Scott Alexander accidentally writing about the function of writing prompts · 2:42am June 21st
From today's Astral Codex Ten:
The left is a New York synagogue. It’s Moorish Revival style, which means its architect was working off a basically false and idealized picture of 1100s al-Andalus. The right is a different New York synagogue, by an architect who was “just doing stuff”.
[Implicit question Scott left unsaid: Why is the architecture inspired by a fake tradition more aesthetically interesting than architecture designed with no constraints?]
I don’t have a great explanation for why this should be true. Some of it might be about creativity: “what’s the best idea you can think of?” might not be a fertile prompt. “How did the Moroccans do it?” might be a great prompt, especially if you know nothing about Morocco and you end up getting it totally wrong. Your imaginary version of Morocco, or the way you fill in the blanks in your idealized version of Morocco, or what happens when Morocco collides with everything else in your brain, might be more interesting than whatever you invent from whole cloth.
From the linked blog:
This is why writing satire is so difficult. If you go too ridiculous, you'll break your readers' suspension of disbelief and still be less insane than things that have actually happened.
Gah, tip of my brain, I'm trying to remember who it was that was going on about how creative misunderstandings of your influences are at the root of great art. T.S. Eliot, maybe?
Oh yeah. "No limits" is almost as bad as "incredibly tight limits" for creativity.
I think that's one of the reasons why fanfiction is such a fertile field. Because while it comes with ready-made characters and setting, and even a ready audience, it also puts limits on you. You can break the limits, but that's just another way of being constrained by them!
5787663
The trouble with fiction is that it must be believable.
Now the thing is, I find both of these to be architecturally interesting, but then I've always been a bit of a sucker for that 1970s brutalist neo-deco sort of look.
There's a certain level of pushing the limits of materials technology in both, but in different directions. The Moorish design is seeking intricacy that would have been prohibitively expensive in the era it references; it harkens to a past that may not exist, in ways that were not necessarily possible for the people who inhabited that likely fictional past, to glorify the ideals they strove towards with their design. The concrete abomination (I say this in love) reaches towards an imagined, idealised future in which technology has overcome the limits of mere function and physics, where concrete can be stretched thin as gauze and moth and rust no longer corrupt, and mankind reaches closer still to the face of god.
Religious buildings often embody the dichotomy of reaching for tradition and also reaching for the impossible future, because of their dual purpose of portraying a solid and unchanging foundational faith, but also portraying the magnificence of the object (or objects) of veneration of the faith and the hoped for future that the faith embodies to its followers. The great cathedrals of the middle ages were at the very forefront of architectural design; so too the great mosques of the islamic golden age, and the temples of Jerusalem, and any other religious structure you might care to mention. Even when they were leaning heavily on traditionalist methods and the aesthetics of a glorified past, they did so in ways that looked forward and exploited all the latest understanding of technology and science, because merely replaying the past is not sufficient to glorify a faith; it must be reimagined and expanded, to capture the heart and direct the mind toward a more heavenly perception.
Also surprised that astral codex is still a thing.
I think you're right to focus on how fertile the prompt is. For most us, the way our horse brains generate horse ideas isn't conducive to generating a list of ideas ordered by goodness. We manipulate and compose the ideas that come to mind more easily, and we hope that at least one of the results is good.
If that's actually how we come up with ideas, then anything that factors the problem of composing a good solution would reduce the depth of thought required to come up with good solutions.
Which time period had the most interesting architecture?
For someone that's studied historical architecture, it's easy to come up with some answer that's useful for generating good architectures. Same withWhich buildings had the most impact on future architecture?
The same is probably true for anything that factors the problem of manipulating a solution into a better one.
A building in the middle of a city block should be distinguished by architecture and highlights, not by its full color palette.
Churches should have pointy things.
Everything is better with ponies.
So for your example, even if you did know how Moroccan architecture got it wrong,
How did Moroccans do it?
could still be a great prompt as long as you're simultaneously considering the manipulations required to fix whatever they got wrong.I would guess that prompts are more useful in cases where the depth-of-thought required to come up with a good solution is higher.