Iceberg Principle

The more the writer knows and chooses not to use, the richer the story feels. This is counterintuitive. Writers often feel that more information on the page means more texture, more depth, more world. In practice, it usually means congestion. The prose that feels thick with life is often the prose that has been stripped — where every visible detail is load-bearing because the writer knows what’s beneath it.

Readers don’t need the appendices. They need to feel, through the prose, that the appendices could exist. Confidence in the prose comes from knowledge behind it. The difference between a detail that lands and one that doesn’t is often not the detail itself but whether it’s supported by a structure of knowledge the writer has and isn’t showing.

Here’s what this looks like in practice. Two writers describe the same fictional neighborhood — the kind of block where two-flats sit next to shuttered storefronts and a dollar store has replaced the butcher that was there since 1957. Writer A has driven through neighborhoods like this, looked up the demographic statistics, read one news article. Writer B grew up in such a neighborhood, or spent six months interviewing residents, or has read everything written about how capital flight and discriminatory lending produced the specific built environment. Writer A’s description will be accurate and dead. Writer B’s will be accurate and alive. The difference is not the words. It’s the ten-to-one ratio of what B knows to what B shows.

A writer building a fictional city doesn’t describe the entire city. They describe what the viewpoint character sees, filtered through what that character notices, at the moment the character is there. But the writer knows the entire city — its history, its economy, its neighborhoods and their relationships, the texture of daily life in different parts of it.

This knowledge shapes the details that do appear. A character walking through a wealthy neighborhood and noticing a specific architectural feature feels different if the writer knows the economic history that produced that architecture. The character doesn’t explain the economic history. The writer doesn’t share it. But it’s there, supporting the detail from below.

Raymond Chandler knew Los Angeles — not just its geography but its class structure, its corruption, its racial geography, its specific quality of sunlight and money and disappointment. Marlowe’s Los Angeles is a felt place because Chandler put an entire city’s worth of knowledge behind every street name and eucalyptus hedge and oil-money house set back from the road. The details are selective. The selectivity works because of what’s behind it.

Tolkien is the deliberate anti-iceberg: his appendices are published, his mythology is visible. But this exception proves the rule. Tolkien is specifically interested in the mythology as mythology — in The Lord of the Rings, the world-building is not concealed support for a character story. The epic scale and the sense of deep history are the point. The published appendices are not evidence that visibility works generally; they’re evidence that Tolkien’s particular project required it. For writers whose world-building is support rather than subject, the iceberg position is the correct one.

The same principle governs character development. A writer creating a character who appears in three scenes should know that character’s complete history — the failed relationships, the childhood fears, the private embarrassments, the compromises they’ve made and justified, the things they’ve never told anyone. Almost none of this appears directly. But the character who acts in those three scenes acts the way a person with that specific history would act, and the reader feels the weight of that history without being told it.

When a minor character in a skilled novel feels three-dimensional, this is why. The writer knows more about them than the scene requires. That excess knowledge shapes what the writer chooses to show — the particular way this character holds themselves, the thing they don’t quite say, the hesitation before a simple yes.

This is the deeper function of Backstory: not material to be dropped into the narrative as exposition, but knowledge the writer holds that shapes every sentence written about that character. The backstory that appears on the page is almost never the useful version. The backstory that lives in the writer’s notes and never reaches the reader at all is the one doing the structural work.

Interiority is where the iceberg principle meets character most directly. In close point of view, the character’s interior is partially visible — but only partially. The writer knows the full interior, including the things the character won’t admit to themselves, the contradictions between what they think they believe and what their actions reveal. The reader gets the surface of that interior. What the writer knows but withholds gives the visible surface its texture.

The iceberg principle has a practical implication for research: always research more than you’ll use. Interview a cardiologist for a scene where a character has a heart attack. Use three details. The remaining knowledge makes those three details accurate — accurate in a specific way that a writer who looked up "heart attack symptoms" on a website wouldn’t achieve. Authenticity is felt by readers even when they have no expertise to evaluate it against.

This is not mysticism. It’s the result of how expertise changes perception. The writer who knows cardiology in depth will select different details than the writer who did a surface search. The expert-informed details have an internal coherence — they fit together the way real things fit together — that surface-research details lack. The reader who has never had a heart attack can still feel the difference between a scene written with authentic medical knowledge and one written from a symptom list.

Research for Fiction covers the practical mechanics. The iceberg principle is the argument for why research should be deep rather than sufficient.

Hemingway’s warning about hollow places is a diagnostic tool. When prose feels thin — when a reader says "I wasn’t convinced" or "it didn’t feel real" — the cause is often that the writer omitted things because they didn’t know them, not because they chose to leave them out. The hollow place is the absence of the knowledge that should be supporting the visible detail.

The symptoms are recognizable. Details that are technically accurate but don’t cohere with each other. World-building elements that appear and disappear without consistency. Characters who behave differently in different scenes without motivation. Settings that feel like sets — correct in their surface features but without the quality of being inhabited. Each of these is a hollow place: a visible surface unsupported from below.

The fix is not to add more prose. It’s to do more work beneath the surface. Add description and you get more description. Develop genuine knowledge of what’s beneath the surface and every sentence you write changes.

This is why the iceberg principle is not a prescription for minimalism. Hemingway wrote with extreme compression. Dickens did not. Tolstoy did not. Le Guin did not. The principle is not about how much to put on the page. It’s about the ratio of what you know to what you show — and the requirement that the ratio never fall below approximately eight to one.

Show Don’t Tell and the iceberg principle are not the same principle, but they emerge from the same understanding. The writer who tells — who narrates directly — is often the writer who doesn’t trust their details to carry meaning. Trust in details requires knowledge behind the details. The writer who shows is trusting that a well-chosen specific detail will communicate more than an explanation, because the detail is supported from below.

Showing fails when the writer doesn’t have enough knowledge to select the right details. In that case, telling becomes a compensatory move — an attempt to deliver meaning directly because the selected specifics aren’t carrying it. The iceberg principle predicts this: a thin iceberg produces showing that doesn’t work, because the visible portion isn’t supported by enough beneath. The solution is not to tell instead of show. The solution is to build the structure beneath the surface until the visible portion can carry weight.