A.I. Identifies the Six Main Emotional Arcs in Storytelling

One of my colleagues in the Master’s program brought my attention to an article about a group of researchers that used a computer to identify the six most common emotional arcs in storytelling. Even if you aren’t interested in computers or A.I., I highly recommend this piece. It proposes an interesting way in which to view our work and raises questions about the value of a human writer.

Here are the six most common emotional arcs in storytelling, according to this article:

  1. Rags to Riches (rise)
  2. Riches to Rags (fall)
  3. Man in a Hole (fall then rise)
  4. Icarus (rise then fall)
  5. Cinderella (rise then fall then rise)
  6. Oedipus (fall then rise then fall)

I’m not going to go into how they compiled this list; the article explains that well enough. I want to explore the implications for writers.

An illustration used to explain Freytag’s Pyramid, retrieved from the Wikipedia entry on Dramatic Structure
No doubt you’re familiar with Freytag’s Pyramid. Teachers drilled it into me in elementary and middle school. If you aren’t familiar with it, Freytag’s Pyramid is a visual representation of what is considered to be the typical plot arc: exposition, rising action, climax, falling action, and denouement, in that order. (I have placed a generic example above.)

Needless to say, this structure is overly-simplified. Plots are made of mini-rises and falls, minor climaxes, sometimes an ending for a beginning. Plots are often more like mountain ranges than pyramids and some stories, particularly flash fiction and other short works, don’t have clearly-defined plots that match this model. That’s how I view the six emotional arcs.

On the one hand, it’s amusing and insightful to study the emotional arcs of your favorite stories in this light. If you determine that they fall under one or multiple of these arcs, you can try and apply that arc(s) to your own works. You can even turn this lens on your better-received works and try to pinpoint what made them successful emotionally.

Of course, such research would take the magic out of the process and play down the emotional arc of your work and the stories you love to read.

Emotional arcs are only this simple in things like fairy tales, which are purposefully simple in order to aid oral storytelling. In fact, Cinderella most caught the attention of Kurt Vonnegut, whose experiment inspired this computer-generated study. If you have anything more complicated and sophisticated than the original fairy tales, these “six most common emotional arcs” won’t fit your story exactly.

When discussing this issue with the longer and more-complicated Harry Potter and the Deathly Hallows, the researchers claim that “the emotional arc associated with each sub-narrative is clearly visible.” In that case, they separate the sub-narratives from the larger narrative, treating them as their own stories, and so one must wonder if these emotional arcs are truly the most common overall or just the most common for certain lengths of stories.

The opposite may also be true: very short fictional works don’t fit these arcs neatly, either. The study only concerns works between 10,000 and 200,000 words long. What, then, can we say about the relation between these common emotional arcs and flash fictions? Or even the traditional short story? Some may fit these arcs but the intricate weaving of narration, plot, character, structure, and form will make it harder to see which stories fit which arcs.

My personal issues with the results aside, the study raises an intriguing–and perhaps disturbing–topic for writers. Namely, can a machine/computer create original writings? The lead author of this study, Andy Reagan, Ph.D. candidate (as of the article’s July 2016 publication), indicates that there are still many problems to resolve before this idea can come to fruition. Nevertheless, many act as though this possibility is very strong. Competitions for writing bots have already started.

Should writers worry? Are we going to be replaced by writing bots that will produce more entertainment more quickly for less cost? No.

In Jorge Luis Borges’s “The Library of Babel”, one could go mad wandering through the library looking for a desired book. You may find the book you want or you may find something similar but not quite same, entirely unrelated, or nonsensical. Just like this library, a writing bot may produce something profound or it may spew something original but ultimately boring. The difference is that a properly-programmed writing bot probably won’t generate any of the nonsensical tomes.
Image of the English language cover, retrieved from The Library of Babel Wikipedia Entry
Reagan mentions that, among other obstacles, the computer would need to create compelling and meaningful characters and dialogue. This area, more than plot or emotional arc, is where human writers will always dominate writing bots. In fact, I would like to add one more aspect to that list: structure.

You can analyze all the stories in the English-speaking world–all the stories in any language–and compile a database of the most successful emotional arcs, plots, character types, etc. That doesn’t mean that you can then spin this data into a unique creation that people will want to read. Instead, the result will be cookie-cutter plots, stock characters, and predictable emotions. The structure will, more likely than not, be the run-of-the-mill linear structure as well.

How do human writers have the potential to avoid such problems? I phrase the question this way because even the best writers don’t avoid all of these faults in everything they write. To err is to human, after all. But how are we able to navigate these obstacles when we do?

The truth is, no one really knows. One story can keep readers’ attention while a similar story loses it. There’s something, however small and indistinguishable, that the writer puts into the story that helps it succeed. I think the advantage to human writers can be traced back to consciousness.

Consciousness makes us aware of the world around us. We’re able to think in ways that machines can’t, gain insights from seemingly nowhere, and feel a connection with fellow living beings that transcends animalistic instinct. It’s my belief that consciousness is also the result of the writer’s instinct. Whenever we have epiphanies, answers to problems with existing works or inspiration for new ones, whose source we cannot identify, we are experiencing something which machines cannot. We suddenly know what’s wrong with a story even though we can’t explain it. A plot, emotional arc, or scene appears fully-formed in our heads and even though we can’t rationalize it, we know it’s right. These instances demonstrate an awareness that far exceeds observation and analysis; it’s consciousness. Until someone can determine what consciousness is and bottle it, writing bots will never be able to do what human writers do.

In answering “why we write”, Robert Coover said that it is “because there is nothing new under the sun except its expression”. (Here’s a link to a video of his entire answer.) Coover’s words act as a double-edged sword: they support writing and other art because the only way to make things new is to express them in a new way but, at the same time, they imply that no idea is original. I agree with Coover in that every idea (for stories and poems) has already been said. What’s unique is our way of expressing these ideas. That is why no “writing bot” could ever truly replace a human writer. It’s not just what we say, it’s how we say it. And if a scientist does construct a computer which captures that inexplicable essence, the reproduction of consciousness won’t be far behind.


Designed by Stephanie Hoogstad circa 2011

Share Your Thoughts