The Shapes of Stories
February 6, 2015 2:52 AM Subscribe
This question makes no sense.
posted by LogicalDash at 3:27 AM on February 6, 2015
posted by LogicalDash at 3:27 AM on February 6, 2015
Ugh, this question has annoyed me since undergrad. In its literal interpretation it's meaningless, like asking "how many different personalities are there" or "how many colours are there". One, infinity, and any other number are all equally valid answers; the answer depends on how inclined the asker is to subdivide a continuum.
Asking and trying to answer the question can be an interesting and useful excercise for students or writers and a way to explore common plot archetypes and frameworks that have been identified in the past, but declaring that you've literally determined "how many" there are is pointless.
And to the extent they actually explain what the research here involves, although it seems interesting enough as a piece of statistical language analysis, I'm very skeptical of the idea that machine analysis could really tell us anything meaningful along the lines Jockers describes or along the lines of the question as formulated here. Even if AI were strong, which it isn't, its reaction to fiction *still* wouldn't necessarily tell us much interesting from a human perspective.
Maybe I'll eat my words when the bestsellers are all written by the Authortron 3000 but that ain't yet.
posted by Drexen at 3:28 AM on February 6, 2015 [12 favorites]
Asking and trying to answer the question can be an interesting and useful excercise for students or writers and a way to explore common plot archetypes and frameworks that have been identified in the past, but declaring that you've literally determined "how many" there are is pointless.
And to the extent they actually explain what the research here involves, although it seems interesting enough as a piece of statistical language analysis, I'm very skeptical of the idea that machine analysis could really tell us anything meaningful along the lines Jockers describes or along the lines of the question as formulated here. Even if AI were strong, which it isn't, its reaction to fiction *still* wouldn't necessarily tell us much interesting from a human perspective.
Maybe I'll eat my words when the bestsellers are all written by the Authortron 3000 but that ain't yet.
posted by Drexen at 3:28 AM on February 6, 2015 [12 favorites]
He hasn't actually descibed what those plots are - the Vice journalist notes this - but Jocker says on his site he's going to do a follow-up post. Will he ever reveal them? And why keep them secret? Does he plan to make money from them somehow? Or do they reveal a horrifying secret about human nature? Or does he just want people to read everything he has to say and not just skip to the payoff?
posted by BiggerJ at 3:29 AM on February 6, 2015 [1 favorite]
posted by BiggerJ at 3:29 AM on February 6, 2015 [1 favorite]
SPOILERS - they will look sort of like this:
\__/
__/\
/\/\/
posted by Artw at 3:34 AM on February 6, 2015 [23 favorites]
\__/
__/\
/\/\/
posted by Artw at 3:34 AM on February 6, 2015 [23 favorites]
declaring that you've literally determined "how many" there are is pointless.
It's not pointless, it reduces things to the point where they can fit in a headline and you can get a nearly-contentless article like this one written about your work and spread around the web.
"Well, it's complicated" would not be a news story. "Six" is.
posted by Wolfdog at 3:36 AM on February 6, 2015 [3 favorites]
It's not pointless, it reduces things to the point where they can fit in a headline and you can get a nearly-contentless article like this one written about your work and spread around the web.
"Well, it's complicated" would not be a news story. "Six" is.
posted by Wolfdog at 3:36 AM on February 6, 2015 [3 favorites]
The six plots are: happiness, unhappiness, unbridled fury, bridled fury, confusion, that special sort of unending despair you get when you've been left on hold for over an hour.
posted by dng at 3:40 AM on February 6, 2015 [6 favorites]
posted by dng at 3:40 AM on February 6, 2015 [6 favorites]
The those really into story analysis by means of upward and downward trending lines I highly recommend Hamlet's Hit Points By Robin D Laws - it's specifically aimed at applying what's learned from doing that to RPGs, but it has general applications.
posted by Artw at 3:45 AM on February 6, 2015 [3 favorites]
posted by Artw at 3:45 AM on February 6, 2015 [3 favorites]
Plot is for genre scrum... proper literary fiction is above all that. It's about emotional arcs or something
posted by fearfulsymmetry at 3:47 AM on February 6, 2015 [1 favorite]
posted by fearfulsymmetry at 3:47 AM on February 6, 2015 [1 favorite]
I guess that I might as well stop writing my second book then - that's ok, I was at an impasse anyways and nobody really reads any more.
posted by Old'n'Busted at 3:53 AM on February 6, 2015
posted by Old'n'Busted at 3:53 AM on February 6, 2015
I wanted to write a snarky comment, but there are only six kinds and I'm too late to this thread.
posted by tykky at 4:04 AM on February 6, 2015 [11 favorites]
posted by tykky at 4:04 AM on February 6, 2015 [11 favorites]
Christopher Booker reckons there are seven.
1. Overcoming the Monster
2. Rags to Riches
3. The Quest
4. Voyage and Return
5. Comedy
6. Tragedy
7. Rebirth
I waded through the (over long) book years ago so I don't remember the specifics terribly well; I remember better the sense that his starting proposition of "all stories fit into one of these boxes" was misleading insofar as every work of fiction has a lot of different stories - or plots - going on at once. Jaws is clearly an overcoming the monster story, but it's also got the mayor's failure to take the town from rags to riches, it's got Quint on a quest, they literally take a voyage and some of them return, there's comedy, there's tragedy, and there's rebirth (Jaws II).
Also: Christopher Booker - nominative determinacy that he wrote a book much?
posted by The Ultimate Olympian at 4:09 AM on February 6, 2015 [8 favorites]
1. Overcoming the Monster
2. Rags to Riches
3. The Quest
4. Voyage and Return
5. Comedy
6. Tragedy
7. Rebirth
I waded through the (over long) book years ago so I don't remember the specifics terribly well; I remember better the sense that his starting proposition of "all stories fit into one of these boxes" was misleading insofar as every work of fiction has a lot of different stories - or plots - going on at once. Jaws is clearly an overcoming the monster story, but it's also got the mayor's failure to take the town from rags to riches, it's got Quint on a quest, they literally take a voyage and some of them return, there's comedy, there's tragedy, and there's rebirth (Jaws II).
Also: Christopher Booker - nominative determinacy that he wrote a book much?
posted by The Ultimate Olympian at 4:09 AM on February 6, 2015 [8 favorites]
I was at an impasse anyways...
"When in doubt, have a man come through a door with a gun in his hand." - Raymond Chandler
posted by The Ultimate Olympian at 4:12 AM on February 6, 2015 [6 favorites]
"When in doubt, have a man come through a door with a gun in his hand." - Raymond Chandler
posted by The Ultimate Olympian at 4:12 AM on February 6, 2015 [6 favorites]
Best bit of the FPP is the Vonnegut video, because... well... Vonnegut.
posted by The Ultimate Olympian at 4:13 AM on February 6, 2015 [2 favorites]
posted by The Ultimate Olympian at 4:13 AM on February 6, 2015 [2 favorites]
This is the sort of thing that people who don't write stories say at parties because they think it makes them look clever, and nobody argues with them because, ugh, how exhausting. We came here to drink, dude, not hear your cribbed Everything Is A Remix posturing.
posted by Ian A.T. at 4:14 AM on February 6, 2015
posted by Ian A.T. at 4:14 AM on February 6, 2015
But I have told you before:
They are seven, and they are:
Man* vs Man
Man vs Woman
Man vs Himself
Man vs Society
Man vs Nature / God / the gods
Man vs Machine
and finally,
Man vs Alien / the Un-natural
Conveniently, you can see them all in a single movie: The Abyss.
* Feel free to substitute the noun of your choice here, of course.
posted by Herodios at 4:17 AM on February 6, 2015 [3 favorites]
They are seven, and they are:
Man* vs Man
Man vs Woman
Man vs Himself
Man vs Society
Man vs Nature / God / the gods
Man vs Machine
and finally,
Man vs Alien / the Un-natural
Conveniently, you can see them all in a single movie: The Abyss.
* Feel free to substitute the noun of your choice here, of course.
posted by Herodios at 4:17 AM on February 6, 2015 [3 favorites]
For eons it's only been possible to chart plotlines as \__/, __/\, and /\/\/, but now, thanks to the invention of emoji, we can do 📈📉📈📉, 👽💋🐓, or even 🐍💭🎿
posted by oulipian at 4:18 AM on February 6, 2015 [9 favorites]
posted by oulipian at 4:18 AM on February 6, 2015 [9 favorites]
Plot is for genre scrum... proper literary fiction is above all that. It's about emotional arcs or something
If you read the article, though, the researcher is actually looking at emotional arcs through "a model that algorithmically abstracts the structure of plot by looking at how the sentiment changes in a story."
And I'm sorry, researcher, but "happy / sad / happy" is not what most people think of when they think of "plot."
posted by drlith at 4:47 AM on February 6, 2015 [1 favorite]
If you read the article, though
And I'm sorry, researcher, but "happy / sad / happy" is not what most people think of when they think of "plot."
posted by drlith at 4:47 AM on February 6, 2015 [1 favorite]
Vonnegut charted plots on a chalkboard when he spoke at my university. He was great.
posted by Songdog at 4:51 AM on February 6, 2015
posted by Songdog at 4:51 AM on February 6, 2015
Actually, I agree with drlith—Vonnegut charted emotional rise and fall, not plot.
posted by Songdog at 4:56 AM on February 6, 2015
posted by Songdog at 4:56 AM on February 6, 2015
Thanks for posting this, the man of twists and turns.
Metafilter: It's all about emotional arcs or something
Literary critics gonna literary critic. Computational analysis is one of the flavors du jour, and it will pass or recede with time, but it actually does tell us things about the formal characteristics of fiction that are really hard to get at with close/normal reading. Very few people have read enough, or have the time to read enough, to duplicate large-scale DH projects on a per-scholar basis. Plenty of people read a hundred novels, some read a thousand, but I think not so many read ten thousand -- and I think few people can remember and think clearly enough to say something meaningful about trends across that entire set.
drlith, I agree with you about the definition of "plot," but I think it's interesting that Jockers is proposing a different understanding of it. Many writers of fiction are careful to distinguish between the plot of a story (what happens) and the structure of a story (how what happens is presented on the page). I've seen that kind of distinction clearly made more in recent years than in past decades, though obvs writers knew about it. What Jockers seems to me to be talking about seems, on some level, akin to what many writers talk about when discussion the "beats" of a story.
For those wondering why Jockers doesn't spill all the beans, well, there are a number of possible reasons. He may indeed be waiting to reveal the full thing in The Inevitable Monograph. Or, given that scholars who "do" digital humanities sometimes shun traditional publication, he may be revealing his work as he goes -- other projects do this, too. Some would also say that this stripe of scholarship is closer to social scientific methods than the "traditional" humanities, so the articles & books tend to focus a little more on incremental change than the paradigm-busting article of book that every humanist lusts after, more or less secretly. (And, of course, many critics of the digital humanities would say there might not actually be any there there, and that's why no payoff.)
posted by cupcakeninja at 4:56 AM on February 6, 2015 [3 favorites]
Metafilter: It's all about emotional arcs or something
Literary critics gonna literary critic. Computational analysis is one of the flavors du jour, and it will pass or recede with time, but it actually does tell us things about the formal characteristics of fiction that are really hard to get at with close/normal reading. Very few people have read enough, or have the time to read enough, to duplicate large-scale DH projects on a per-scholar basis. Plenty of people read a hundred novels, some read a thousand, but I think not so many read ten thousand -- and I think few people can remember and think clearly enough to say something meaningful about trends across that entire set.
drlith, I agree with you about the definition of "plot," but I think it's interesting that Jockers is proposing a different understanding of it. Many writers of fiction are careful to distinguish between the plot of a story (what happens) and the structure of a story (how what happens is presented on the page). I've seen that kind of distinction clearly made more in recent years than in past decades, though obvs writers knew about it. What Jockers seems to me to be talking about seems, on some level, akin to what many writers talk about when discussion the "beats" of a story.
For those wondering why Jockers doesn't spill all the beans, well, there are a number of possible reasons. He may indeed be waiting to reveal the full thing in The Inevitable Monograph. Or, given that scholars who "do" digital humanities sometimes shun traditional publication, he may be revealing his work as he goes -- other projects do this, too. Some would also say that this stripe of scholarship is closer to social scientific methods than the "traditional" humanities, so the articles & books tend to focus a little more on incremental change than the paradigm-busting article of book that every humanist lusts after, more or less secretly. (And, of course, many critics of the digital humanities would say there might not actually be any there there, and that's why no payoff.)
posted by cupcakeninja at 4:56 AM on February 6, 2015 [3 favorites]
There are way more than six plots, I mean just off the top of my head:
Man vs TSA
Dan Brown vs the thesaurus
Man vs inability to decide whether one loves the vampire or the werewolf
The Man vs Bitcoin
Man vs games journalism
The list goes on and on.
posted by Behemoth at 4:58 AM on February 6, 2015 [22 favorites]
Man vs TSA
Dan Brown vs the thesaurus
Man vs inability to decide whether one loves the vampire or the werewolf
The Man vs Bitcoin
Man vs games journalism
The list goes on and on.
posted by Behemoth at 4:58 AM on February 6, 2015 [22 favorites]
I don't know about the theory but I was surprised to discover that 'pout' had been classified as a negative word. I usually find it rather heartening.
posted by Segundus at 4:58 AM on February 6, 2015
posted by Segundus at 4:58 AM on February 6, 2015
Man vs Monomyth
posted by Artw at 5:02 AM on February 6, 2015 [3 favorites]
posted by Artw at 5:02 AM on February 6, 2015 [3 favorites]
Here are my four:
1. Bad things happen
2. Good things happen
3. A mixture of good and bad things happen
4. Nothing much happens at all
posted by vacapinta at 5:05 AM on February 6, 2015 [7 favorites]
1. Bad things happen
2. Good things happen
3. A mixture of good and bad things happen
4. Nothing much happens at all
posted by vacapinta at 5:05 AM on February 6, 2015 [7 favorites]
Lumpers vs. splitters. But it's like the Kama Sutra, where #479 is just like #478 only with your fingers crossed.
posted by jfuller at 5:19 AM on February 6, 2015
posted by jfuller at 5:19 AM on February 6, 2015
The article doesn't go into much technical detail, but as far as I can tell, this is what this method is doing:
1. Breaking the book into a series of short passages.
2. Counting the number of "positive" and "negative" words in each section.
3. Giving each section an overall positive/negative emotional score based on those counts.
4. Using a complicated method to fit a "sentiment" function to those data points.
5. Running a clustering algorithm to partition those functions into buckets based on similarity.
This is an interesting exercise but it's pretty far away from finding some sort of essential six structures of plot. The biggest problem is that this method does not actually understand what it's reading beyond a very simplistic level. There are many ways authors indicate the "high point" of a plot other than using positive/negative vocabulary, and even the word counting method will be pretty noisy and random. The raw graphs will tend to contain a lot of wild swings of emotion that don't match the subjective experience of an actual reader who understands what's being read.
The solution to this is to fit a smoother function to the data, but that takes things even further away from the actual content of the text because the fitted function is adding structure to the data that isn't really there. And finally the clustering algorithm just gives a set of reasonably similar buckets, rather than some sort of immutable truth like the article implies, and those buckets can easily change if the algorithm or its parameters are changed slightly. You could similarly cluster last names into six groups where Smith, Garcia, and Patel are in cluster 3, but that doesn't mean that there are exactly six kinds of last names. That's why you can't really list what the six plots are that result from this method, clustering just puts similar things together in an efficient way, it doesn't assign easy to understand high level labels to each cluster.
posted by burnmp3s at 5:21 AM on February 6, 2015 [2 favorites]
1. Breaking the book into a series of short passages.
2. Counting the number of "positive" and "negative" words in each section.
3. Giving each section an overall positive/negative emotional score based on those counts.
4. Using a complicated method to fit a "sentiment" function to those data points.
5. Running a clustering algorithm to partition those functions into buckets based on similarity.
This is an interesting exercise but it's pretty far away from finding some sort of essential six structures of plot. The biggest problem is that this method does not actually understand what it's reading beyond a very simplistic level. There are many ways authors indicate the "high point" of a plot other than using positive/negative vocabulary, and even the word counting method will be pretty noisy and random. The raw graphs will tend to contain a lot of wild swings of emotion that don't match the subjective experience of an actual reader who understands what's being read.
The solution to this is to fit a smoother function to the data, but that takes things even further away from the actual content of the text because the fitted function is adding structure to the data that isn't really there. And finally the clustering algorithm just gives a set of reasonably similar buckets, rather than some sort of immutable truth like the article implies, and those buckets can easily change if the algorithm or its parameters are changed slightly. You could similarly cluster last names into six groups where Smith, Garcia, and Patel are in cluster 3, but that doesn't mean that there are exactly six kinds of last names. That's why you can't really list what the six plots are that result from this method, clustering just puts similar things together in an efficient way, it doesn't assign easy to understand high level labels to each cluster.
posted by burnmp3s at 5:21 AM on February 6, 2015 [2 favorites]
Man v. Whatever , etc, are conflicts, not plots.
Researcher is working from the old Russian formalist concepts of fabula (the events, in chronological order) and syzuhet (their artistic arrangement) and using Big Data to analyze words ("happier" versus "sadder") as a key to recognizing syuzhet. The claim is the method reveals a limited number of shapes (six, or sometimes seven), although researcher has not revealed those shapes, yet (presumably that's for the next chapter -- Book Publication).
posted by notyou at 5:23 AM on February 6, 2015 [1 favorite]
Researcher is working from the old Russian formalist concepts of fabula (the events, in chronological order) and syzuhet (their artistic arrangement) and using Big Data to analyze words ("happier" versus "sadder") as a key to recognizing syuzhet. The claim is the method reveals a limited number of shapes (six, or sometimes seven), although researcher has not revealed those shapes, yet (presumably that's for the next chapter -- Book Publication).
posted by notyou at 5:23 AM on February 6, 2015 [1 favorite]
Ugh, this question has annoyed me since undergrad. In its literal interpretation it's meaningless, like asking "how many different personalities are there" or "how many colours are there". One, infinity, and any other number are all equally valid answers; the answer depends on how inclined the asker is to subdivide a continuum.
It's a continuum, but when you look at how existing stories are actually distributed in that continuum, there may be distinct clusters. And if there are, that's an extremely interesting result.
posted by automatronic at 5:31 AM on February 6, 2015
It's a continuum, but when you look at how existing stories are actually distributed in that continuum, there may be distinct clusters. And if there are, that's an extremely interesting result.
posted by automatronic at 5:31 AM on February 6, 2015
There is only one plot: unlikely hero is given a magical tool by an older mentor and then defeats the bad guy. Every single story ever fits into this plot. For example:
Star Wars -- Unlikely Hero: Luke. Older Mentor: Obi Wan Kenobi. Magical Tool: Light Saber. Bad Guy: The Death Star.
The Wind in the Willows -- Unlikely Hero: Mole. Older Mentor: Badger. Magical Tool: Hot food and dry clothes. Bad Guy: Toad's self-destructive impulses.
Flowers in the Attic -- Unlikely Hero: Cathy. Older Mentor: Chris. Magical Tool: Incest. Bad Guy: Their grandmother.
Remembrance of Things Past -- Unlikely Hero: The Narrator. Older Mentor: His mother. Magical Tool: A cookie. Bad Guy: Time.
Really it's not hard to see this guys.
posted by kyrademon at 5:59 AM on February 6, 2015 [8 favorites]
Star Wars -- Unlikely Hero: Luke. Older Mentor: Obi Wan Kenobi. Magical Tool: Light Saber. Bad Guy: The Death Star.
The Wind in the Willows -- Unlikely Hero: Mole. Older Mentor: Badger. Magical Tool: Hot food and dry clothes. Bad Guy: Toad's self-destructive impulses.
Flowers in the Attic -- Unlikely Hero: Cathy. Older Mentor: Chris. Magical Tool: Incest. Bad Guy: Their grandmother.
Remembrance of Things Past -- Unlikely Hero: The Narrator. Older Mentor: His mother. Magical Tool: A cookie. Bad Guy: Time.
Really it's not hard to see this guys.
posted by kyrademon at 5:59 AM on February 6, 2015 [8 favorites]
Mega Shark vs Giant Octopus
posted by jason_steakums at 6:07 AM on February 6, 2015
posted by jason_steakums at 6:07 AM on February 6, 2015
Obviously, this was written by someone who never saw the movie Head...
posted by Alexandra Kitty at 6:19 AM on February 6, 2015 [1 favorite]
posted by Alexandra Kitty at 6:19 AM on February 6, 2015 [1 favorite]
While he isn't using the word plot in a traditional sense, he caught my eye. Looking at the visual structure of a story is really helpful.
A lot of authors use the shrunken manuscript method to figure out if everything is balanced. I looked at these graphs and the first thing I thought was, "This is so cool, I can use this to make my books better."
Of course, I may just be hoping there's a magical tool that can help me understand why some of my titles sell better than others. While the macro answer is always some are better than others, this could help understand why.
posted by headspace at 6:24 AM on February 6, 2015
A lot of authors use the shrunken manuscript method to figure out if everything is balanced. I looked at these graphs and the first thing I thought was, "This is so cool, I can use this to make my books better."
Of course, I may just be hoping there's a magical tool that can help me understand why some of my titles sell better than others. While the macro answer is always some are better than others, this could help understand why.
posted by headspace at 6:24 AM on February 6, 2015
(It will not, however, stop me from writing a reply on Mefi where I use the word "better" sixty-seven times in a row.)
posted by headspace at 6:31 AM on February 6, 2015
posted by headspace at 6:31 AM on February 6, 2015
The "How Many Plots?!?" framing of these articles is stupid clickbait targeted at infuriating people who love books, but the way Jockers is looking at story structure is really interesting. As a writer and reader it feel like a look under the hood of the writer's craft. And that's particularly valuable, because even great writers are often blind to their own tools.
posted by 256 at 6:44 AM on February 6, 2015 [1 favorite]
posted by 256 at 6:44 AM on February 6, 2015 [1 favorite]
Man or Astro-man?
posted by Panjandrum at 6:45 AM on February 6, 2015 [1 favorite]
posted by Panjandrum at 6:45 AM on February 6, 2015 [1 favorite]
The number of plotlines cannot be determined because it's a continuum, much like there are no colors because the spectrum of visible light is a continuum, and putting a number such as "six" on plotlines is nonsense but somehow the distinction between tragedies and comedies still makes sense.
posted by Pyrogenesis at 6:51 AM on February 6, 2015 [1 favorite]
posted by Pyrogenesis at 6:51 AM on February 6, 2015 [1 favorite]
...also there are no general categories or sets and therefore words such as "tree" are meaningless because every individual tree is completely unique and Vladimir Propp was full of shit.
posted by Pyrogenesis at 6:59 AM on February 6, 2015 [1 favorite]
posted by Pyrogenesis at 6:59 AM on February 6, 2015 [1 favorite]
Seems to me Denis Johnston proposed there were seven:
Achilles (the fatal flaw)
Circe (the spider and the fly)
Romeo and Juliet (two lovers meet and then lose each other)
Tristan (the love that cannot be)
Faust (the debt that must be paid)
Cinderella (unrecognized virtue)
Orpheus (the gift taken away)
That said, because he lived into the latter half of the twentieth century, he later added an eighth: the indomitable hero, but I forget who his archetype was there. Could be James Bond, could be Superman, could even be Indiana Jones, I suppose.
posted by ricochet biscuit at 7:19 AM on February 6, 2015
Achilles (the fatal flaw)
Circe (the spider and the fly)
Romeo and Juliet (two lovers meet and then lose each other)
Tristan (the love that cannot be)
Faust (the debt that must be paid)
Cinderella (unrecognized virtue)
Orpheus (the gift taken away)
That said, because he lived into the latter half of the twentieth century, he later added an eighth: the indomitable hero, but I forget who his archetype was there. Could be James Bond, could be Superman, could even be Indiana Jones, I suppose.
posted by ricochet biscuit at 7:19 AM on February 6, 2015
SPOILERS - they will look sort of like this:
\__/
__/\
/\/\/
Go home Laurence Sterne, you're drunk.
posted by yoink at 7:26 AM on February 6, 2015 [4 favorites]
\__/
__/\
/\/\/
Go home Laurence Sterne, you're drunk.
posted by yoink at 7:26 AM on February 6, 2015 [4 favorites]
How does one categorize Italo Calvino's If on a Winter's Night a Traveler? Man versus the shimmering mercury of an ever-shifting narrative?
posted by Turkey Glue at 7:57 AM on February 6, 2015
posted by Turkey Glue at 7:57 AM on February 6, 2015
You're in luck, Turkey Glue. Calvino actually did it himself.
posted by Pyrogenesis at 8:23 AM on February 6, 2015 [5 favorites]
posted by Pyrogenesis at 8:23 AM on February 6, 2015 [5 favorites]
Fantastic! And it's as delightfully complicated as I remember.
posted by Turkey Glue at 8:28 AM on February 6, 2015
posted by Turkey Glue at 8:28 AM on February 6, 2015
Six? Pffft. Amateur.
There's only one plot:
1/ Nothing worth telling you about was happening.
2/ Something worth telling you about happened.
3/ It finished.
posted by yoink at 8:44 AM on February 6, 2015
There's only one plot:
1/ Nothing worth telling you about was happening.
2/ Something worth telling you about happened.
3/ It finished.
posted by yoink at 8:44 AM on February 6, 2015
Saying there are only so many plots is like saying there are only so many melodies.
Once you introduce timbre, it's never quite so easily reduced.
posted by sonascope at 8:53 AM on February 6, 2015
Once you introduce timbre, it's never quite so easily reduced.
posted by sonascope at 8:53 AM on February 6, 2015
No. Saying there are only so many plots is like saying there is a distinction between the major and the minor key. Otherwise yes. Timbre is a good parallel.
posted by Pyrogenesis at 9:27 AM on February 6, 2015
posted by Pyrogenesis at 9:27 AM on February 6, 2015
No. Saying there are only so many plots is like saying there is a distinction between the major and the minor key.
Except that it's possible to give reasonably rigorous definitions of major and minor keys such that two independent analysts asked to sort the same selection of scores into "major" and "minor" will make the same selections. That isn't possible with any of these "there are only so many plots" schemata. With the right hermeneutic legerdemain any plot can be shown to be an example of any of the N ur-plots.
posted by yoink at 9:31 AM on February 6, 2015 [1 favorite]
Except that it's possible to give reasonably rigorous definitions of major and minor keys such that two independent analysts asked to sort the same selection of scores into "major" and "minor" will make the same selections. That isn't possible with any of these "there are only so many plots" schemata. With the right hermeneutic legerdemain any plot can be shown to be an example of any of the N ur-plots.
posted by yoink at 9:31 AM on February 6, 2015 [1 favorite]
There is one plot.* Someone wants something but something stands in the way.
* Unless I'm not feeling overly reductionist on a given day, in which case there are more.
posted by Zed at 9:48 AM on February 6, 2015
* Unless I'm not feeling overly reductionist on a given day, in which case there are more.
posted by Zed at 9:48 AM on February 6, 2015
That isn't possible with any of these "there are only so many plots" schemata.
I won't defend this particular division, but what you're saying is far, far from being obvious. If that were the case, then both Propp, to whom I already linked above, but perhaps more importantly, most of the work of Claude Levi-Strauss [pdf] - you know, that arguably the important antrhopologist of the 20th century - would be nonsense. And maybe it is, but the reasons for it are far from obvious, and definitely absent from this thread.
posted by Pyrogenesis at 9:53 AM on February 6, 2015
I won't defend this particular division, but what you're saying is far, far from being obvious. If that were the case, then both Propp, to whom I already linked above, but perhaps more importantly, most of the work of Claude Levi-Strauss [pdf] - you know, that arguably the important antrhopologist of the 20th century - would be nonsense. And maybe it is, but the reasons for it are far from obvious, and definitely absent from this thread.
posted by Pyrogenesis at 9:53 AM on February 6, 2015
what you're saying is far, far from being obvious. If that were the case, then both Propp, to whom I already linked above, but perhaps more importantly, most of the work of Claude Levi-Strauss [pdf] - you know, that arguably the important antrhopologist of the 20th century - would be nonsense.
You are aware, I hope, that Levi-Strauss claimed that Propp was an idiot who got everything wrong--and that Propp thought pretty much the same about Levi-Strauss.
Which, you know, pretty much proves my point.
posted by yoink at 9:57 AM on February 6, 2015
You are aware, I hope, that Levi-Strauss claimed that Propp was an idiot who got everything wrong--and that Propp thought pretty much the same about Levi-Strauss.
Which, you know, pretty much proves my point.
posted by yoink at 9:57 AM on February 6, 2015
Which, you know, pretty much proves my point.
The point being that two scientists disliking each other demonstrates that they were both completely wrong to begin with, and as such there's no need to argue for why they were wrong, because it was all some verbose trickery anyway?
I, uh, may have to disagree on this point.
posted by Pyrogenesis at 10:03 AM on February 6, 2015
The point being that two scientists disliking each other demonstrates that they were both completely wrong to begin with, and as such there's no need to argue for why they were wrong, because it was all some verbose trickery anyway?
I, uh, may have to disagree on this point.
posted by Pyrogenesis at 10:03 AM on February 6, 2015
Sorry, can't stay serious while reading story about Jockers discussing literature.
posted by benito.strauss at 10:05 AM on February 6, 2015
posted by benito.strauss at 10:05 AM on February 6, 2015
The point being that two scientists disliking each other demonstrates that they were both completely wrong to begin with
It isn't that they disliked each other, it's that they thought the other person's classification scheme was incorrect. They literally cannot both be right--which you seem to want to believe.
You hold them both up as Important Authorities (although neither, in fact, has that kind of status in their respective fields today--Important Historical Figures, yes, but nobody much in either folklore studies or anthropology is doing work that simply applies Propp's or Levi-Strauss's frameworks), but they are, precisely, like two musicologists who can't agree at all on which pieces are in a minor key and which are in a major key.
And more to the point, no two structuralist anthropologists back in the good old days agreed about how to apply structuralist analysis to any given culture. It's all very well coming up with schemata, but in all aspects of human signifying behavior the devil remains firmly lodged in the interpretive detail.
posted by yoink at 10:17 AM on February 6, 2015
It isn't that they disliked each other, it's that they thought the other person's classification scheme was incorrect. They literally cannot both be right--which you seem to want to believe.
You hold them both up as Important Authorities (although neither, in fact, has that kind of status in their respective fields today--Important Historical Figures, yes, but nobody much in either folklore studies or anthropology is doing work that simply applies Propp's or Levi-Strauss's frameworks), but they are, precisely, like two musicologists who can't agree at all on which pieces are in a minor key and which are in a major key.
And more to the point, no two structuralist anthropologists back in the good old days agreed about how to apply structuralist analysis to any given culture. It's all very well coming up with schemata, but in all aspects of human signifying behavior the devil remains firmly lodged in the interpretive detail.
posted by yoink at 10:17 AM on February 6, 2015
I piss all over his six plots.
Then I buy him some ice cream.
Then I punch him in the face.
posted by sexyrobot at 10:36 AM on February 6, 2015
Then I buy him some ice cream.
Then I punch him in the face.
posted by sexyrobot at 10:36 AM on February 6, 2015
> This next image shows Dan Brown’s blockbuster novel The Da Vinci Code. Notice how much more regular the fluctuations are. This is the profile of a page turner.
This feeling-based "look at the graphs!" analysis makes the work hard to take seriously. There are mathematical methods to quantify a statement like "time series A has more regular fluctuations than time series B." The author should use them.
> But the full power of these transformed plots does not sit simply in visualization. The values that inform these visualizations can now be compared. In a follow up post, I’ll discuss how I measured and compared 40,000+ plot shapes and then clustered the resulting data in order to reveal six common, perhaps archetypal, plot shapes.
From the author's code on github:
I hope this step was more than "run principal component analysis on the first 3 Fourier coefficients," because yeah... no shit, that will make lots of data sets look the same.
posted by scose at 11:00 AM on February 6, 2015 [3 favorites]
This feeling-based "look at the graphs!" analysis makes the work hard to take seriously. There are mathematical methods to quantify a statement like "time series A has more regular fluctuations than time series B." The author should use them.
> But the full power of these transformed plots does not sit simply in visualization. The values that inform these visualizations can now be compared. In a follow up post, I’ll discuss how I measured and compared 40,000+ plot shapes and then clustered the resulting data in order to reveal six common, perhaps archetypal, plot shapes.
From the author's code on github:
get_transformed_values <- function(raw_values, low_pass_size = 3, ...) {
values_fft <- fft(raw_values)
keepers <- values_fft[1:low_pass_size]
...
}I hope this step was more than "run principal component analysis on the first 3 Fourier coefficients," because yeah... no shit, that will make lots of data sets look the same.
posted by scose at 11:00 AM on February 6, 2015 [3 favorites]
1. He fixes the cable.
posted by Cookiebastard at 11:22 AM on February 6, 2015 [1 favorite]
posted by Cookiebastard at 11:22 AM on February 6, 2015 [1 favorite]
Matthew Jockers says six.
Actually, according to the last link, it's not Matthew Jockers who came to the conclusion, but a computer.
From the first link: “There’s no reason why the simple shapes of stories can’t be fed into computers.” There are probably still writers who find that statement provocative.
Of course, you can feed shapes of stories into computers. in fact, you can feed everything into computers. Only an idiot would find that statement provocative. What you cannot expect is that the computer has some deeper insight based on informed reasoning applied to the junk that you feed into it. In fact, in the sort of statistical modeling described here, computers will only give you the equivalent of coffee dregs and it is up to the human experimenter to "interpret" the results. The same is incidentally true for any computer models on the future, such as climate change. In any model of minimum complexity, there are so many knobs you can turn (i.e. parameters to adjust) that you can predict anything depending on where you set those knobs. In practice, most computer models are tweaked (by making "assumptions") so that the result is something that "looks about right," so they are heavily tainted by the preconceptions of the people who come up with the model.
Ugh, this question has annoyed me since undergrad. In its literal interpretation it's meaningless, like asking "how many different personalities are there" or "how many colours are there". One, infinity, and any other number are all equally valid answers; the answer depends on how inclined the asker is to subdivide a continuum.
In the present case, six (or seven) roughly coincides with the number of items that can be committed relatively easily into short-term memory by most people, and I bet that's no accident. If the computer had said "there are 34 different plots", then most readers would probably react with "Yeah? 34? Why 34? Why not 33 or 35?" But six? I bet most people's reaction is "I guess that sounds about right" without giving it any further thought.
How many colors are there? The answer is usually six or seven, depending on who you ask, even though it is a totally arbitrary division of a continuous spectrum.
So here's my theory, which shall henceforth be known as Sour Cream's Law: When people are asked to subdivide a continuous spectrum into smaller categories, they will come up with 5 to 7 categories most of the time.
posted by sour cream at 7:40 AM on February 7, 2015 [3 favorites]
Actually, according to the last link, it's not Matthew Jockers who came to the conclusion, but a computer.
From the first link: “There’s no reason why the simple shapes of stories can’t be fed into computers.” There are probably still writers who find that statement provocative.
Of course, you can feed shapes of stories into computers. in fact, you can feed everything into computers. Only an idiot would find that statement provocative. What you cannot expect is that the computer has some deeper insight based on informed reasoning applied to the junk that you feed into it. In fact, in the sort of statistical modeling described here, computers will only give you the equivalent of coffee dregs and it is up to the human experimenter to "interpret" the results. The same is incidentally true for any computer models on the future, such as climate change. In any model of minimum complexity, there are so many knobs you can turn (i.e. parameters to adjust) that you can predict anything depending on where you set those knobs. In practice, most computer models are tweaked (by making "assumptions") so that the result is something that "looks about right," so they are heavily tainted by the preconceptions of the people who come up with the model.
Ugh, this question has annoyed me since undergrad. In its literal interpretation it's meaningless, like asking "how many different personalities are there" or "how many colours are there". One, infinity, and any other number are all equally valid answers; the answer depends on how inclined the asker is to subdivide a continuum.
In the present case, six (or seven) roughly coincides with the number of items that can be committed relatively easily into short-term memory by most people, and I bet that's no accident. If the computer had said "there are 34 different plots", then most readers would probably react with "Yeah? 34? Why 34? Why not 33 or 35?" But six? I bet most people's reaction is "I guess that sounds about right" without giving it any further thought.
How many colors are there? The answer is usually six or seven, depending on who you ask, even though it is a totally arbitrary division of a continuous spectrum.
So here's my theory, which shall henceforth be known as Sour Cream's Law: When people are asked to subdivide a continuous spectrum into smaller categories, they will come up with 5 to 7 categories most of the time.
posted by sour cream at 7:40 AM on February 7, 2015 [3 favorites]
Point in case: There are 5 continents if you count the Americas as one (that's why we have 5 Olympic rings), 6 if you count them as two and 7 if you include Antarctica.
posted by sour cream at 7:45 AM on February 7, 2015
posted by sour cream at 7:45 AM on February 7, 2015
The Onion: Literary Study Finds All Modern Narratives Derived From Classic ‘Alien Vs. Predator’ Conflict
posted by 1970s Antihero at 12:08 PM on February 10, 2015 [3 favorites]
posted by 1970s Antihero at 12:08 PM on February 10, 2015 [3 favorites]
« Older Ex-Military Leaders Debate Senators On Closing US... | The Secret Life of Tim Hunkin's Personal Novelty... Newer »
This thread has been archived and is closed to new comments
posted by Artw at 3:13 AM on February 6, 2015 [7 favorites]