top of page
Search

The Difference Between Generating and Creating: Why Most of What We Call Creativity Is Fluency in Disguise

  • 2 days ago
  • 6 min read

There is a distinction that rarely gets made in creativity discourse, and it has become more urgent as generative AI colonizes the vocabulary of "original thinking." The distinction is this: producing something unfamiliar is not the same as thinking creatively. Arranging known elements in a novel sequence - which is precisely what most people do when they believe they are being creative - is better described as combinatorial fluency. It is useful. It is pleasurable. But it is not creativity in any meaningful sense of the word.

Arthur Koestler identified the deeper mechanism in "The Act of Creation" (1964), calling it bisociation - the moment when two independent, self-consistent planes of thought collide and the collision is not resolution but illumination. The critical word is "independent." Bisociation does not describe a person drawing on two ideas they already habitually associate. It describes a mind holding two genuinely incompatible frames simultaneously - not resolving the contradiction, but producing something out of the tension between them. Most creativity training, most brainstorming methodologies, and essentially all of what AI currently calls "creativity" operate within a single associative network at a time. They are fast. They are productive. They are not bisociative.

The problem is one that psychologists Posner, Strike, Hewson and Gertzog identified in a different context - what they called "conceptual change theory." Deep cognitive restructuring requires a prior state of genuine dissatisfaction with one's existing framework. Without that dissatisfaction, new inputs are simply absorbed into old structures. The result feels creative to the person producing it because something new came out, but structurally the mind has not reorganized at all. This is why a person can spend years in "creative work" and find that their output, while varied, traces the same underlying topology of ideas over and over. The vocabulary expands; the cognitive architecture does not.

Creativity, as opposed to fluency, requires what the psychologist Else Frenkel-Brunswik described as tolerance of ambiguity - not as a personality trait to be admired, but as a trainable capacity to remain in a state of unresolved cognitive tension long enough for something genuinely new to form. Janusian thinking, as identified by Albert Rothenberg in his studies of Nobel laureates and major artists, is the simultaneous conception of two contradictory ideas as both valid and operative. Not "I considered both sides." Not "I synthesized opposing views." Both things, unresolved, held in the same mental space at the same time. Einstein reconciling curved and flat models of the universe. Dylan holding protest and popular music as mutually incompatible and refusing to dissolve the tension into one or the other. The output of Janusian thinking does not feel like a compromise. It feels - and this is Rothenberg's observation - like a rupture.

This matters enormously in language learning, where the confusion between fluency and creativity has done particular damage. Advanced learners plateau not because they lack vocabulary or grammatical competence, but because they have become exceptionally good at producing fluent, contextually appropriate, cognitively safe language. Their interlanguage has stabilized into an echo chamber. They can express familiar thoughts with increasing elegance. They cannot yet think in the second language in genuinely unfamiliar territory - which is where cognitive growth actually happens. The neuroscientific framing is straightforward: fluency reduces cognitive load because it draws on consolidated neural pathways. Real creative thinking demands the activation of what researchers call "spreading activation" across distant conceptual nodes - and the more fluent a person becomes, the more effort it takes to reach those distant nodes rather than defaulting to well-worn associative chains. Fluency, perversely, can become the enemy of creative development unless it is regularly disrupted.

The philosopher Viktor Shklovsky described defamiliarization - ostranenie - as the fundamental function of art: to make the familiar strange again, to restore the sensation of perception where habit has replaced it. Shklovsky was writing about literature in 1917, but the cognitive principle he identified is now supported by a substantial body of work in predictive processing. The brain operates as a prediction machine, constantly modeling expected inputs and suppressing conscious attention to confirmed predictions. Novel stimuli that violate strong predictions produce what researchers call "interesting failures" - moments when the predictive model breaks down and the system is forced to generate new hypotheses. These failures are not errors to be corrected. They are the neurological signature of creative potential. Most educational environments, and most conventional creativity training, do almost everything possible to prevent them.

This is where randomness becomes something more than a gimmick. The cognitive benefit of what might be called "grandomness" - genuinely improbable stimulus pairings rather than merely uncommon ones - is that they create the conditions for low-probability combinatorics: deliberate collisions between elements with a very low statistical probability of co-occurring in natural language or ordinary thought. High-probability associations (the ones produced by most brainstorming, most AI prompts, and most creative exercises) are cognitively cheap precisely because the pathways already exist. Low-probability pairings force the construction of new pathways - or, in Koestler's terms, the bridging of genuinely independent matrices. The difference between a creative insight and a clever arrangement is exactly the difference between a new bridge and a well-maintained existing road.

The practical consequence is uncomfortable: most of what is marketed as creativity training does not train creativity. It trains creative output - which is different. Output-oriented training produces people who can reliably generate more ideas, who are faster at brainstorming, who are comfortable with ambiguity as a concept while remaining deeply uncomfortable with it in practice. The distinction matters because it predicts what happens when genuine cognitive complexity is demanded - in a high-stakes negotiation, in a real interview, in a context where no scaffold exists and something genuinely unfamiliar must be navigated in real time. Fluent people become rigid. Creatively trained people become generative.

Grandomastery (https://grandomastery.com) was built explicitly around this distinction. Its theoretical architecture draws on bisociation, defamiliarization, spreading activation, and what Boden called "transformational creativity" - not to combine these frameworks decoratively, but because each addresses a specific mechanism that output-oriented training ignores. Activities like Random Abstractions (https://grandomastery.com/abstractions) and Random ISM (https://grandomastery.com/ism) do not ask for fluent recombination of familiar ideas. They require holding genuinely incompatible conceptual frames simultaneously and producing something out of the collision - not a synthesis, not a compromise, but a response to the unresolved tension. The Random Presentation activity (https://grandomastery.com/presentation) goes further: it trains diachronic creativity - the ability to maintain creative coherence across a sequence of unrelated and unpredictable prompts over time - rather than the synchronic creativity that most exercises develop (one prompt, one response, done).

The broader cultural dimension of this problem is not often discussed. We are in a period where the primary institutional pressure on creative practice is toward legibility, reproducibility, and measurability - all of which systematically favor fluency over creativity. PISA's creative thinking assessment, for all its improvements in 2022, still evaluates creativity within standardized formats and predetermined domains. Rubric-based scoring by definition penalizes responses that fall outside the rubric - which is precisely where bisociative, transformationally creative responses tend to fall. The message this sends is not subtle: creativity that cannot be categorized and scored is not real creativity, it is error. This is educationally catastrophic, and the arrival of AI that can generate rubric-satisfying creative output at scale makes it more so, not less.

The philosopher Charles Peirce's concept of abductive reasoning - the generation of the most plausible explanation from incomplete information - is perhaps the best single description of what real creative thinking does that fluency cannot. Abduction does not extrapolate from known patterns. It proposes. It leaps. It accepts that the leap may be wrong while insisting that the leap is necessary. It is irreducibly human not because humans are better at it than machines in any simple computational sense, but because it requires genuine engagement with uncertainty as a productive state - not a problem to be resolved, but the condition under which something genuinely new can form. Training that does not put people repeatedly into that state, that does not demand the construction of meaning where no obvious meaning exists, does not train creativity. It trains the appearance of it.

The question worth sitting with is not "how do I become more creative?" but "how do I stop mistaking fluency for creativity?" The answer has to begin with genuine cognitive discomfort - not as a self-improvement metric, but as a design principle for how we learn, teach, and practice thinking.



Recent Posts

See All

Comments

bottom of page