How Reading Helps You Solve Problems Faster: The General, the Tumor and the Elephant

In 1980, two psychologists handed a group of students a medical problem that almost nobody solves. A patient has an inoperable tumor deep in the body. A ray strong enough to destroy it will destroy everything it passes through on the way. A ray weak enough to spare healthy tissue will not touch the tumor. What do you do? Roughly one student in ten found the answer. Then Mary Gick and Keith Holyoak tried something else. Before handing over the problem, they had a second group read a short story about a general and a fortress.

The general who could not march his army down one road

The story runs like this. A dictator rules a small country from a fortress at its center. Many roads lead to it, and all of them are mined. A large force moving down any single road would be destroyed. A rebel general wants the fortress, and he has the men to take it. He divides his army into small groups, sends each group down a different road, and times their arrival so they converge on the fortress together. Small enough on each road to pass safely. Large enough at the target to win.

The answer to the tumor problem is the same answer. Send several weak rays from different angles so they meet at the tumor and sum there, harmless everywhere along the way.

Students who read the fortress story before seeing the medical problem solved it at three times the rate of students who did not. Thirty percent instead of ten. And when the researchers added one sentence telling them the story might help, the solution rate rose to somewhere between seventy-five and ninety percent, depending on the version of the experiment.

Sit with the gap between thirty and seventy-five, because that gap is the whole subject of this article. The students in the middle group had the answer. They had read it minutes earlier. Most of them did not notice they were holding it.

Knowing something and reaching for it are two different skills

This is the finding that survived forty years of follow-up work, and it is less flattering than the usual account of reading. Reading does not hand you solutions. It stocks you with structures, and those structures sit quietly until something makes you connect one to the problem in front of you. Gick and Holyoak broke the process into three steps: noticing that a connection exists, mapping the parts of the old situation onto the new one, and generating the parallel answer.

Steps two and three are reliable. People are good at them. Step one is where almost everyone fails.

The reader's advantage, then, is not simply that they have read more. It is that a reader who has met a hundred structures has a hundred candidates competing for attention when a new problem arrives, and the odds that one surfaces rise accordingly. The two-story condition makes this concrete. Students who read two different stories built on the same underlying shape solved the problem at fifty-two percent with no hint at all, because two examples let them see the shape rather than the story. One example gives you an anecdote. Two give you a pattern.

A six-year-old, an elephant, and a boat

Around the year 200, Sun Quan sent Cao Cao an elephant as a gift. Cao Cao, then the most powerful man in northern China, wanted to know what it weighed. His officials had no scale that could hold it. One proposed building an enormous steelyard. Another proposed cutting the animal into pieces and weighing the pieces, which solves the measurement problem and creates a worse one.

Cao Cao's son Cao Chong, who was five or six, suggested walking the elephant onto a boat, marking the waterline on the hull, leading the elephant off, then loading stones into the boat until it settled to the same mark. Weigh the stones. You have the weight of the elephant, and the elephant is unharmed.

The anecdote appears in Chen Shou's Records of the Three Kingdoms, compiled in the third century, and Chinese schoolchildren still learn it. Historians have argued for a long time about whether it happened. The scholar Chen Yinke proposed that the tale traveled in from an earlier Buddhist story rather than originating at Cao Cao's court, and the debate has never fully closed. That argument matters less than it appears to. Whether the boy did it or a storyteller invented him, the story has been carried for eighteen centuries because of what it demonstrates: the substitution of one measurable quantity for another. In 2017 a team of Chinese researchers published a method for reconstructing a three-dimensional shape by lowering an object into water and tracking the displacement, and named the method after him.

What you already know outweighs how well you read

In 1987, Donna Recht and Lauren Leslie took over an empty classroom at Marquette University and built an eighteen-by-twenty-inch model of a baseball field, complete with four-inch wooden players. They brought in sixty-four seventh and eighth graders one at a time, sorted in advance along two lines: strong or weak readers, and high or low knowledge of baseball.

Each student read the same passage describing half an inning, then used the model to act out what had happened.

The weak readers who knew baseball outperformed the strong readers who did not. Reading ability, measured on its own, barely predicted who could reconstruct the inning. Knowledge did. Recht described what that knowledge was doing as scaffolding: a structure that lets a reader hold new information in place instead of processing it word by word and losing the shape of it.

The study is heavily cited and it has also been argued over, which is the reason to trust the wider picture rather than the single result. It has not been cleanly replicated. Dan Reynolds has questioned whether a test of baseball idioms measures what we mean by knowledge of baseball at all. Timothy Shanahan, who accepts the underlying principle as settled, still cautions against building a pedagogy on one study from 1988. Recht and Leslie themselves concluded something narrower than their popularizers do: that strategy instruction and knowledge-building deserve equal weight, not that strategies are useless.

The box that people cannot see as a shelf

A failure mode runs underneath all of this, and Karl Duncker named it in the 1930s. He gave people a candle, a book of matches and a box of thumbtacks, and asked them to fix the candle to the wall so no wax dripped on the table. The answer is to empty the box, tack it to the wall, and stand the candle in it. Most people cannot see it. They are stuck on the box as a container, because that is what it was doing when they were handed it. Duncker called this functional fixedness. Hand people the same tacks in a pile with the box beside them, and the solution rate roughly doubles.

Two later findings sharpen the point. Stanford undergraduates given the problem in writing solved it twenty-three percent of the time. Underline the nouns in the problem statement, so that the word box is pulled loose from the phrase box of tacks, and the rate rises to around half. And in 2000, Tim German and Margaret Defeyter found that six- and seven-year-olds show functional fixedness while five-year-olds largely do not. The younger children have not yet learned firmly enough what objects are for.

That is the cost of expertise, and it is why Cao Cao's ministers lost to a child. Everything you know well becomes a groove. Reading widely is one of the few habits that keeps cutting new grooves after the old ones have set.

What this means for the reading you actually do

Four things follow, and none of them require you to read about problem-solving.

Range does more work than volume. Two structurally similar stories from different worlds taught the pattern better than one story told twice. A biography, a folk tale and a piece of popular science will furnish more usable structures than three novels of the same kind.

Fiction is not the soft option. The fortress was a story. The elephant was a story. Narrative is how the mind stores relationships between parts, which is exactly the form an analogy needs in order to travel.

The step to practice is noticing. Faced with a hard problem, the useful question is not what should I do, but what does this remind me of. That single question is what the researchers' hint supplied, and it was worth forty-five percentage points.

Reading with a child builds their scaffolding, not only their skill. The baseball study points at something parents rarely hear: comprehension is not a general-purpose muscle you can strengthen in isolation. Every subject a child reads about becomes scaffolding for every future text that brushes against it. A child who has read about volcanoes, medieval kitchens and migrating birds will read better than their reading level predicts.

How this applies at your kitchen table

When a child comes to you stuck, on homework, on a friendship, on how to fit everything into a school bag, the reflex is to give the answer or to give a strategy. The research suggests a third move: ask what it reminds them of. If they read something last week with the same shape, they are holding the answer and do not know it, exactly as those undergraduates did not know it.

Adults are no different. The general and the fortress sat in those students' memory, minutes old, and thirty percent reached for it. The rest needed one sentence of permission to go and look.