Forty minutes on the same paragraph. Three opening lines written and deleted. The cursor is doing its thing, the coffee has gone cold, and the honest diagnosis is not that you have no ideas — it's that you have the same four ideas you had yesterday, and you already know how each one ends. The wolf. The raven. The clever fox. The wise old bear who dies in act two.
That is not a shortage of imagination. It's a search problem: left to yourself, you keep sampling from the small pile of images you already find appealing, which is exactly the pile that got you stuck. The way out is to take the choice away from yourself — hand it to something that has never heard of your taste and does not care about it.
Which is the actual, unglamorous job of a random animal generator. Not novelty. Forced sampling from a set you did not curate.
An arbitrary constraint outperforms a blank page
This isn't just vibes. In a 2017 study in Psychology of Aesthetics, Creativity, and the Arts, Catrinel Haught-Tromp had participants write two-line rhymes for an occasion — a birthday, a farewell. When they had to work in a given noun, the rhymes came out rated more creative than when they were free to write whatever they liked, and the benefit carried over: practice with the constraint left people more inventive afterwards. She calls it the Green Eggs and Ham hypothesis, after the book Dr. Seuss famously produced on a bet that he couldn't write one using fifty distinct words.
Writers have been exploiting this on purpose for a long time. The Oulipo group built a whole practice out of it — Georges Perec wrote La Disparition, a full-length novel, without ever using the letter e, and Gilbert Adair then translated the thing into English under the same handicap. Brian Eno and Peter Schmidt sold Oblique Strategies as a deck of cards precisely because a card you draw beats a suggestion you choose.
The mechanism is the same in all three cases: the constraint has to arrive from outside. If you pick your own animal, you will pick a wolf, because a wolf already matches the mood you're stuck inside. Press the button on a list of roughly 360 species you didn't assemble and you get a tapir, and now you have a problem you've never solved before.
The value is in the animals you'd never have chosen. If you find yourself rerolling until something "fits", you've quietly turned the generator back into a menu, and the whole effect evaporates. Take the third draw.
Worked example: the draw comes up pangolin
Say you pull a card and it's a pangolin. The instinct is to write a scaly guy. Don't. The move is to take each literal, factual trait, one at a time, and ask what that trait would mean if it belonged to a person, a culture, or a society.
Here's the real pangolin, and what each fact gives you:
| Pangolin fact | Literal reading | What it hands your character |
|---|---|---|
| The only mammal covered in keratin scales — the same protein as your own fingernails | Armour that is made of dead body parts | Someone whose protection is built entirely out of what they've already lost |
| Eats essentially nothing but ants and termites | Absurdly narrow diet, total dependence on one food | A specialist. Magnificent in one domain, helpless the moment it disappears |
| Nocturnal, secretive, near-solitary | Actively avoids being witnessed | Not shy — deliberately unobserved. A character with a reason to be unwitnessed |
| Rolls into a sealed ball under threat; the name is from Malay pengguling, roughly "one who rolls up" | Defence by total withdrawal | Their defining verb is close. Every conflict scene now has an obvious, character-true failure mode |
| That defence works on leopards. It does not work on a human being, who simply picks the ball up | The perfect defence against the wrong threat | The tragedy is pre-installed. You don't have to invent one |
| Over 895,000 trafficked between 2000 and 2019; commercial trade in wild-caught pangolins banned by CITES in 2016 | Hunted for the armour itself | The thing that keeps them alive is the reason they're hunted |
Read the right-hand column as a block and you have a character sheet: a hyper-specialist who armours themselves in their own losses, avoids being seen, responds to every crisis by shutting completely, and is targeted because of that armour — with a defence that is flawless against the danger they were built for and worthless against the one that actually showed up.
That is a protagonist. It took ten minutes and none of it came from you deciding to be interesting. It came from reading a real animal literally and refusing to reach for metaphor until the facts ran out.
The takeaway: don't ask what the animal symbolises. Ask what it does, then ask who does that. Symbolism gives you the wolf again. Behaviour gives you something nobody's written.
Four other ways to spend a draw
Build the culture, not the creature
Draw once, then design a fictional society around the animal's actual biology rather than its appearance. The starfish is the cleanest example in the pool, and nothing about it is what you'd guess: it has no brain and no blood, running instead on a water vascular system that drives hundreds of hydraulic tube feet by pressure alone. Many species feed by pushing their stomach out through the mouth, digesting the prey outside the body, then pulling the stomach back in. And a severed arm regrows — in some species the arm regrows the entire animal.
Don't design starfish-people. Design a society with no central authority, because there is nothing in the anatomy that could hold one: decisions emerge from pressure distributed across hundreds of small actors, and nobody can point at where a decision was made. A society that does its digesting in the open, on top of the thing being consumed, with no private processing of anything. A society where a severed piece becomes a whole person — so what does exile mean, when the exiled fragment simply becomes a complete member somewhere else? What does inheritance mean? That's three hours of genuinely novel worldbuilding out of one card.
Force a chimera for character design
Set the count to 3 and draw once. Then assign each animal a different design layer instead of blending them into mush:
- Silhouette from the first — say a wombat: low, blocky, absurdly wide, built like furniture.
- Surface from the second — a cuttlefish: skin that rewrites its own pattern in under a second.
- Movement and limbs from the third — a praying mantis: folded raptorial forearms, a triangular head that swivels to track you.
A squat, heavy, immovable body wearing skin that will not hold still, with two folded blades it hasn't used yet. That's a concept sheet. Blending the three into one furry blob is not — the layer assignment is what does the work, and it's why this beats "draw a cool monster."
One animal a day, twenty minutes, no negotiating
The oldest art-practice habit there is, and the reason it fails is decision fatigue: you open the sketchbook, spend eleven minutes choosing a subject, and by the time you've chosen, the will to draw is gone. Draw one animal, start the timer. Each card ships with a reference photo, so you're not falling down an image-search hole either. The point isn't the drawing — it's that on day nine you're rendering a barnacle, and rendering a barnacle teaches you things a ninth face does not.
Write the fable backwards
Aesop's animals aren't animals; the fox reasons like a person who wants grapes. The far more interesting version runs in reverse — start with a documented behaviour that has no moral attached, and find out what it argues for.
Take the pangolin's curl again. Optimal against every predator it evolved alongside, useless against a threat that arrived recently and can simply bend down. That's a fable about a defence perfectly tuned to a danger that no longer exists — and it's a much sharper piece than anything you'd have produced by opening a document and typing "a fable about pride."
FAQ
Does this only work for fantasy and sci-fi?
No, and it's arguably stronger outside them, because in a realist piece the animal never appears on the page. It's a generator for behaviour, not a casting call. Draw a great blue heron and you get a character who stands motionless in a conversation and strikes exactly once — that's a difficult uncle at a dinner party, no speculative element required. Crime and literary fiction take this fine. The only genre where it gets awkward is one where you were already going to write an animal, since then the draw competes with your plot instead of feeding it.
How do I stop the result feeling random instead of intentional?
Two rules. First, no rerolling — the second you start shopping for a good animal you're back to picking from your own taste. Second, run the fact-to-behaviour translation before you decide whether the draw is any good. Almost every "bad" animal is one you simply haven't read literally yet; a sea urchin sounds like a dud until you find out it chews with a five-toothed jaw mounted underneath its body. Forced-ness is a symptom of stopping at the animal's surface, not of the animal being wrong.
Is this useful for visual artists or only writers?
More useful for visual artists, if anything, because the reference photo is right there on the card and you can start drawing inside ten seconds. The layered-chimera exercise above is a character-design method first and a writing method second, and the daily draw is straightforwardly the most reliable sketchbook habit I know. Game designers land in between — an unfamiliar animal's real behaviour is a surprisingly good source of enemy movesets and creature AI.
What if the animal is one I know nothing about?
That's the best possible outcome. Fifteen minutes of reading about an Australian swiftlet will hand you more usable specifics than an hour of staring at a wolf you already think you understand. Unfamiliarity is the feature — the animals you can already picture are precisely the ones that produce the design you'd have made anyway.
None of this is about inspiration, which is not a thing you can schedule. It's about removing one decision — what is this about — so the remaining decisions have somewhere to push against. The forty minutes you spent deleting opening lines weren't a failure of imagination; they were forty minutes spent choosing, when you could have let a button choose and gone straight to the work.
Next time the page won't move, draw six animals, take the third one, and give it ten minutes of literal reading before you decide it's wrong.
Cover photo by Anete Lusina on Pexels.