Maybe there's some technical definitions for the words used to describe the films, but match between archetype and movie is incomprehensible to me. I'm not seeing a single one where the film is where I'd expect based on the descriptions
If you aren't a filmmaker that can be tough. Filmmakers think in beats - for everyone else it is like looking at the skeleton and trying to guess the shape of the animal.
Consider the field of comparative mythology described by Joseph Campbell in "The Hero with a Thousand Faces". [0] As a side quest, check out James Frazer's "The Golden Bough". [1]
On a more critical note: separating a story (diegesis) from the telling (narrative structure) is like reducing animals to skeletons, disregarding the sinews, fascia, nerves, flesh, and fluids that make up an animal.
For example, describing Alain Robbe-Grillet's "La Jalousie" as a story about a man who suspects his lover of adultery is reductive to the point of atrocity. That particular novel tells the story using pioneering metafictive techniques and reducing it to its narrative types would yield very little insight.
The Campbell Structure and the Save the cat structure are the most used structures in hollywood. Audiences are also bored as heck :)
The Campbell structure is the Sa-dominant and is the linear ascent.
You shouldnt seperate a story from a structure at all. You should find the right bones for the story. Not suggesting otherwise in anyway.
PS: if by any chance, you come across the script for that film, do share. Would love to see what it says. Im curious myself now. If i had to guess, reading the synopsis of the film, i'd say : https://arc.quanten.co/archetype/anishchit
It could be interesting to consider nested structures
e.g. in a series there is likely a structure to the season as well as a structure to each episode, as well as potentially longer form structure over multiple seasons
The system does identify story threads (http://arc.quanten.co/showcase/series) - for eg, if you scroll down, you'll realize that there are two story arcs in that episode and that is typical for tv shows, there is a running theme, and then the arc for the season or episode.
The system identifies it, we havent started naming it yet, though we can at some point. I need to work with more showrunners to understand their thinking and process on it. Know anyone?
This is a very ingenious and interesting idea, presented in an esthetically pleasing way. However two related things make me curious. As the identified registers and arcs are very un-orthogonal how useful they are for your intended users? For example how their workflow/decision making process would look like if they discover that their script matches "The Crucible" in 78% and "The Graceful Endurance" in 82%? Also who is an actual target group? Creators themselves or rather agents, producers, editors and other execs or "gatekeepers"?
Most scripts go through a process called "Development" where they are written, and rewritten so many times.
You'll be surprised to know that the first version of Iron man, starring Robert Downing Jr started with the comment from Stan lee as "I don't care for the script" and had to be completely rewritten. So doing 4-5 rewrites is quite normal.
The benefit with a system like this is, you can decide which archetype best tells your story, and also you can benchmark versions of the story to see how close / further away you are as you do your rewrites.
Film executives work with something called comps, what they believe are their comparables - in terms of similar stories and similar structures. The structure of your comp could give you some hints of where you might want to lean towards - or you can go the opposite direction too.
As someone who is building a tool, we take a conscious decision to be at best descriptive + suggestive, but never prescriptive. present the data, let the creators decide how they want to take it forward. I believe that is where originality will thrive.
I have friends who are filmmakers and they'd write me off if i had any other thoughts :)
The primary audience for these are producers and development executives on the studio side, and independent filmmakers if they want a tool to evaluate and give them insights to build on it.
This is called the clock template and helps writers to think about timing - where the midpoint falls, and what has to come before, after etc. Doesn't help much with the overall story structure and beats though.
In the late 19th century, Georges Polti claimed that all human drama can be boiled down to exactly 36 situations [0]. These are much more granular and include things like:
I remember reading that a while back. Thank you for sharing that again.
While there are six general shapes, I've noticed that there are nuances to it. Also the medium significantly changes the form. Just as youtubers creating content follow a different format, compared to tiktokers - even within produced content, I have noticed that Film, Franchise Films, Prestige Drama, Network TV, Mini Series, Shorts etc have nuances.
The paragraph in the beginning reminded me of the 5-step story structure I was taught at school, and I just noticed that it is only featured on the French Wikipedia page [0]. In my experience it worked quite well for classical linear stories, and highlighting it in a text back at school also scored a lot of marks during exams, so now I am somewhat trained at recognizing it.
You are spot on. The simpler version of this is the three step story structure - setup, conflict, resolution. Which is what is used in most pitches etc.
But as stories get more complex, with multiple stories weaving in and also as you bring different genres in, some structures are better than others for different stories.
While I have figured out 15 so far, I want to take the WGA 101 screenplays of all time, which goes all the way from Casablanca, - and i want to see how some of these structures have evolved and are evolving over time.
For eg, since the past 2-3 years, leaving an open end (like in the case of Project Hail Mary in a new universe) shows up in 12% of films, compared to less than 1% before that. Those kind of insights are interesting.
The music analogy is very pleasant. Let me share another stab with similar inspiration. The parallel I'd make is with harmony. For instance in LOTR Dominant is when Frodo and Gollum struggle at the cliff of Mount Doom, they lose control of the Ring that makes an upwards arc spinning. Max tension. Fast forward, Frodo returns to Shire, music is at home again. Tension resolves. (Tonic). So setup, conflict, resolution would be pre-dominant, dominant and tonic. Subplots are secondary dominants.
This is a very interesting concept. I see in your replies to other comments that you are looking at movies from different cultures, which would be a great test of your idea. Once you have sufficiently advanced, it would be great to look at theatre too. I have a hypothesis that movie-writing began to diverge from theatre-writing in the very late 20th century in terms of structure and writing with the rise of the blockbuster and the emphasis on spectacle, and we lost something after that.
Doesn't the format by itself implicitly change the structure of the content? I know some friends who are in theatre, and even when they do 4 shows a week, there are variations that they make constantly that no two shows are the same.
My musician friends who are scholars in indian music tell me that there is a difference between a written raaga and a performed raaga and in a performed raaga, the actor has the right to improvize on that.
Im trying very hard to not go into the rabbit hole which might become purely academic :) But i would think that if we take how youtube content is structured, or tiktoks are, each format would lend to a new structure.
Thats my sense, i could be wrong. What do you think?
I guess I am talking about analysing the written works more than the final acted piece (e.g. a book of Sam Shepard plays vs their performance). That would be good enough, I think.
I totally get why you would want to avoid the rabbit hole but your work is super interesting and I hope that you do get the luxury of being able to dive into adjacent formats and comparing them.
I dont want to say anything and get into trouble like Chalamet :)
But yes, I do believe that if AI is going to get so good at things that we are all going to have free time, we are going to have more time for entertainment. It is either going to be the arena or protests - and theatre might reclaim its glory days.
Ive been watching some of the shows by the National Theatre via streaming and do enjoy them.
Adjacency wise, a few startups have asked if they can use this framework to finetune their storytelling. Im still thinking.
Another view is that theatre is local, low-infrastructure storytelling. It's been so important to us as a species, we don't want to forget it in preference of only following storytelling distributed by corporations.
It should be good now. I spotted a few readability issues on some places, so wanted to push out a quick update so that i dont inconvenience everyone. You can check now.
The question we started off with was - if there are scales and raagas for music, is there something similar for storytelling. What goes well after what beat.
That took us through a journey.
Building Quanten Pulse, which led to Quanten Arc (real data, that led to a model), which then allowed us to create a benchmark database of more than 400 films.
So if you breakdown 400 hollywood blockbusters, and break them scene by scene, map emotions and durations, and character arcs, what is the patterns that you see - and if you step back, do you see clusters of patterns that resonate well.
Most people in hollywood write stories in two structures - predominantly. It is either Save the Cat, or the Heroes journey. But what if you don't want to save cats or go on the journey? (imagine if someone telling a musician, you have two scales - thats it).
We took a peek into the 400 and found 15 different narrative structures that work well. I have a feeling as we expand - into regional cinema, and different formats, we will find more.
I find this notion discombobulating every time it pops up. Just because a particular nuance of an emotion doesn't have its own precise word in the local language, it doesn't mean that the locals don't experience it.
Emotions are universal. Even if some hypothetical language has a particular term for an emotion that in English would fall somewhere between "guilt" and "shame", it doesn't mean that English-speakers don't often experience it; they simply lack a term with the exact nuance, because it rarely matters that much, and we can express the idea with the help of a longer sentence.
If one’s goal is universal appeal, sure. If one’s goal is to capture a very specific time and place and culture, that exact nuance with that name could be very important. Proust spent hundreds of pages in trying to do this sort of thing.
> Emotions are universal. they simply lack a term with the exact nuance
You are mistaking culture for language here. That's mistake number one.
Mistake number two is assuming that a language is merely a purely biological response you can easily map to. Emotions as we conceptualise them, exist in a sociocultural context.
You say emotions are universal but, are they? Have you ever experienced what an edo period Samurai was going through after failing his lord? Ever experienced the feeling of passing your rite of passage in an amazonian tribe. No. You can surely interpret those situations through your own lens and experience feelings about your interpretation but that doesn't mean you are feeling what they are feeling. You could have your own interpretation of what falling from a high altitude feels like. It doesn't mean that is going to match the emotion of someone who has actually jumped.
Mistake three is assuming that your own cultural context (which you have ignored) has the same emotional interpretation of a situation as any other context. A situation that in a cultural context might elicit feelings of belonging, in another might elicit feelings of entrapment, anxiety or lack of freedom.
The very idea that everyone experiences the same emotions is itself a cultural byproduct of a culture that often sees itself as the mirror of the world rather than as an additional perspective of it.
To be fair with you, I have only mapped English Films as of now. I havent had much finding enough Indian film scripts (and am interested in Japanese / Korean films over time), but i do know that the way stories are told - there is a big difference between the way stories are told in europe vs how they are told in Hollywood (even they are both in english).
Not pitching, but giving some context here : The first product we built is an audience analytics platform. We use near eye headsets while audiences are watching films during test screenings, to collect occulometric and biometric data to map engagement.
Emotions are a messy thing to map and rely on. For eg, there was a film that we were testing, and three of the 50 audience members had a totally off-beat engagement signal compared to the rest of the audiences - and this was a scene in the park. So after the film was done, I had flagged the moderator to ask those three audience members what happened in the scene. So he casually brought up what parks meant to them. And one of them said, it was a weekend routine to spend time in the park with his mother whom he lost recently - so the minute he saw that park shot, thats all he could think of, and had nothing to do with the film.
So if you really base a narrative structure on emotions, we will have to baseline every single person on earth for all of their experiences, and wade through that - and it would be impossible to produce a mass media outcome.
Stories hinge on emotion (absolutely) but beyond that it is also a journey, where the audience participates, pays attention, and engages - we see this with occulometric and biometric data. And to truly engage, audiences must understand the words being spoken - i could watch a portuguese film without subtitles and as much as i could guess maybe 10% of the narrative, i wouldnt have a clue whats going on - let alone participate in it.
Our system currently cannot accomodate for languages other than english (we are doing some tests with German, French, Spanish and Hindi - but i wouldnt rank the confidence levels on those to be high enough yet). Beta, at best.
Thank you for asking that very crucial question. I appreciate it.
Seems like a cool idea but it's kinda hard to tell without seeing a whole movie or two fully broken down into those scale steps. Maybe it's there behind a paywall.
Yep, that's definitely on the list. The only issue that I am battling with is how to take something as a positive signal to build these patterns on. Because if i take the entire universe of films, there is probably every variation of arcs. But some have worked / resonated with audiences and some haven't. The way European Cinema has been funded (predominantly through governments) mean that they do the film circuit and then disappear - and the filmmakers are off to make the next film grant. How do i find the signal to identify whats a good film or not.
There was this snide remark that someone in hollywood made where they said, they make movies whereas Europe makes (art) cinema.
I havent figured out how to resolve that yet.
But yes to korean, japanese films - that's very much on the list.
This is all very interesting and I appreciate your commitment to serious research.
> How do i find the signal to identify whats a good film or not.
What is your the definition of 'good'? Without having that, answering the question of a signal is impossible, of course.
Popularity is not necessarily a signal of resonance. It's arguably a consequence of distribution; marketing; appeal to the population of potential movie goers, often thought to be males in their late teens and early 20s going in groups; non-challenging content; etc.
Many films considered 'great' have not been popular, which of course is true in many artforms. Also, what matters more in your measurement, a film that resonates a little for many, or one that resonates a lot for few?
(Sorry if these questions are answered on the website; it's not loading for me.)
I apologise. I had to push a quick update on the site since some folks had readability issues. Please do try again - hopefully its working.
So curiosity aside, I do keep reminding me what problem am I solving and for whom? Auteurs don't care for anyone's opinions and they don't need the tools nor will pay for it.
I definitely do want to study "great" films - not just popular, but films that are timeliness. The next thing on my agenda is to benchmark the WGA 101 Best Screenplays of the 21st century and after that benchmark the WGA 101 Best Screenplays of all time (which starts with Casablanca).
Even if there are no users for it, I'd be curious to find out how things have evolved and what stands out. But as a nimble startups, don't think auteurs and art house filmmakers would be customers. And they shouldnt - they should do what they do, and go with their gut.
so raagas are like scales? i thought it was just a type of music where the same songs are played by every artist, like blues, so maybe i don't get this idea at all. but is it about the order of scenes in movies? or like which scenes are "allowed" in a movie of a particular genre? in any case, are you familiar with the Aarne Thompson Uther index?
Raagas are like scales (which is a massive reduction though) because they are a bit more complicated. Each raaga has time, emotion attached to it as well - there are raagas you can use in a time of a day (in the morning, but not evening for a evening song etc). But yes, scales are the closest it comes to.
There are 72 major raagas - called mela kartha raagas - those are the root raagas, and there are combinations and permutations done that generates the janya raagas - which is children raagas (there are thousands of those - and different artists can create variations on these).
Most films in Hollywood have narrative beats - its 7-8 beats. Each tv show for eg, has 5-6 beats. Most micro drama episode has 2-4 beats. Its quite structure that way.
If you take a structure like Save the cat, or Heroes' journey, the order of the beats are also quite well laid out - just that those two structures dont cover the span of stories, and rest is all quite undocumented.
Im trying to work backwards - and quite aware that i am probably identifying derivatives than the root, but even the derivatives can be quite useful to guide others from generating engaging content.
I am not aware of Aarne Thompson's work. I'm looking it up right now...
Check the sample page for Anora, it might help : http://arc.quanten.co/showcase/film
(scroll down to the bottom to see the narrative beat structure)
On a more critical note: separating a story (diegesis) from the telling (narrative structure) is like reducing animals to skeletons, disregarding the sinews, fascia, nerves, flesh, and fluids that make up an animal.
For example, describing Alain Robbe-Grillet's "La Jalousie" as a story about a man who suspects his lover of adultery is reductive to the point of atrocity. That particular novel tells the story using pioneering metafictive techniques and reducing it to its narrative types would yield very little insight.
[0] https://en.wikipedia.org/wiki/The_Hero_with_a_Thousand_Faces
[1] https://en.wikipedia.org/wiki/The_Golden_Bough
The Campbell structure is the Sa-dominant and is the linear ascent.
You shouldnt seperate a story from a structure at all. You should find the right bones for the story. Not suggesting otherwise in anyway.
PS: if by any chance, you come across the script for that film, do share. Would love to see what it says. Im curious myself now. If i had to guess, reading the synopsis of the film, i'd say : https://arc.quanten.co/archetype/anishchit
e.g. in a series there is likely a structure to the season as well as a structure to each episode, as well as potentially longer form structure over multiple seasons
The system identifies it, we havent started naming it yet, though we can at some point. I need to work with more showrunners to understand their thinking and process on it. Know anyone?
https://en.wikipedia.org/wiki/The_Seven_Basic_Plots#The_plot...
Edit: typo fix + "execs" added
You'll be surprised to know that the first version of Iron man, starring Robert Downing Jr started with the comment from Stan lee as "I don't care for the script" and had to be completely rewritten. So doing 4-5 rewrites is quite normal.
The benefit with a system like this is, you can decide which archetype best tells your story, and also you can benchmark versions of the story to see how close / further away you are as you do your rewrites.
Film executives work with something called comps, what they believe are their comparables - in terms of similar stories and similar structures. The structure of your comp could give you some hints of where you might want to lean towards - or you can go the opposite direction too.
As someone who is building a tool, we take a conscious decision to be at best descriptive + suggestive, but never prescriptive. present the data, let the creators decide how they want to take it forward. I believe that is where originality will thrive.
I have friends who are filmmakers and they'd write me off if i had any other thoughts :)
The primary audience for these are producers and development executives on the studio side, and independent filmmakers if they want a tool to evaluate and give them insights to build on it.
https://blog.quanten.co/why-quanten-arc-does-not-tell-you-wh...
https://channel101.fandom.com/wiki/Story_Structure_101:_Supe...
http://blog.janicehardy.com/2018/07/the-plot-clock-structure...
Appreciate you sharing it.
Slaying of a kinsman unrecognized.
Disaster.
Falling prey to cruelty or misfortune.
The Enigma (solving a mystery).
[0] https://en.wikipedia.org/wiki/The_Thirty-Six_Dramatic_Situat...
Stability Incitement Pursuit Reversal Crisis Revelation Climax Resolution (sometimes) Epilogue / New Beginnings
While there are six general shapes, I've noticed that there are nuances to it. Also the medium significantly changes the form. Just as youtubers creating content follow a different format, compared to tiktokers - even within produced content, I have noticed that Film, Franchise Films, Prestige Drama, Network TV, Mini Series, Shorts etc have nuances.
[0]: https://fr.wikipedia.org/wiki/Sch%C3%A9ma_narratif (https://fr.wikipedia.org/wiki/Sch%C3%A9ma_quinaire is also describing the same thing)
But as stories get more complex, with multiple stories weaving in and also as you bring different genres in, some structures are better than others for different stories.
While I have figured out 15 so far, I want to take the WGA 101 screenplays of all time, which goes all the way from Casablanca, - and i want to see how some of these structures have evolved and are evolving over time.
For eg, since the past 2-3 years, leaving an open end (like in the case of Project Hail Mary in a new universe) shows up in 12% of films, compared to less than 1% before that. Those kind of insights are interesting.
Thanks for sharing that link.
https://postimg.cc/nM9cTkpt
My musician friends who are scholars in indian music tell me that there is a difference between a written raaga and a performed raaga and in a performed raaga, the actor has the right to improvize on that.
Im trying very hard to not go into the rabbit hole which might become purely academic :) But i would think that if we take how youtube content is structured, or tiktoks are, each format would lend to a new structure.
Thats my sense, i could be wrong. What do you think?
I totally get why you would want to avoid the rabbit hole but your work is super interesting and I hope that you do get the luxury of being able to dive into adjacent formats and comparing them.
But yes, I do believe that if AI is going to get so good at things that we are all going to have free time, we are going to have more time for entertainment. It is either going to be the arena or protests - and theatre might reclaim its glory days.
Ive been watching some of the shows by the National Theatre via streaming and do enjoy them.
Adjacency wise, a few startups have asked if they can use this framework to finetune their storytelling. Im still thinking.
Anyway the site is too clever for its own good and crashes out with a "We hit an error" modal overlay on Safari on Mac, so I'll never know.
The question we started off with was - if there are scales and raagas for music, is there something similar for storytelling. What goes well after what beat.
That took us through a journey.
Building Quanten Pulse, which led to Quanten Arc (real data, that led to a model), which then allowed us to create a benchmark database of more than 400 films.
So if you breakdown 400 hollywood blockbusters, and break them scene by scene, map emotions and durations, and character arcs, what is the patterns that you see - and if you step back, do you see clusters of patterns that resonate well.
Most people in hollywood write stories in two structures - predominantly. It is either Save the Cat, or the Heroes journey. But what if you don't want to save cats or go on the journey? (imagine if someone telling a musician, you have two scales - thats it).
We took a peek into the 400 and found 15 different narrative structures that work well. I have a feeling as we expand - into regional cinema, and different formats, we will find more.
Tell me what you think : https://arc.quanten.co/archetype
PS: While we started with Hollywood, we are starting to do this analysis for Bollywood films too (though finding scripts has been difficult)
How do you deal with emotions that only exist (as their own concepts) in certain cultures (saudade in Portugal, hygge in Scandinavia)?
Emotions are universal. Even if some hypothetical language has a particular term for an emotion that in English would fall somewhere between "guilt" and "shame", it doesn't mean that English-speakers don't often experience it; they simply lack a term with the exact nuance, because it rarely matters that much, and we can express the idea with the help of a longer sentence.
You are mistaking culture for language here. That's mistake number one.
Mistake number two is assuming that a language is merely a purely biological response you can easily map to. Emotions as we conceptualise them, exist in a sociocultural context.
You say emotions are universal but, are they? Have you ever experienced what an edo period Samurai was going through after failing his lord? Ever experienced the feeling of passing your rite of passage in an amazonian tribe. No. You can surely interpret those situations through your own lens and experience feelings about your interpretation but that doesn't mean you are feeling what they are feeling. You could have your own interpretation of what falling from a high altitude feels like. It doesn't mean that is going to match the emotion of someone who has actually jumped.
Mistake three is assuming that your own cultural context (which you have ignored) has the same emotional interpretation of a situation as any other context. A situation that in a cultural context might elicit feelings of belonging, in another might elicit feelings of entrapment, anxiety or lack of freedom.
The very idea that everyone experiences the same emotions is itself a cultural byproduct of a culture that often sees itself as the mirror of the world rather than as an additional perspective of it.
Not pitching, but giving some context here : The first product we built is an audience analytics platform. We use near eye headsets while audiences are watching films during test screenings, to collect occulometric and biometric data to map engagement.
Emotions are a messy thing to map and rely on. For eg, there was a film that we were testing, and three of the 50 audience members had a totally off-beat engagement signal compared to the rest of the audiences - and this was a scene in the park. So after the film was done, I had flagged the moderator to ask those three audience members what happened in the scene. So he casually brought up what parks meant to them. And one of them said, it was a weekend routine to spend time in the park with his mother whom he lost recently - so the minute he saw that park shot, thats all he could think of, and had nothing to do with the film.
So if you really base a narrative structure on emotions, we will have to baseline every single person on earth for all of their experiences, and wade through that - and it would be impossible to produce a mass media outcome.
Stories hinge on emotion (absolutely) but beyond that it is also a journey, where the audience participates, pays attention, and engages - we see this with occulometric and biometric data. And to truly engage, audiences must understand the words being spoken - i could watch a portuguese film without subtitles and as much as i could guess maybe 10% of the narrative, i wouldnt have a clue whats going on - let alone participate in it.
Our system currently cannot accomodate for languages other than english (we are doing some tests with German, French, Spanish and Hindi - but i wouldnt rank the confidence levels on those to be high enough yet). Beta, at best.
Thank you for asking that very crucial question. I appreciate it.
http://arc.quanten.co/showcase/film (Anora) http://arc.quanten.co/showcase/series (The Pitt S01E01)
There was this snide remark that someone in hollywood made where they said, they make movies whereas Europe makes (art) cinema.
I havent figured out how to resolve that yet.
But yes to korean, japanese films - that's very much on the list.
> How do i find the signal to identify whats a good film or not.
What is your the definition of 'good'? Without having that, answering the question of a signal is impossible, of course.
Popularity is not necessarily a signal of resonance. It's arguably a consequence of distribution; marketing; appeal to the population of potential movie goers, often thought to be males in their late teens and early 20s going in groups; non-challenging content; etc.
Many films considered 'great' have not been popular, which of course is true in many artforms. Also, what matters more in your measurement, a film that resonates a little for many, or one that resonates a lot for few?
(Sorry if these questions are answered on the website; it's not loading for me.)
So curiosity aside, I do keep reminding me what problem am I solving and for whom? Auteurs don't care for anyone's opinions and they don't need the tools nor will pay for it.
I definitely do want to study "great" films - not just popular, but films that are timeliness. The next thing on my agenda is to benchmark the WGA 101 Best Screenplays of the 21st century and after that benchmark the WGA 101 Best Screenplays of all time (which starts with Casablanca).
Even if there are no users for it, I'd be curious to find out how things have evolved and what stands out. But as a nimble startups, don't think auteurs and art house filmmakers would be customers. And they shouldnt - they should do what they do, and go with their gut.
There are 72 major raagas - called mela kartha raagas - those are the root raagas, and there are combinations and permutations done that generates the janya raagas - which is children raagas (there are thousands of those - and different artists can create variations on these).
Most films in Hollywood have narrative beats - its 7-8 beats. Each tv show for eg, has 5-6 beats. Most micro drama episode has 2-4 beats. Its quite structure that way.
If you take a structure like Save the cat, or Heroes' journey, the order of the beats are also quite well laid out - just that those two structures dont cover the span of stories, and rest is all quite undocumented.
Im trying to work backwards - and quite aware that i am probably identifying derivatives than the root, but even the derivatives can be quite useful to guide others from generating engaging content.
I am not aware of Aarne Thompson's work. I'm looking it up right now...