Divergent thinking
Generating many varied possibilities from a single starting point.
What it means
Divergent thinking is the cognitive ability to generate numerous, varied, and original responses to an open-ended prompt, producing a broad search across possibilities rather than converging on one answer. It is commonly indexed by fluency (number of ideas), flexibility (range of categories), originality (rarity), and elaboration (detail), and is treated as a key component, though not the whole, of creativity. Divergent thinking is fostered by deferring judgment, broad attention, positive affect, and remote association, and it precedes the convergent evaluation that selects and refines ideas. Critics note that scores on divergent-thinking tests predict real creative achievement only modestly, since execution and domain knowledge also matter. It remains a central construct for understanding and training ideation.
Where the construct came from
J. P. Guilford introduced the idea in his 1950 address to the American Psychological Association, arguing that the field had neglected creativity in favor of the single-answer intelligence testing that dominated it. In his structure-of-intellect model he split divergent production, which fans out toward many answers, from convergent production, which narrows toward the one correct answer. E. Paul Torrance then operationalized the idea as the Torrance Tests of Creative Thinking, still the most administered creativity battery, and the alternate-uses task, listing uses for a brick or a paperclip, became the workhorse measure. The framing mattered because it made creativity look testable and trainable, and it seeded the assumption, later found only partly true, that counting a person's ideas estimates their creative potential.
How it is scored
Scoring is where divergent thinking gets hard. Fluency is a simple count, but flexibility, originality, and elaboration require judgment, and originality is traditionally credited by rarity within a sample, so the same answer's score shifts with the comparison group. Hand-scoring is slow and inconsistent between raters, which pushed the field toward automation. Beaty and Johnson's SemDis platform scores originality as the semantic distance between the prompt and the response across several language models; that tracked human raters only modestly. Fine-tuned large language models later did far better, with Organisciak and colleagues reaching roughly r = .8 against human judges. A persistent trap is that fluency inflates every other index, since people who list more ideas also tend to hit more original ones, so scores must be adjusted for sheer output before they mean much.
What the evidence shows
Divergent-thinking scores predict real-world creative achievement, but weakly. Kim's 2008 meta-analysis put the correlation at about r = .22, slightly ahead of IQ's r = .17, and a more recent meta-analysis by Said-Metwaly and colleagues reported a similarly small link, meaning these tests explain only a small share of the variance in what people actually create. Runco and Acar argue this is partly the wrong yardstick: the tests measure potential, not accomplishment, and potential only converts to achievement through domain knowledge, persistence, and the convergent work of selecting and finishing an idea. So the construct is reliable and modestly valid as a snapshot of ideational ability, but a high score is a starting point rather than a forecast of creative output.
Limits and honest caveats
Several cautions follow. Divergent thinking is close to necessary but far from sufficient; ideation that never converges produces volume, not value. The tests are largely domain-general, yet real creativity is domain-specific, so a fluent brick-uses scorer need not be an inventive engineer. Time on task matters too: give people longer and their later ideas tend to be more original, so short administrations undersell slow thinkers. And the emphasis on producing many options can mislead teams into treating idea count as progress. In practice the usable core is procedural rather than diagnostic: separate generation from evaluation, defer judgment while ideas accumulate, then switch deliberately into convergent mode to prune and develop. The generating half is what divergent thinking names; the disciplined pruning half decides whether any of it matters.
Examples
Asked how many uses a brick could have, a high diverger lists a doorstop, weapon, paperweight, heat store, chalk, and art object across very different categories.
A team told to name twenty ways they could lose their biggest customer, judgment suspended, gets past the three obvious answers into the strange ones that turn out to be the real risks.
Naming a new coffee shop, one person offers three variations on 'Bean'; another offers a street name, a bird, a joke and a Portuguese word. Same fluency, far more flexibility.
A design instructor makes students sketch ten different chairs before choosing one, forcing them past the first obvious form into shapes no one would reach by refining the first sketch.
Working a differential diagnosis, a clinician deliberately lists every condition that could fit the symptoms, common and rare, before narrowing, so a treatable outlier is not lost to the first plausible label.
First described in J. P. Guilford (1950s); Torrance tests.
Key references
- Said-Metwaly, S., Taylor, C. L., Camarda, A., & Barbot, B. (2024). Divergent thinking and creative achievement—How strong is the link? An updated meta-analysis. Psychology of Aesthetics, Creativity, and the Arts, 18(5), 869–881. doi.org/10.1037/aca0000507
- Organisciak, P., Acar, S., Dumas, D., & Berthiaume, K. (2023). Beyond semantic distance: Automated scoring of divergent thinking greatly improves with large language models. Thinking Skills and Creativity, 49, 101356. doi.org/10.1016/j.tsc.2023.101356
- Beaty, R. E., & Johnson, D. R. (2021). Automating creativity assessment with SemDis: An open platform for computing semantic distance. Behavior Research Methods, 53(2), 757–780. doi.org/10.3758/s13428-020-01453-w
- Runco, M. A., & Acar, S. (2012). Divergent thinking as an indicator of creative potential. Creativity Research Journal, 24(1), 66–75. doi.org/10.1080/10400419.2012.652929
- Kim, K. H. (2008). Meta-analyses of the relationship of creative achievement to both IQ and divergent thinking test scores. The Journal of Creative Behavior, 42(2), 106–130. doi.org/10.1002/j.2162-6057.2008.tb01290.x