What Creativity Actually Is

The Problem with How We Talk About Creativity

The popular framing of creativity treats it as a personality attribute possessed in varying degrees — you are or aren’t creative, and the creative ones among us produce art, inventions, and ideas while the rest execute and manage. This framing is empirically wrong and practically damaging. It positions creativity as fixed and innate rather than as a process that can be understood, cultivated, and improved.

The cognitive science of creativity doesn’t support the attribute view. Creativity is a process — specifically, a process of combining existing mental representations in ways that produce something new and useful. The mental representations (concepts, experiences, knowledge structures, sensory memories, procedural skills) are the raw material. The combination process is what creativity refers to. Both the material and the process can be improved by deliberate work.

This is the combinatorial theory of creativity, associated most clearly with Margaret Boden’s work and with Steven Johnson’s Where Good Ideas Come From. It is not the only framework — there are genetic/evolutionary theories, social/network theories, and domain-specificity arguments — but it is the most empirically grounded and practically generative.

Divergent and Convergent Thinking

J.P. Guilford’s 1950 presidential address to the American Psychological Association introduced divergent thinking as the cognitive counterpart of convergent thinking and argued that standard intelligence tests — which measure convergent thinking almost exclusively — were missing a critical component of human cognitive ability.

Convergent thinking moves toward a single correct answer. A math problem has a solution; a logic puzzle has a solution; a grammar question has a correct answer. IQ tests and academic examinations measure convergent thinking heavily. It is the cognitive mode that education systems optimize for.

Divergent thinking generates multiple possible answers, associations, or uses from a single starting point. The Alternative Uses Task (Guilford’s measure): how many uses can you think of for a brick? A convergent answer: building a wall. Divergent answers include building walls, breaking windows, propping doors, scratching surfaces, as a paperweight, as a weapon, as a musical instrument (scraped), as a heat sink, as a marker — the measure scores fluency (number of uses), originality (statistical rarity of uses), flexibility (number of distinct categories of use), and elaboration (detail in the description).

Divergent thinking is associated with creative output but is not identical with it. Creative work requires both divergent generation of possibilities and convergent selection and refinement of the best ones. The creative process is alternation between these modes: expand the possibility space, then select and develop within it. Pure divergent thinking produces many ideas; pure convergent thinking produces solutions to well-posed problems. Creativity requires both in sequence.

The Combinatorial Core

Where do new ideas come from? The uncomfortable answer, documented in historical case studies and cognitive research, is that they come from existing ideas in new combinations. Darwin’s natural selection combined Malthusian population pressure with Humean notions of variation — separately available to many contemporaries, combined by Darwin. The telegraph combined the electrical wire with the communication use-case of the pneumatic tube. Newton’s calculus combined coordinate geometry with the method of exhaustion from Archimedes. Every account of creative insight, examined closely, reveals prior components.

This is not debunking. The combination is the creative act. The insight that two existing ideas, connected in the right way, produce a third idea that neither contains is non-trivial. It requires that you have both components in your available conceptual space, that you encounter the right juxtaposition, and that you recognize the connection’s potential. Any of these can fail; the successful combination is still an achievement.

The practical implication is that creative output is dependent on the range and quality of inputs. Wider reading, broader experience, deeper expertise across multiple domains — all expand the combinatorial space of available connections. T-shaped knowledge (deep in one domain, broad across several) produces richer raw material for combination than narrow depth alone.

Remote Associates and the Aha Moment

Sarnoff Mednick’s Remote Associates Test (RAT) operationalizes the combinatorial view. Three seemingly unrelated words are given, and the task is to find a fourth word that connects all three. (Pine / Crab / Sauce → Apple.) The test measures the ability to find non-obvious links between remote concepts — which correlates with creative performance on other tasks.

The neuroscience of insight moments — the “aha” experience — consistently implicates the right anterior temporal lobe. EEG and fMRI studies show a distinctive burst of high-frequency gamma activity in this region approximately 300 milliseconds before subjects report a sudden solution to a compound remote associate problem. This neural signature corresponds to the subjective experience of insight: a solution that arrives whole, not through step-by-step reasoning.

Mark Jung-Beeman and John Kounios found that the state immediately before an insight solution is characterized by reduced visual cortex activity (looking inward rather than at the external environment) and increased activity in the right hemisphere’s default mode regions. The insight moment is preceded by a kind of internal quieting — attention turning away from the problem space toward the broader associative field where the solution was waiting.

The Role of Expertise

Creativity in any domain is constrained by domain knowledge. Boden’s concept of P-creativity (personally novel — new to the creator) versus H-creativity (historically novel — new to history) is useful here. P-creativity is achievable without expertise; H-creativity, in most domains, requires it.

The reason: the combinatorial space of a domain is only accessible to someone who knows the domain’s existing ideas well enough to see the gaps and possibilities. A physicist who doesn’t know the existing literature can’t make a contribution that’s historically novel because they won’t know that the combination they’ve produced hasn’t been produced before. A musician who doesn’t know harmonic theory can’t deliberately extend or violate it; they can only accidentally stumble across violations. The combinations that are original at the historical level require knowing what has and hasn’t been combined before.

This is the expertise paradox: deep domain knowledge provides the raw material for creativity but can also constrain it. Experts know too well what’s “already been done,” what violates domain norms, what the respected approaches are. The creative use of expertise requires some capacity to treat the domain’s constraints as provisional — to ask “what if this rule didn’t apply?” — while still knowing the rules well enough for the question to be meaningful.

What This Changes About Practice

The combinatorial view of creativity has direct implications for practice.

Read widely across domains, not just in the domain you’re working in. The connections that produce H-creative ideas often come from importing concepts from outside the domain into it. Darwin imported Malthus; Feynman imported the toy problems his father posed as puzzles; the designers of the Macintosh imported calligraphy aesthetics. Cross-domain reading is not a diversion from serious work — it is the feedstock of serious creative work.

Expose yourself to diverse problems before you need to solve a specific one. The incubation period (discussed in the next entry) is a period during which remote associations are made below the threshold of conscious attention. A richer store of prior exposures means more material for the incubation process to work with.

Don’t dismiss partially formed ideas. The combination that produces insight often requires a long chain of intermediate steps, any of which might be invisible at the time. An idea that seems unproductive today may be the missing component that makes a different idea click two years from now.

The creative process is not separate from learning and doing. It is what learning and doing produce when accumulated over time in a mind that’s actively looking for connections.