The Beginning
I'm doing a massive worldbuilding project by gamifying incremental changes over time
The Beginnings:
What is this?
I’ve decided to start a worldbuilding project from scratch. Like most of the people who might read this, I’m an avid reader and have loved worldbuilding since I read JRR Tolkien’s Appendices. For a long time, I’ve thought how an avian species would look at the world differently, how their history would evolve, and how their mythology and fiction would as well. I’ve decided to finally put my curiosity into practice, and figure it out.
What do I mean ‘from scratch?’
In one sense, it means nothing. Unless you’re working off of a pre-existing world, every worldbuilding project inherently starts from scratch — that is, there wasn’t anything to start from. In another sense, it’s impossible, since everyone comes at any worldbuilding with all the baggage that comes from being immersed in storytelling since childhood. Anything I write is informed by all that I’ve read in the past.
In my case, however, I’m using the phrase ‘from scratch’ to represent a project that is going to worldbuild a species and society from the very beginning of pre-history. My project starts with an avian species first developing sentience in a small corner of a large world, like the first Homo Sapiens developing sentience in Africa. I will develop the history, mythology and language of this species along with their history, as they expand outwards and develop societies and culture.
The idea is that incremental changes over time will create a more believable and rich tapestry of a world than hand-crafting each part individually.
Isn’t this a hopelessly over-large project suffering from delusions of grandeur?
Yes. Fundamentally, this will end up having exponential complexity, as the species spreads out across the planet, if I try to simulate each part of the species. Of course, most worldbuilding projects are hopelessly optimistic, and that’s part of the fun of it.
There are two things that I will be doing that will help constrain the difficulty of this. First, I will try to use some sort of gaming mechanics (rolling dice to determine outcomes) to prevent having to decide every small decision in the outcomes of the avian species. Second, I will use ChatGPT and other AI tools to help come up with concepts. For the most part, I won’t merely copy-paste outputs from AI tools. In my experience, they’re generally not quite at that point yet, though who knows what will happen in the future. Instead, I’ll use the outputs as inspiration and a base to work from.
The World:
I started off by using the donjon fantasy world generator to create a world map. The map comes complete with hexes filled with different types of terrain. The map I created is below:
From several attempts, I noted that the generator tends to create maps with either one major landmass and the rest ocean, or one major inland sea and the rest land, depending on how you set it up.
As you can see, the map is definitely usable, and the major advantage is that it took only about three button-clicks to create it. However, it also has a few major disadvantages, first among them the fact that it doesn’t really look realistic. Plausibly that’s because we are conditioned to think of a planet to look like our own, with major continents. Nonetheless, the large mountain range taking up the entire southeast of the planet, and the ocean taking up the entire northwest, just looks unlikely somehow. In addition, the lack of major separations between any portions of the map also limit story-telling and cultural development possibilities. It also just produces a .jpg file for the map, which limits my possibilities for later editing and changes.
In the end, although the donjon map would have been usable, I decided against using it. If I’m going to make a major project of the scale that I’m planning to, it would be too limiting to do so, and I would probably regret it later.
My second attempt was made using FractalTerrains 3, a map generator published by ProFantasy Software. Where the donjon generator was simplicity itself, FT3 is incredibly complex. I had a lot of fun playing around with it, and it definitely offered ways to get closer to what I had in mind. Below are two of the maps I generated with it — where donjon took about five minutes, each of these took an hour or two. You’ll notice the large craters in several portions of both maps - FT3 includes the ability to send asteroids into the map as well. Features like that were the reason I took so long - although you can generate a map very quickly, there are tons of different tweaks you can make, including the ability to change the map projection you use.
Although I could have kept on playing with this for untold hours more, I decided to call it a day and stick with the first of the two FT3 maps I generated. Here it is again in a cool Mollweide projection:
Each time I look at the map, I think - ‘oh, maybe I should make a few more tweaks,’ but I decided to just stick with what I have and move on.
The Biome:
Next, I decided to create the home biome of the species that I would follow throughout my worldbuilding journey. I had already decided that it would be an avian species - part of my inspiration was thinking through how a bird-based species would differ from humans in their mentality and culture.
I also decided that it made sense that they would start off in a mountainous biome. There is a mountain range nearly on the equator in the middle of the central continent, and I decided that would be the starting place for my new species.
Specifically, they’ll reside on the eastern side of the mountain range, on the leeward side of the prevailing winds. This area is very hot, and lacks the rainfall that inundates the western side, leaving it drier and more like scrublands. To the south and west of the mountain range is an area of tropical forests (where the rain is dumped prior to this mountain range), while to the east is a savannah.
I used ChatGPT to help me brainstorm a set of creatures for the biome that my new species would inhabit. I used the following prompt:
“I'm worldbuilding a sc-fi world. I'm going to create creatures for each biome in the world I've made. Help me brainstorm the following things for a mountainous biome on the equator: 1 major predator 2 minor predators 1 scavenger 1 major herbivore 2 medium herbivore 2 small herbivore 2 staple food (either for humans or for animals) 2 usable resource”
Although my project is closer to fantasy than sci-fi in inspiration, using the words ‘fantasy’ or just leaving it as world without any modifier, gave me beasts that were too magical and fantastic in nature for my liking.
Using this prompt and minor variations, I received answers like the following:
“Certainly! Here's a conceptualization for a mountainous biome located on the equator in a sci-fi setting:
Major Predator: Solarwing Griffin:
Description: A large raptor-like creature with leathery wings that have solar-absorbing scales. The scales store sunlight and allow it to release searing beams of light or warmth during its hunts.
Behavior: Known to soar high above the peaks, it uses thermal updrafts and solar beams to disorient and hunt its prey.
Minor Predators: 1. Equatorial Lynx:
Description: A medium-sized feline with heat-dissipating fur and retractable claws.
Behavior: Agile and stealthy, this lynx hunts during the cooler hours, using the rugged terrain to its advantage.
2. Cragscale Vipers:
Description: Snakes with scales resembling the mountainous terrain, equipped with heat-seeking pits.
Behavior: They bask in the equatorial sunlight, ambushing warm-blooded prey using their heat detection and venomous bite.
Scavenger: Peakbone Vulture:
Description: A large bird with an elongated neck and keen eyesight.
Behavior: Soars between the mountain ranges, efficiently locating and feeding on carcasses left by other predators.
Major Herbivore: Plateaumast Elephant:
Description: A robust herbivore with pillar-like legs, broad feet for stability, and a trunk adapted to grasp high-altitude vegetation.
Behavior: Roams in familial groups, using its strength to move rocks and access hidden mountain flora.
Medium Herbivores: 1. Cliffgrazer Goats:
Description: Sure-footed goats with rubbery hooves and long, spiral horns used for both defense and digging.
Behavior: Often found on precarious ledges, feeding on mosses and hardy mountain plants.
2. Ridgeback Deer:
Description: Deer with a ridge of spines along their back that can collect and condense moisture from the air.
Behavior: Typically move in small herds, using their spines to capture morning dew or fog for hydration.
Small Herbivores: 1. Stonechip Rabbits:
Description: Small mammals with stone-like fur patterns, large ears for dissipating heat, and strong hind legs.
Behavior: Burrow into the mountainside for shelter, feeding on roots and mountain herbs.
2. Lichenlicker Rodents:
Description: Tiny, nimble creatures with long, bristly tongues.
Behavior: Climb sheer rock faces to lick and consume lichens and tiny insects.
Staple Foods: 1. Mountain Tubers:
Description: Hardy root vegetables that grow in high-altitude soils, packed with nutrients.
Usage: Consumed by various herbivores and can be harvested by humans as a rich source of carbohydrates.
2. Skyberries:
Description: Berries that grow on the elevated edges of cliffs, filled with juicy, water-rich pulp.
Usage: Eaten by a variety of creatures and humans for hydration and nutrition.
Usable Resources: 1. Sunstone Crystals:
Description: Luminescent gems found within mountain caves, charged by equatorial sunlight.
Usage: Can be used as a natural light source, in technology, or as a decorative element.
2. Equator Moss:
Description: A moss that grows at high altitudes, known for its ability to filter and purify air.
Usage: Harvested for bio-filtration systems, breathing apparatus, or medical applications due to its oxygen-enhancing properties.
Feel free to use these ideas as a starting point and adjust them to better fit the intricacies and lore of your sci-fi world. The key is to visualize how the unique conditions of an equatorial mountain range might shape the evolution and behavior of its inhabitants.”
You can see both the promise of some of these ideas, and the issues with them. Most of them are direct analogies to Earth creatures, the names are kind of boring, and some of the additional aspects like the equator moss that can be “harvested for bio-filtration systems” seems to be just adding a sci-fi gloss without thinking about it. This makes sense, since large language models don’t really ‘think’ the way we use the word, they instead search for associations between words, which somehow has the emergent property of looking similar to thinking.
As I’m writing this post, I realize that I could have added something like “avoid direct analogies to terrestrial creatures” to my prompt, and gotten more distinct results. In fact, when I just tested it, I noted a distinct improvement, and that will be my prompt for generating biomes going forward. Nonetheless, I’ve already created the first biome, using the above prompt.
I generated four iterations of the above prompt, and then for each creature, I scanned through the different results looking for inspiration, and transcribed my own interpretation of the best one.
Lastly, I used Stable Diffusion, an AI art generator, to create pictures of some of the animals in this biome. In general, I’ve found that SD is great at creating something with a general vibe of what you’re looking for, but that it takes a great deal of effort to create exactly what you are looking for if you have a specific idea in mind. Luckily, I didn’t really know what these creatures would look like, so I just generated dozens of possibilities and then picked which ones I liked the most. If I try to create additional pictures of these creatures, however, I will find it difficult to recreate the exact likeness.1
My notes on the biome look like this, along with the pictures for those creatures I generated:
Equatorial mountain biome - hot, rocky soil, relatively sparse tree cover
Major Predator: Large raptor-like creature with leathery wings and four legs, uses thermals to stay in the air for hours
Minor Predator: medium-sized feline with heat-dissipating fur and retractable claws, hunts during cooler hours
Minor predator: Snakes with scales, infrared senses, basks on cliff faces during the day and hunts at dusk
Scavenger: Mammal with sturdy legs and keen sense of smell, looks for carcasses but will attack a wounded creature if it thinks it can get away with it
Major Prey: Large herbivore with sturdy legs, hump on back to store water, thick leathery skin, can go days without water
Medium Prey: solitary mammals with hooves that can grip nearly sheer surfaces, go to hard-to-reach vegetation
Medium Prey: small creatures with large ears, primarily nocturnal, jump from rock to rock
Minor Prey: Small mouse with fur that holds heat in (to prevent being spotted by IR senses) - hold heat and then dissipates it when it finds water (or feels safe?). Build underground homes
Minor Prey: small lizard that feeds on insects and takes shelter under rocks
Food: Mountain Grass - tough, fibrous plant, can be used for grain
Food: Mountain tuber - root vegetable that grows beneath the rocky soil
Resource - Stone that emits a low glow, can be used for lighting
Resource - plants with durable fibers, reflect the sun’s rays
I explicitly did not name these creatures, as I want to name them based on the my avian species’ perception of them. So stay tuned for the next post, when I create my new species and their initial language.
As a side-note, here are a few other generations that appeared during my attempts. These weren’t what I was looking for, but I may use them for a future creature. Some of these are very weird looking.
Super interesting! I wonder what you might discover when trying to argue for the capacity of linguistic emergence in avian species rather than taking that as an axiom. Maybe there is a biochemical basis for your contention that finds obstacles in today’s human-dominated era on Earth. Then, you would only need to negatively construct your fictional world.