. The soles breezed across the floor, cutting the mist with rhythmic motions. I then turned the ring on my finger and watched my father rise, soil shedding from his skin. His shaved face and clean hands stood against the paling crowd. This impressed the people who stood before me, as did the fact that my tongue did not bleed from the needle it held.

Mother, my fog, my war goddess, licked her fingers and placed them to my face, wiping the thick layer of dirt away. Then I truly began to look like my father’s son, in form, face, and color.

Now, I wait atop the mountain, fallen, waiting, waiting for my son, in a time of war.

What is this story’s literary value? I was pleased with it but not proud. I composed the damn thing but wasn’t sure I understood it. So I requested a critique from my writers group; most of us had met at last year’s spec-fic workshop at IIT-Kanpur (India). Suresh Jois, a writer with a neuroscience background, remarked that:

“I think its an excellent instance of capturing the utter and total dessication of the pre-, intra- and post-war existential condition from an individual perspective….The narrator is inheritor, recruit, participant, perpetrator, victim, undertaker, chronicler, bard all rolled into one. There are no constants, no directions, no references, except this hallucinatory, Dali-esque montage of events and objects passing by and thru the narrator.”

Abha Iyengar saw mythic aspects to the story:

“I think of this as a warrior’s tale, the hero who has to go fulfill his destiny. And of the cyclical nature of things. And how this war is something that is a constant, it will never go away.”

Pervin Saket felt that:

“Its got this uncomfortably gripping quality that made me wish I hadn’t known it, but also like I have always known it. I would call this one a toxic story.”

Swapna Kishore thought the story, while deeply felt, made little sense:

“The story was raw in emotion, but otherwise made little sense to me. It seemed to mix multiple archetypes, but I’m not an expert at literary analysis, and cannot pin down why it seemed a bit of a mishmash. Not that I don’t like surreal, yet the images seemed sort of odd, and I could not relate to them. They seemed to have all the right elements, but the stringing together didn’t work for me, even in a surreal way. But yes, the emotion and a sense of a story was definitely there. I also found the tenses somewhat confusing.”

So on and so forth. I agreed with all the comments. The story was indeed fragmentary, confusing. It did read like an attempt at representing a traumatized mind. There were clear indications of a quest, in fact, the stuff with bleeding tongues and mothers and fathers suggested an unsuccessful retelling of Oedipus. Specifically, it seemed to say that sons go to war over the motherland because they can’t go to war with their fathers over their mothers. Oedipus was blinded by a stake because he killed his father and bedded his mother. Here the needle stitches silent the son’s tongue, because he will not speak the truth about the “fog of war.” Some such thing.

Like I said, the story had left me pleased but not proud. I thought the story was unclear, but perhaps unclear in an useful way since it suggested many interpretations.

Now, it’s odd that I should speak of my story in this detached manner. But it is not my story. It was mostly written by a program based on Propp’s formalist theories.

The formalist, Vladimir Propp, had studied some hundred odd Russian folktales and found they all followed a story-line with a sequence of thirty-one events or “functions.” Not all of the events were necessarily present, but the order of the events was always the same. Brown University has an automatic story generator based on Propp’s scheme. You check the event boxes, click “generate,” and voila! there’s a story. I started with one such automatically generated story. It didn’t make much sense, but by adding four or five strategic lines and tweaking a few words, an interpretation became possible.

The Oedipus Hex experiment wasn’t intended as a gotcha. Swapna was right the story seemed cobbled together, but the others had been able, as expert readers, to imagine their way out of that difficulty. It wasn’t a Sokal-type effort. The experiment was the outcome of an earlier discussion the group had had on the future of interactive fiction. To me, the experiment suggests several things.

  1. A story’s meaning is not intrinsic to the text.
  2. A story may acquire its interpretative richness from a few critical lines.
  3. Propp got the basic (western) folktale structure right. If his model– based on Russian folktales– resonates for us Indians, then he may have even tapped into something universal.
  4. It’s possible to use a program to “compose” a story rather than “write” it.

And a final point:

  • Identifying a program as the author takes away most of its literary value.

I’m not sure why this is, but I can feel no pride in the final story, weird and mythic though it is. I didn’t feel this way when I solved complicated math problems using mathematical software like Mathematica or Maple. The process is very similar. You enter a problem. The program gives an answer. You fiddle with it, enter new questions. Eventually you have the answer you were looking for. Sometimes it’s worth writing it up and trying to get it published. In your name. No scientific paper ever lists Mathematica as an author. It’d be like listing trees as a co-author.

So will we ever give a Booker award to a program? Will we be willing to give it to a human working with a program? Before this experiment, I would have said “no!”. Now, it’s more like “no.” We allow keyboards and wordprocessor software. Why not AI “helper” programs? I was astonished that so little work was required to take an automated story and make it presentable. The Brown University generator is a village idiot in the world of generative text. Imagine what we could do with Deep Blue versions in ten, fifteen, hundred years? What will it do to poetry, genre fiction, and any other form of text where the structure is relatively rigid?

This is a story many of us could have written. We could have written it without knowing anything about Proppian generators. By a strange coincidence, Abha Iyengar, sent me a poem by Marina J. Neary that proved this exact point. Here it is:

“Do not go up that mountain, little Josh!”
His mother cried and trembled in despair.
“When you’re alone, who will be there to wash
Your bloody knees and stroke your sweaty hair?
Don’t listen to your father; he tells lies.
I swear to you by all that lives and moves,
This world does not deserve your sacrifice.
Sweet child, don’t throw your body to the wolves!
Just walk away, disown your destiny.
Don’t be ashamed to fail this wretched test.
Forsake your friends and come along with me.
Fathers have plans, but mothers know what’s best.”
But Josh looked up and saw the dark clouds gather,
And up the mountain went to meet his father.

Marina Neary is an award-winning poet, playwright, dancer, choreographer, an expert on Victor Hugo, and certainly has no need of Proppian generators. Yet here we are.

If we begin to compose stories as well as write them, then what will it do to the careful distinction we maintain between a literary work’s worth and its biography? I’m not sure we can maintain this distinction. The “aura” of text will surely get affected. Walter Benjamin in his essay The Work of Art in the Age of Mechanical Reproduction (1936) defined the “aura” of an art object to be that which was mechanically irreproducible in it.

Art objects are all about aura. Say you are at an Arts & Crafts fair. You see a bicycle seat and handlebars arranged to represent a bull’s head. Then it’s just that: cute, cool and fun. But what if you’d come across the arrangement in the posh Galerie Louise Leiris in Paris, and learn Picasso had done the arranging and that it was a key artifact of the “found art” movement?

I began by telling you possible histories for a story, any story. Both readers and writers work in the belief that work’s origins– its biography– is independent of its literary value (the fallacy of which Shaw pointed out, with his usual prescience, in Fanny’s First Play). But what Oedipus Hex shows us, I think, is that the human touch will become part of a story’s aura. At that point, a written story will truly become an art object. It also means reading will become an art form, even faintly oracular. After all, what is truly mechanically irreproducible– at least for now– is the act of reading.

Indeed, we can even see an ancient parallel here. Our silicon offspring, children of logic and fiction, have begun to reach for their mother. Perhaps the story will turn out differently this time around.


  1. says

    Yeah, extremely interesting post. Thank you.

    In something like the Proppian generator, I think you have to give some credit to the programmer. I mean, the sentences/paragraphs that the computer puts together were inputted by a human who designed them to have some form of literary resonance, and to fit together regardless of the order they’re put in. So maybe it’s a three-way coalition between programmer, computer and user.

    Heh. Maybe the same thing happens when we write stories. We pick up sentences, paragraphs, ideas and store them in our heads as we live our lives. Then we arrange them in a suitable order based on the concept we want to express in our fiction. Then we tweak them until they make good reading.

  2. says

    Aishwarya: thanks.

    Dylan: Yes, the programmer is a new character in the creative mix. Unless, programs turn out to be the best programmers. :) There’s some indication that programming is not really a human activity, but like chess, is just something we can play on a small-scale. The Brown generator simply combines chunks of pre-cooked text based on your settings. But the kind of generators that Chris Crawford and the interactive storytelling gang have in mind appear to be much more ambitious.


  3. Jaya Bhattacharji Rose says

    Anil, fabulous piece. Its an interesting perspective – creator, critiques, commentator etc all rolled in one. Useful link to preserve for future creative writing gathering, editors and writers.

  4. Bevivek says

    Fabulous piece Anil.

    The day that Propp’s stuff truly becomes semantically close to the Neary, de Aura she is de gone.


    The semantics of Marina Neary’s story are in the story itself unlike in the generated piece where you have to string together an interpretation using all the schemata you can lay your sweaty mental literary hands on. Give the Neary piece to 5 people and they are likely to have similar interpretations. If you take the 2 passage, beyond the fact that both involve a rite of passage for a son up a mountain, they have little else in common. So there is a certain sleight of hand you employ when you write, “Marina Neary is an award-winning poet, playwright, dancer, choreographer, an expert on Victor Hugo, and certainly has no need of Proppian generators. Yet here we are.”

    Actually we are not.

    There are several generating programs which generate syntactically gorgeous but semantically rubbish prose. I am too lazy to send you links but will at a pinch.

    The problem IMO is the lack of schemata or whatever is the current cognitive term for both thematic and hierarchical episodic structures which lead to story understanding and generation in the other direction.

    We are still far from Rybka 4 for story generation.

    Yours truly

    Rybka 5

  5. says

    Rybka 5: Agreed, the syntax/semantics disconnect is the major issue. It’s always been the issue. But AI has traditionally focused on phrase structure/transformational grammar, not story grammar. If there is such a grammar– and I think there is– then we should be able to generate stories that are hard to distinguish from human-produced stories *in specific contexts*. But I’m seeing it more as a collaborative effort between a human and something that’s a freaky mishmash between Doug Lenat’s Cog, the BNC corpus, Henry James’ left femur and a souped up Proppian model. It’s the difference between expecting Mathematica to come up with a theory (which it can in some specific contexts) and using it as an assistant. Allow me to quote myself from an related discussion.

    We have two great examples of software taking over what used to be an exclusively human (creative) domain, namely, (1) chess and (2) symbolic computing. Both problems involves working with combinatorial trees. We also have one great example where software has had very little success, namely, the game of Go. Go has only 2 rules, involves both sides of the brain and is deeply strategic. The last 2 points are particularly interesting. Computers have a hard time with pattern recognition and games of strategy; the first needs a body, and the second needs one to know “why”. Despite appearances, Chess is not a strategic game; it’s a tactical game.

    So here’s a hypothesis:

    Stories that are more Go-like than Chess-like cannot be generated automatically.

    I don’t quite know what that means, but I’m thinking chess-like stories would include folktales, genre pulp and whodoneits.