20 comments

[ 3.3 ms ] story [ 59.6 ms ] thread
> Here, by classifying the emotional arcs for a filtered subset of 1,327 stories from Project Gutenberg’s fiction collection

Why couldn't they have found 10 more!!

Check out the Vonnegut YouTube link in the doc's references. Talk about the original master class.
Yeah he described these types of arcs way, way back when.
The article talks about composing these shapes using variable coefficients to fit the resulting emotional arcs. I'm not convinced this isn't just Fourier decomposition on arbitrary continuous functions.

Maybe there's insight there regardless? Hard to say. But the title could easily be "continuous functions are approximated by Fourier series with six terms" in my opinion.

>just Fourier decomposition

Yes and their 6 components are actually more like three components if you allow the sign of the amplitude to go negative.

Interestingly, there doesn't seem to be much variation in phase, so it's akin to a real amplitude (though the highest-frequency component does look a bit like it may be time-reversed rather than simply inverted)

Ha! - this was from my friend's research group at Adelaide Uni.

It's surprising how often it continues to pop up.

I was reading a book about story structure, and it has some complicated analysis steps and I was thinking it would be fun to throw an LLM at a book and have it give you premise, theme, structural elements, similarly structured works, etc.
Northrop Frye in "The Great Code: The Bible and Literature" has done a related analysis in the 80s.

I think he identifies 2 primary Biblical narratives - U-shaped plot and inverted U plot - and their role in shaping Western literary and cultural tradition.

U-shaped plot is a descent into a difficult or chaotic situation that represents a fall from a higher state, then proceeding through a series of complications and conflicts that further the sense of alienation or suffering until there is a turning point that leads to a rise in fortune or circumstance, with a positive resolution and the protagonist achieving some form of redemption, success, or happiness.

This could be a literal descent, like being cast out of Paradise, or the Prodigal Son, Job, or Joseph of Egypt (thrown in a well, sold as slave, then in prison), or a metaphorical one, such as losing one's way morally or spiritually.

Inverted U plot examples: king Saul, Samson.

These archetypal narrative structures have been reflected and reinterpreted in literature throughout history.

Samson arguably is U shaped. He goes from hero to prison (and blinded), but wins a huge victory at the end (though it kills him).
Very similar idea to Joseph Campbell's "The Hero with a Thousand Faces". [1] That almost all mythological plots tend to be downward U decent into challenges / temptations and return from the abyss, with an upward U return portion where the hero returns through the known and once again receives the call to adventure or another temptation.

[1] https://en.wikipedia.org/wiki/The_Hero_with_a_Thousand_Faces

Why stop at 6? Why not rise-fall-rise-fall and fall-rise-fall-rise? And so on?
Are you by any chance an epic fantasy author? Brandon Sanderson perhaps? ;)
Combinatorics gives the same six possible shapes as a basis set if stories are limited to at most four emotional states:

1. low, high

2. high, low

3. low, high, low

4. high, low, high

5. low, high, low, high

6. high, low, high, low

So, I guess it's mostly saying that there are at most four states... but the Harry Potter graph in the paper looks like 17 states would explain the data better.

Part of the issue is the natural binary focus. If all states only have a positive / negative interaction with each other, then there can be no prime states other than 2. Silly Fry, 2's don't exist, only 1 and 0.

From a computer perspective, those are avoided kind of like how all games only want texture in 2 powers, because all other sizes are inconvenient. So everything that's ingested from media like video games is always arriving in a power of 2 format.

This is likely simply a statistical artifact because they have no controls.

It would have been very simple to add controls to the book. Include books that should have no arc. For example, a compilation of sonnets or non-fiction.

If you dig through their data in the appendix, you'll see that they did include this but never analyzed it. And the results if you do are really bad for the paper.

https://www.gutenberg.org/ebooks/38887 Why should this book have an arc? It's non-fiction.

https://www.gutenberg.org/ebooks/19994 is a collection of short stories. Again they claim it has an arc.

https://www.gutenberg.org/ebooks/22495 this is literally a collection of puns, that's it. Yet it has an amazing arc according to their data. How is this even remotely plausible?

https://www.gutenberg.org/ebooks/1041 they claim Shakespeare's sonnets have an arc. But.. people can't even agree about the order they're in and in what way/if it's coherent. How can there be an arc?

Seems very dubious.

I vaguely recall Thomas Kuhn saying that normal science is finding out the things you already know using different equations
I have to say that whenever anyone comes along and renders anything to such a basic model I find the model is not interesting and bores me, it's the details that are notable.
That’s cause the product we’re consuming isn’t the conscious content. It’s the unconscious content. We think we respond to plot but we actually respond to the textural elements of narrative: all those little details that we barely notice. Whether or not a story successfully creates the illusion that you’re invested in the plot rather than the details, is the difference between a successful story and a failed story.

It’s a bit like putting icing on a cake. We may think we want to eat the cake because the icing looks good, but the fact of the matter is 90% of that cake is cake batter. If the batter is bad, then the cake is bad and if the batter is good, then the cake is good. The icing is only valuable in that it creates the illusion of a cohesive whole cake, allowing you to misattribute the deliciousness of the cake to a combination of flavors rather than the individual parts.