The Emotional Arcs of Stories Are Dominated by Six Basic Shapes

This content has been archived. It may no longer be accurate or relevant.

From The Computational Story Laboratory (citations omitted):

The power of stories to transfer information and define our own existence has been shown time and again. We are fundamentally driven to find and tell stories, likened to Pan Narrans or Homo Narrativus. Stories are encoded in art, language, and even in the mathematics of physics: We use equations to represent both simple and complicated functions that describe our observations of the real world. In science, we formalize the ideas that best fit our experience with principles such as Occam’s Razor: The simplest story is the one we should trust. We also tend to prefer stories that fit into the molds which are familiar, and reject narratives that do not align with our experience.

We seek to better understand stories that are captured and shared in written form, a medium that since inception has radically changed how information flows. Without evolved cues from tone, facial expression, or body language, written stories are forced to capture the entire transfer of experience on a page. A often integral part of a written story is the emotional experience that is evoked in the reader. Here, we use a simple, robust sentiment analysis tool to extract the reader-perceived emotional content of written stories as they unfold on the page.

We objectively test the theories of folkloristics, specifically the commonality of core stories within societal boundaries. A major component of folkloristics is the study of society and culture through literary analysis. This is sometimes referred to as narratology, which at its core is “a series of events, real or fictional, presented to the reader or the listener”, who further define narrative and plot. In our present treatment, we consider the plot as the “backbone” of events that occur in a chronological sequence. We first find an analogous definition in Aristotle’s theory of the three act plot structure: A central conflict emerges in act one, followed by two major turning points in acts two and three before concluding with a final resolution. While the plot captures the mechanics of a narrative and the structure encodes their delivery, in the present work we examine the emotional arc that is invoked through the words used. The emotional arc of a story does not give us direct information about the plot or the intended meaning of the story, but rather exists as part of the whole narrative. This distinction between the emotional arc and the plot of a story is one point of misunderstanding in other work. Through the identification of motifs, narrative theories allow us to analyze, interpret, describe, and compare stories across cultures and regions of the world. We show that automated extraction of emotional arcs is not only possibly, but can test previous theories and provide new insights with the potential to quantify unobserved trends as the field transitions from data-scarce to data-rich.

. . . .

We consider a range of these theories in turn while noting that plot similarities do not necessitate a concordance of emotional arcs.

  • Three plots: In his 1959 book, Foster-Harris contends that there are three basic patterns of plot (extending from the one central pattern of conflict): the happy ending, the unhappy ending, and the tragedy. In these three versions, the outcome of the story hinges on the nature and fortune of a central character: virtuous, selfish, or struck by fate, respectively.
  • Seven plots: Often espoused as early as elementary school in the United States, we have the notion that plots revolve around the conflict of an individual with either (1) him or herself, (2) nature, (3) another individual, (4) the environment, (5) technology, (6) the supernatural, or (7) a higher power.
  • Seven plots: Representing over three decades of work, Christopher Booker’s The Seven Basic Plots: Why we tell stories describes in great detail seven narrative structures:
    • – Overcoming the monster (e.g., Beowulf ).
    • – Rags to riches (e.g., Cinderella).
    • – The quest (e.g., King Solomons Mines).
    • – Voyage and return (e.g., The Time Machine).
    • – Comedy (e.g., A Midsummer Night’s Dream).
    • – Tragedy (e.g., Anna Karenina).
    • – Rebirth (e.g., Beauty and the Beast).

In addition to these seven, Booker contends that the unhappy ending of all but the tragedy are also possible.

  • Twenty plots: In 20 Master Plots, Ronald Tobias proposes plots that include “quest”, “underdog”, “metamorphosis”, “ascension”, and “descension”.
  • Thirty-six plots: In a translation by Lucille Ray, Georges Polti attempts to reconstruct the 36 plots that he posits Gozzi originally enumerated. These are quite specific and include “rivalry of kinsmen”, “all sacrificed for passion”, both involuntary and voluntary “crimes of love” (with many more on this theme), “pursuit”, and “falling prey to cruelty of misfortune”.

The rejected master’s thesis of Kurt Vonnegut—which he personally considered his greatest contribution— defines the emotional arc of a story on the “Beginning– End” and “Ill Fortune–Great Fortune” axes. Vonnegut finds a remarkable similarity between Cinderella and the origin story of Christianity in the Old Testament, leading us to search for all such groupings. In a recorded lecture available on YouTube, Vonnegut asserted: “There is no reason why the simple shapes of stories can’t be fed into computers, they are beautiful shapes.”

. . . .

For a suitable corpus we draw on the freely available Project Gutenberg data set. We apply rough filters to the entire collection in an attempt to obtain a set of 1,737 books that represent English works of fiction.

. . . .

Using principal component analysis, we find broad support for six emotional arcs:

  • “Rags to riches” (rise).
  • “Tragedy”, or “Riches to rags” (fall).
  • “Man in a hole” (fall–rise)
  • “Icarus” (rise–fall).
  • “Cinderella” (rise–fall–rise).
  • “Oedipus” (fall–rise–fall).

See the rest in the following embedded document.

(Trigger Warning: There is math.)

SixStoryArcs

12 thoughts on “The Emotional Arcs of Stories Are Dominated by Six Basic Shapes”

  1. Agree Ashley

    Id rather set the hawks to fly than explain how the alas are made

    Folklore should have been earthy, muddy, roots dragging in the dirt. Instead it became a, well, as you see…

  2. This is a bit like quantifying painting by analyzing the light frequency of the paints used.

    It tells you everything about the quality of the paint, and nothing about the art.

    Still, very clever.

  3. We use the standard linear algebra technique Singular Value Decomposition (SVD) to find a decomposition of stories onto an orthogonal basis of emotional arcs. (p.3)

    In the holy name of Gauss, what the hell does this gobbledygook mean?

    • Not gobbledygook, just a language you don’t speak.

      I’m tempted to say that Google and Wikipedia are your friends if you want a translation but, if you don’t speak the language in the first place, the wiki entry on SVD, though clear enough, probably won’t help.

      Basic translation: we’ve done all these standard statistical tests and if you accept our way of measuring emotions our conclusion (in section IV on page 6) is that there are six core emotional arcs. The rest is for their fellow academics, and there are not even any particularly illuminating equations.

      But as Ashley implies this hammer of analysis mostly misses the point and the conclusion is not exactly a surprise, though there is little harm in saying that statistics supports intuition.

      • Oh, I speak the language of mathematics or at least the dialects of statistics, abstract algebra, complex analysis, differential and integral calculus, differential equations, Boolean logic, and geometry (Euclidean and non-Euclidean). I figured Singular Value Decomposition was a technique unfamiliar to me. What made me choke was an orthogonal basis of emotional arcs.

        I used to work a lot in 4-space geometries. Don’t ask why. I got to where I could calculate 4-space orthogonal vectors in my head. Those six words I highlighted above are gobbledygook.

        What the frell is an orthogonal to an emotional arc? A logical tangent? If it is an arc in the mathematical sense, then it does not have an orthogonal. A vector tangent to the arc at a point can have one. So did they mean the set of orthogonals to the first derivative of the emotional arc? Of what possible use could that conglomeration be?

        I got a feeling that every last iota of result depends on the values they assigned to the variables at the beginning. And they pulled those out of the air. You can prove anything if you can make up your data.

        Color me skeptical and unimpressed.

        YMMV.

        • antares, my apologies for interpreting your gobbledygook comment incorrectly by assuming without evidence that it indicated mathematical ignorance.

          I took emotional arcs just to be their fancy name for the data set they created by running their Hedonometer (I don’t like that name!) on the windowed data sets from the books being studied. I freely admit that I was not motivated enough to work through the detail of the SVD to see if they were really finding a decomposition … onto an orthogonal basis, it just sounded like the typical jargon one sees in such papers.

          I sympathise with your scepticism – though the actual conclusion seems kind of obvious and probably as good as the various ideas floating around about the number of plots, etc. I find myself oscillating between give them an ignoble … and I guess someone has to do the analysis to prove the obvious.

      • There is a point to these academic studies.
        They are looking for “handles” they can use with the optimizing databases being hyped as AI.

        They figure that if they identify the right variables they can cook up algorithms that will let “AI” craft readable stories. They’ll eventually succeed in getting an AI to spit out a halfway readable litfic narrative, maybe in a decade or so.

        Nobody will be impressed but their peers.

        They’d be better advised to direct their efforts towards jokes (there’s a big market for cheap laughs) or even better, limericks. 😉

        • Too late for limericks, Felix. A quick search turned up several pages of generators.

          Amusingly, one of the top results might interest PG (and antares) – https://case.law/gallery/limericks

          P is where Foster stood talking to Swann.
          This appeal is from an order of Hon.
          The joints were hard.
          In this regard.
          Would it not have been culpable mis con?

          (Keep clicking the new rhyme button, and it might eventually kick out something sensible. Or not.)

  4. PG I really don’t like trigger warnings especially when they are about maths. A good equation is worth a thousand words (though possibly not in most novels).

    • Ah, motifs! I remember having those drilled into my head in religion class. I think I had to write about flood motifs, which led to me reading about Quetzalcoatl vs. Lord of Mictlan, and the Epic of Gilgamesh. I should probably re-read Gilgamesh; the translation I read seemed a little dull.

      Thanks for the link!

Comments are closed.