Back to the table

33 Beginner Wb Worldbuilding Worldbuilding

Worldbuilding

Designing the world your story happens in.

What it is

Worldbuilding is deciding what is true in the world your story happens in. You do it before you draft, and you keep doing it while you draft. The ground it covers is wider than most writers expect.

The scale can be anything. A galaxy with rival empires needs worldbuilding. So does a village of four hundred people, where everyone knows who owns the bakery, two families haven't spoken since 1998, and the bus timetable does its own thing on Sundays. The job is the same at both sizes. You decide what exists, what the rules are, who holds power, and what the people who live there take for granted without ever thinking about it.

Every genre needs this, not only fantasy and science fiction. A crime novel set in a real city and a romance set in a small town still have to settle what is possible, who holds power, what the locals assume without a second thought, and what happens to someone who breaks a rule. In those books the answers come from research instead of invention. They still have to be settled.

Worldbuilding and setting are different things. Setting is where and when one story takes place: this house, this street, this year. Worldbuilding is the set of decisions underneath that make the house and the street behave the same way every time you go back to them. Those same decisions would still hold if a sequel moved three hundred miles away. Put simply, setting is what the reader sees, and worldbuilding is what you know.

Worldbuilding is also the front door to eight more specialist pages. Setting covers place and time. Lore covers the myths and secrets a world tells about itself. Magic system covers rules and costs, and technology does the same job in science fiction. Factions covers organized groups competing for power. History covers the timeline behind today's grudges. Culture covers how a people live and what they believe from day to day. Maps covers geography that doesn't cheat. One rule runs through all eight, and it starts here: build what the story needs, and only hint at the rest. Orson Scott Card makes the same case in the first section of How to Write Science Fiction & Fantasy, listed below. He tells you to work outward from the story you mean to tell, instead of inward from a finished reference file.

That rule splits the job into two halves.

The first half is deciding enough. Your main character is going to try, fail and try again, and every attempt happens inside the rules of the world. If you haven't decided a rule, you'll end up making it up to suit each scene. Readers feel that looseness even when they can't say what's wrong. So settle four things early: what people are allowed to do, what they are forbidden to do, who enforces it, and what happens to someone who defies it. Decide enough that you never have to loosen a rule to make a scene work.

The second half is keeping nearly all of it off the page. Most writers skip this half, and most of the skill lives in it.

J.R.R. Tolkien gave the idea its classic statement. He gave a lecture at the University of St. Andrews in March 1939, and the essay version, "On Fairy-Stories", first appeared in print in 1947. In it he called the writer's invented world a Secondary World, and he called the act of making one sub-creation. What he asked of that world was "inner consistency of reality". He did not ask for completeness. A world can be unfinished, as long as it behaves as if it were finished.

Ernest Hemingway gave the second half its picture in Death in the Afternoon (1932): the iceberg. The part under the water gives weight to the part you can see. Worlds work the same way, with one twist. Readers can tell when there's nothing underneath. Material you've worked out gives off confidence even if you never mention it. Material you haven't worked out comes across as fuzzy, and fuzzy reads as flimsy. Backstory applies the same principle to characters.

So what does "hinting" look like on the page? You are exact about small things and quiet about big ones. Mystery for its own sake has nothing to do with it.

Picture a grumbling local who complains, in passing, about the new route of the Bridge District streetcar line since the flood. Nothing gets explained. The reader learns nothing about the streetcar network, the districts or the flood. Yet that one remark tells the reader three things. The city has a transit system. Some authority can change it. And there was recent bad weather, with an aftermath people are still living with.

Hinting comes down to five techniques you can use again and again:

Lincoln Michel gave the two halves a sharper pair of names in Electric Literature in April 2017, in the essay listed below. Worldbuilding, in his sense, is a complete and logical account of a place. Worldconjuring creates the feel of a place from a few well-chosen signals and lets the reader fill in the rest. His view is that every story needs conjuring, while full building pays off only for a minority of books. Most novels use some of each, and the genre decides the mix. The case against worldbuilding gets its full hearing further down.

Contemporary and historical fiction count too. A crime novel set in present-day Manchester, a romance in a small town and a novel set in 1911 all involve worldbuilding. The decisions come from reality instead of your imagination, but you still have to make them and keep them consistent. Informed readers police those rules harder than fantasy readers do, because they can check. How a warrant gets issued, how a nurse's shift pattern shapes her week, the real driving time from one place to another: these are all worldbuilding facts. Research takes the place of invention. Mistakes cost you more embarrassment, too, because someone will write in.

Why it matters

Constraints make plot. When writers get stuck, they tend to ask what else their world could hold. The better question is what the world makes impossible. Each impossibility is a wall, and plot is what happens when characters have to find a way around walls. Give your world instant travel to anywhere and you remove the cost of distance, along with a whole class of obstacle. Make healing free and injuries stop meaning anything. A world where everything is allowed gives you a story where nothing is hard, and a story where nothing is hard is a book without conflict. Timothy Hickson's On Writing and Worldbuilding: Volume I, listed below, is strongest right here, where it follows a world's rules into the events of the plot. The value of a world lies in what it forbids. A long inventory of what it contains is worth far less.

Readers forgive strangeness, but they don't forgive exceptions. Consistency is the thing readers police hardest. They will go along with almost any oddity, and they won't forgive a rule that breaks. Suppose the dead can be contacted, but only on the third night after burial, every single time, and it always costs something. Readers believe that. Now suppose a world just like this one except for a single rule, and in chapter twenty-nine that rule gives way because the plot needs it to. Readers stop trusting the book, and the whole experience of reading depends on trust. An odd world that keeps its rules beats a plausible one that bends them.

A good world can be spent again and again. Storycosm's whole approach rests on this, and the reason is as much commercial as artistic. A sound world can carry a series, then a second series, then standalones, then crossovers. You pay for the construction once and earn on it over years of publishing. A world used that way is what a storycosm means. Once a world outlives a single book, your memory stops being enough, and the facts need a filing system. That filing system is the continuity bible.

Sometimes the world is what readers came for. Epic fantasy readers buy books with appendices, glossaries and maps on purpose. In LitRPG, a detailed explanation of how the game mechanics work is part of the fun. Space opera sells scale. In cozy fantasy, the feel of one small place is what readers are paying for. For these readers, an explanation of the system is the thing they bought, and it never feels like an interruption. So know which kind of book you're writing. The answer comes from your reader promise, and craft purity can't settle it for you. A rigid system dropped into literary fiction and a fuzzy system in a LitRPG book are the same mistake made in two directions.

The case against worldbuilding has a point. In January 2007, the novelist M. John Harrison called worldbuilding "the great clomping foot of nerdism" in a post on his blog, Uncle Zip's Window. His charges were that it's unnecessary, that it tries to cover every inch of a territory that doesn't exist, and that it puts the writer in control where writer and reader ought to be working together. The post started a lasting argument in science fiction and fantasy, and scholars have written about it since. Harrison is right about one way worlds fail. Some are so completely fixed that the reader has nothing left to do, and the novel turns into a guided walk through a model village. But he is describing a hazard, and the craft is bigger than its hazards. A working genre novelist can take note of the danger and carry on. The danger is real, and several of the common mistakes further down grow out of it.

Real-world examples

Dune (1965): three fixed rules that drive the plot

Frank Herbert's Dune was published by Chilton Books in August 1965, after a serial run in Analog that began in December 1963. People remember it for its abundance: the ecology, the religion, the politics of ruling families, the invented vocabulary. Underneath all of that sits a short list of three fixed rules.

The first rule is that there are no thinking machines. In the world's history, a war called the Butlerian Jihad destroyed computers and robots, and the ban still stands. So the world has Mentats, humans trained to do the work computers would have done. Information becomes scarce, personal and open to corruption, and plots can turn on who knows what.

The second rule is personal shields, which make beam weapons pointless. A shield stops anything moving faster than a set speed. When a lasgun beam hits a shield, the explosion is as likely to kill the shooter as the target. So armies fight up close with blades, and the big final scenes are knife fights instead of artillery battles. One chosen limit produced one particular kind of scene.

The third rule concerns spice. Spice exists on one planet only and can't be made anywhere else, and interstellar travel depends on it. That gives the whole galactic economy a single physical chokepoint. Whoever holds that one desert planet holds everything, and the fight over it is the plot.

Herbert didn't start with a system. In 1957 he went to Florence, Oregon, on a magazine assignment to cover a US Department of Agriculture project that planted beach grass to hold the sand back from roads and buildings. He looked at the dunes from the air, took photographs and walked them. He wrote the piece up as "They Stopped the Moving Sands", and it was never published. What he took away was a theme: people stepping into a hostile system and getting consequences they never asked for. The Siuslaw Public Library holds the archive of his Dune research. The world grew from a question, and the rules grew from what the question needed.

Tolkien's legendarium (around 1914 to 1973): the world that came first

J.R.R. Tolkien is the best advocate the world-first side has, and he is also the clearest warning. He said himself that his world wasn't built to serve a story. Letter 165 in his published Letters is dated June 30, 1955, and went to Houghton Mifflin, his American publisher. In it he explained that the invented languages were the foundation of the whole project. The stories were made to give the languages a world to live in. The name came first and the story after. Coming from the finest practitioner the craft has had, that settles one question. Putting the world first can be the main creative act, and the blanket claim that world-first is always a mistake is wrong.

The cost was large. Tolkien started his legendarium, a whole invented mythology and history, around 1914, and he never finished it. The Hobbit came out in 1937 and sold well. He then offered his publishers the material that became The Silmarillion. They turned it down and asked for another hobbit book, and that request is the only reason The Lord of the Rings (1954 to 1955) exists. Afterward he went back to the legendarium. In the late 1950s he worked on its philosophy instead of its stories, and when he died in 1973 it was still unpublished. His son, Christopher Tolkien, assembled and edited The Silmarillion for publication in 1977, and he wrote some linking passages of his own to bridge gaps his father had left open.

So the most elaborate invented world of its era produced two finished novels in its maker's lifetime, and the second one exists because a publisher asked for it. None of this argues against depth. It shows that a world can swallow a career, and that books tend to get finished when an editor is waiting for them.

The Time-Turner in Harry Potter (1999 to 2003): a power that had to be taken back

J.K. Rowling introduced the Time-Turner, a device for going back in time, in Harry Potter and the Prisoner of Azkaban (1999). It gave that book a tidy payoff at its climax. She then spent the next two books containing it. Writing on the official Harry Potter site, she says that in that book she went "far too light-heartedly into the subject of time travel." She also says she doesn't regret the Time-Turner, but that it opened up a great many problems. If wizards can rewind their troubles, later plots have no ground to stand on. So she made repairs. She added a rule limiting how far back anyone can safely go. Hermione gives hers back. And in Harry Potter and the Order of the Phoenix (2003), every remaining Time-Turner is smashed during the battle in the Department of Mysteries, which takes the power out of the world for good.

The Time-Turner is one of the most instructive worldbuilding stories in modern popular fiction, and its lesson has nothing to do with failure. Readers police consistency, and plausibility matters to them much less. Once a capability exists in your world, it exists in every later scene, whether you think about it or not, and readers will ask why nobody used it. You have three honest options. You can live with what it does to your story. You can price it so high that it's rarely worth using. Or you can write it out of the world where readers can watch it go. Quietly forgetting it isn't one of the options, yet that's what most drafts do.

Here's a quick test. For every new ability in your world, note what it would fix if someone used it at the climax. If the answer is "everything", you have a Time-Turner, and you need to deal with it right away.

The Pillars of the Earth (1989) and Elmore Leonard's researcher: worldbuilding by research

Ken Follett writes thrillers and historical novels, not fantasy. The Pillars of the Earth (1989) runs to 400,000 words, and its plot is the building of a cathedral in twelfth-century England. Follett's account on his own site says the interest came before the book. He read about medieval architecture, then visited cathedral towns, Lincoln and Winchester among them, and spent days inside the buildings working out how they were put up. The novel took him three years and three months to write.

The lesson carries over to any kind of book. According to Follett's account on his own site, his writer friends saw right away that the building of the cathedral was the backbone of the story. A cathedral took thirty years or more to build, long enough to cover the adult lives of the characters. The central fact of the setting turned out to be the frame of the plot. Follett also treats the cathedral as an active force. Its existence drives the plot instead of just giving the plot a place to happen. Ask the same of any setting you've researched. Does it make things happen, or does it only decorate?

Elmore Leonard, the crime novelist, went further and kept a full-time researcher. Gregg Sutter worked for him from January 1981 to the end of Leonard's life. Sutter's job was to go and look. He embedded with detectives in Atlantic City, did research in Cuba, and chased down specific, checkable, slightly ridiculous details. Those details give Leonard's novels their grounded feel. You don't need a researcher to learn from this. The texture readers praise as "effortless" in contemporary crime is nearly always somebody's legwork.

Terry Pratchett's Discworld (1983 to 2015): one world across forty-one novels

Terry Pratchett wrote forty-one Discworld novels, from The Colour of Magic (1983) to The Shepherd's Crown (2015), which came out after his death. He reused one world across a whole career. If you want to build a storycosm instead of a single book, Discworld is your case study, and several of its features are worth copying.

First, the Disc is organized around institutions instead of one plot, so it can hold very different kinds of story. The Watch books are police procedurals. The witches books are rural folklore. The Moist von Lipwig books are about institutions being founded. The reading-order guides on Pratchett's own site give a list built around characters alongside the publication order, and they say the books can be read in any order. Mystery fans can come in through the City Watch, and fantasy fans through the witches. The world works because it has several separate doors into it. A world designed around one conflict tells one conflict. A world built from places, jobs and institutions can support hundreds of stories.

Second, the world modernizes over time. A police force, a newspaper, a postal service, paper money and a railway all arrive on the Disc, and each arrival is the subject of a novel. Pratchett raised the world's modernity where another writer would have raised the menace, and his characters argue about what each change costs.

Third, the world needed a reference book. The Discworld Companion, compiled with Stephen Briggs, came out in 1994, after more than a dozen novels, and was revised after that. Long-lived worlds outgrow anyone's memory. Keeping track is the job of a continuity bible, and the time to start one is book two, not book eleven.

An everyday scenario: the 60,000-word world file

Here is a made-up case, though a common one. A writer decides in September to write a fantasy novel. By March she has a 60,000-word world file. It holds a calendar with named months and a leap-year rule, three currencies with exchange rates, nineteen noble houses with sigils, a constructed language of about four hundred words with its own consonant-shift rule, two schisms in the state religion, and a timeline that goes back nine hundred years. The work is good. Two people on a forum have called it the best-built world they've seen there.

The novel stands at 1,400 words. She has started three different first chapters and abandoned all of them.

She isn't lazy, and calling it procrastination misses something. The work is real, and some of it will end up in the book. The trouble is that the world file gives her a feedback loop the novel doesn't. Every hour on the file produces something visible she can finish: one more house, one more god, one more decade. Nothing in it can be wrong yet, because nothing in it is committed. An hour on the novel produces pages she isn't sure about. Offered certain progress or uncertain progress, everyone picks certain.

Here's the sign that tells preparation from avoidance. Preparation settles what the next three chapters will need. Avoidance settles what no chapter has asked for yet. Her leap-year rule is avoidance. Her currency conversions count as preparation only if a smuggling plot turns on them. Otherwise they're avoidance too.

Her fix is a deadline and a cut in scope. She caps the world file at one page and keeps only the rules that chapters one to five use. From then on, anything a scene needs mid-draft gets invented on the spot, put in square brackets, and settled in revision. She finishes the first draft in five months. About a fifth of the world file turns out to be true of the finished book, and her second novel goes faster.

Try it yourself

You don't have to answer every question about your world before you start writing. Patricia C. Wrede's well-known list of worldbuilding questions, listed below, says at its start that you don't need to answer all of them, or any of them, before you draft. The eight steps here decide just enough, and the template after the last step gathers your answers onto one sheet.

  1. Write a one-page world sheet before anything else. One page is a hard limit, so nothing spills onto a second. On it, put where and when the story happens, the three rules of this world that differ from what a reader expects by default, the power structure and how it keeps itself in place, what ordinary people fear, and what is impossible. If it won't fit, you're writing an encyclopedia. Cut it back until it's a sheet.
  2. Name three impossibilities. Finish the sentence "In this world, nobody can ______" three times. Under each one, write the scene it makes possible. A constraint that produces no scene hasn't earned its place. If your middle sags, try this exercise before any other.
  3. Run the Time-Turner test on every capability. List what your world allows that the real world doesn't: magic, a technology, an institution, a law, a talent. For each one, write if this were used on my climax, what would happen? If the answer is "it would solve it", the capability needs a cost, a limit or a removal. Choose one now and make it visible on the page.
  4. Do a street-level pass. Pick one scene and add three details only someone who lives there would know: a price, what a certain smell means, an idiom, a piece of red tape, a shortcut everyone uses. When a world feels thin, look first for missing ground-level texture, and only then at thin cosmology, meaning the big picture of gods, creation and how the universe works. Sensory detail is where that texture comes from.
  5. Keep a consistency ledger while you draft. Keep one running list in whatever tool you draft in. Every time you state a rule, a distance, a price, a date or a name, log it with its chapter. Don't stop in the middle of a scene to check anything, and do all the checking in revision. By book two, the ledger grows into your continuity bible.
  6. Set a hint budget for chapter one. Decide how many unexplained proper nouns chapter one will contain, then count them. Five is generous, and ten is a lot. Make sure none of them gets explained in that chapter. Explaining one later, once the reader wants to know, is exposition doing its job. Explaining it at once is an info dump.
  7. Timebox the building. Give yourself a fixed period for building, a week or two for example. Then write your first draft, building only what the scenes demand as you go. When a scene needs something new, make it up, put it in brackets, and tidy it later. This one habit may do more than any other to separate writers who own a world from writers who finish a book.
  8. Check the breadth before you commit. Ask whether this world could carry five more books, and more than one kind of story. Does it have locations, institutions and trades you haven't used yet, or only a single conflict? If it has only one, widen it now. Breadth is what separates the setting of one novel from a storycosm.

World: (place, time, one line) Three rules that differ from ours: ______ / ______ / ______ Impossible here: ______, which makes this scene possible: ______ Who holds power, and how: ______ What ordinary people fear: ______ Capabilities that could solve my climax, and their cost: ______ Unexplained proper nouns allowed in chapter one: ___ (counted: __) Building stops on: (date)

Common mistakes and misunderstandings

Opening with a lore dump. The first chapter explains the history, the magic, the politics and the map, and the story starts on page nineteen. Science fiction workshops have names for this. The Turkey City Lexicon, a workshop primer for science fiction writers edited by Lewis Shiner, revised by Bruce Sterling and hosted by SFWA, calls it the info-dump. Its dialogue version is "As You Know, Bob": characters telling each other things they both already know, purely for the reader's benefit. Underneath sits a wrong assumption, that readers have to understand before they can be interested. The reverse is true. Interest buys the patience that understanding needs. Give readers a character to care about first, and the explanation after. Exposition covers how to deliver it.

A world with no texture at street level. The cosmology is grand, but the scenes happen in a bare room between people who could be swapped for one another. The Turkey City Lexicon calls this White Room Syndrome and treats it as a failure of imagination, not of planning. It shows up most at the start of a book, while the writer is still finding their feet. To fix it, leave the worldbuilding alone and add one concrete, physical, local detail to every page. Readers believe in a place because of the details a visitor on foot would notice.

Rules that bend when the plot wants them to. Of all the mistakes across the worldbuilding pages, this one probably does the most damage to trust, and it usually happens by accident. A cost set up in chapter four gets waived in chapter thirty-one because the ending needed it. A journey that took eight days now takes two. A character does something the rules forbade, and nobody remarks on it. Readers may not catch the breach, but they feel the softening, and it comes back to you as "the ending didn't feel earned". Your defenses are the consistency ledger and one decision made before you write the ending: the ending has to be reachable under the rules as you stated them.

Worldbuilder's disease. Call it that when the building never stops and the drafting never starts. Use the test from the scenario above. Does this piece of work settle a question the next three chapters will ask? If yes, it's preparation. If no, it's avoidance. Other signs are building things only you will ever see, and feeling relief when a work session grows the world document instead of the manuscript.

Building a world before you know whose story it is. A world isn't a story. "A city where nobody can lie" is a fine situation, but it holds no novel until someone who lives there wants something the ban denies them. This mistake hits hardest for writers who came into fiction through the worldbuilding door. The world feels finished, and the premise doesn't exist yet. Here's a warning sign: you can talk about your world for ten minutes and about your main character for thirty seconds. Fix that imbalance first.

Mistaking a data table for a culture. Nineteen houses, each with a sigil and a motto, make an inventory. A culture is what people do without thinking, what disgusts them, what they find funny and what they'd be ashamed to admit. Culture and factions go wrong the same way when they're named and catalogued but never lived in. One family arguing at the dinner table does more than a list of twelve houses.

Thinking contemporary and historical fiction are exempt. They aren't, and they're harder, because readers can check. The failure just looks different. Instead of broken impossibilities, you get generic blankness: a town that could be any town, a job described by someone who has clearly never done it. For realist writers, research is the worldbuilding. Research is also a handy place to hide from drafting, so the same time limit applies.

Inventing words for the sake of it. Every invented word costs the reader some working memory. A page can carry about three. Twelve is too many. Spend invented words on things that have no English equivalent and that matter to the plot, and use ordinary words everywhere else. Tolkien is the exception, because the languages were his goal. In fiction driven by plot, coined words slow the reading down.

Explaining a mystery that was working. The unexplained streetcar reroute works. Nine pages on flood hydrology would not make it better. When an atmospheric detail is working, you'll be tempted to turn it into information, and giving in usually makes it worse. Build it and understand it, then leave it under the water.

Related elements