Combinatorial Creativity: Why Wide Reading Produces Better Ideas
Most creative output is recombination. The ideas, frameworks, and images that go into a "new" creation almost always existed somewhere before. What's new is the combination, the application, or the connection between things that hadn't been connected yet.
Cognitive scientist Margaret Boden, in The Creative Mind (1990), divides creativity into three types: combinational (recombining familiar ideas in unfamiliar ways), exploratory (working within an existing conceptual space), and transformational (changing the space itself). The first is the most common and the easiest to develop deliberately. Arthur Koestler had described the same mechanism earlier in The Act of Creation (1964), calling it "bisociation": the moment two unrelated frames of reference intersect in a single mind.
The implication is straightforward. If creativity is recombination, the inputs matter. The breadth and depth of what you've read, watched, and thought about sets the ceiling on what you can recombine.
What the Research Shows
Studies of expert creative output consistently find that high performers draw from broader and more varied source material than their peers.
Dean Keith Simonton at UC Davis has spent four decades studying the careers of prominent scientists, artists, and writers. One pattern shows up repeatedly: the most creative figures demonstrate wider reading habits across disciplines, not just deeper expertise within their own. The novel insights tend to come from people who had recently been engaging with material outside their primary field.
Howard Gruber's case studies of Darwin reach a similar conclusion. Darwin's notebooks during the years preceding On the Origin of Species show him reading across geology, animal husbandry, economics (Malthus on population), and travel writing. Natural selection emerged from connections across those domains, not from work confined to any one of them.
This is the rationale behind Charlie Munger's well-known argument for a "latticework of mental models" drawn from many disciplines. The latticework gives you more frames to apply, and more potential intersections between frames. A single field, however deeply mastered, gives you one toolkit. A dozen fields, even shallowly held, give you dozens.
The Retention Problem
Reading widely is necessary but not sufficient. Material you've forgotten can't be recombined.
Hermann Ebbinghaus published the first studies of forgetting in 1885. The core finding has held up across more than a century of replication: without reinforcement, we lose roughly half of newly acquired information within a day, and the loss continues for weeks before flattening. Most of what you read disappears.
Modern memory research has added an important wrinkle: the technique most readers rely on to fight this — re-reading the text — produces almost no improvement in long-term retention. Henry Roediger and Jeffrey Karpicke at Washington University demonstrated in a 2006 study that students who studied a passage once and then took a recall test outperformed students who re-read the same passage four times, even when the recall test offered no feedback. The act of retrieval itself is what consolidates memory.
For creative thinkers this matters specifically because forgotten material isn't available for combination. You can't bisociate two ideas if you can only access one of them. The breadth of your reading determines your potential for creative connection. The depth of your retention determines how much of that potential you can actually use.
Why Non-Fiction Is Particularly Hard
Non-fiction compounds the problem. A novel rewards immersion, and you remember the plot because the story has its own structure that does memory work for you. A book of philosophy, history, or science is dense with conceptual material that has no narrative scaffolding. Most non-fiction readers finish a book and find that within a few weeks they retain a vague impression and a few quotable lines — not the actual ideas they came to absorb.
The research on what does work is fairly clear:
Active engagement during reading. Highlighting alone is mostly performative; the act of selecting passages doesn't reliably improve retention. Writing margin notes that paraphrase or question the argument does. So does pausing periodically to summarize what you just read in your own words.
Retrieval at spaced intervals. Once a chapter is finished, the question isn't whether you understood it. The question is whether you can reconstruct its central claims a week later, then a month later, then six months later. Spaced retrieval — the practice formalized in tools like Anki and earlier by Sebastian Leitner's flashcard system in 1972 — is the most consistently supported intervention in the memory literature.
Discussion or teaching. Explaining material to someone else forces you to organize it in a way that passive reading does not. The "protégé effect" has been replicated across dozens of studies: people who learn material with the expectation of teaching it later retain more than people who learn it for a test.
Chapterly is one of the recent tools built specifically for serious non-fiction readers around these principles. It uses AI tutoring after each chapter, spaced retrieval drawn from your own highlights, and a deliberate focus on retention rather than completion. The premise is the one the research supports: finishing a book doesn't mean learning from it.
How This Connects to Creative Practice
Once retained material is actually accessible, the work of creative recombination becomes possible. A few practical implications follow.
Reading across domains beats reading deep into one. Analogical reasoning — the capacity to see structural similarity between distant subjects — is one of the most heavily studied predictors of creative insight. Dedre Gentner's research at Northwestern shows that people draw stronger analogies when they have material from multiple domains stored in retrievable form. A historian who has also read systems biology has structural patterns available that a pure historian does not.
Notes that you actually revisit produce compounding returns. A commonplace book, a Zettelkasten, or any system that brings old reading back into contact with current thinking changes the math of retention. Niklas Luhmann, the German sociologist who attributed his unusual productivity to his slip-box system, claimed his ideas mostly came from collisions between notes he had written years apart.
Generative exercises depend on stored material. Idea generation techniques — SCAMPER, random word association, divergent thinking prompts — all assume you have something in memory to work with. The exercises don't manufacture inputs. They recombine what's already there. Readers who retain little have less to work with, regardless of how good their technique is.
The relationship between reading and creativity is not the romantic version where sufficient quantity of books somehow produces inspiration. It's mechanical. You read to acquire material. You retain to keep that material available. You recombine to create. Each step compounds the next, and skipping the middle one is why most people who consider themselves widely read still feel stuck for ideas.
Ready to train your creativity? Try science-backed exercises that measure and improve your creative thinking. Start a Free Exercise