Techniques for generating novel ideas, reframing problems, and escaping fixed mental models. Covers lateral thinking (de Bono), Six Thinking Hats, random stimulation, provocation operators (PO), analogical transfer, assumption challenging, and divergent-then-convergent thinking cycles. Use when the goal is to produce new options or perspectives rather than evaluate existing ones.
Critical thinking is often conflated with evaluation — testing arguments, checking evidence, detecting biases. But equally important is the generative side: producing ideas worth evaluating in the first place. Creative thinking is the discipline of escaping fixed mental models, generating novel options, and reframing problems so that previously invisible solutions become visible. This skill catalogs the core techniques, grounded primarily in de Bono's lateral thinking framework and complementary approaches.
Agent affinity: de-bono (lateral thinking, Six Hats, PO operators), dewey-ct (reflective thinking that enables creative flexibility), paul (integration)
Concept IDs: crit-charitable-interpretation, crit-decision-frameworks
| # | Technique | Purpose | Origin |
|---|---|---|---|
| 1 | Divergent then convergent | Separate idea generation from evaluation | General |
| 2 | Six Thinking Hats | Role-play different thinking modes | de Bono |
| 3 | Lateral movement | Escape the dominant pattern | de Bono |
| 4 | Provocation (PO) | Use a deliberately absurd statement to generate new paths | de Bono |
| 5 | Random stimulation | Use an unrelated word or image as a seed | de Bono |
| 6 | Assumption surfacing | List what you are taking for granted, then negate each | General |
| 7 | Analogical transfer | Borrow a solution structure from another domain | Gick & Holyoak |
| 8 | Reframing | Restate the problem in different words to shift perspective | Schön |
| 9 | SCAMPER | Substitute, Combine, Adapt, Modify, Put to other uses, Eliminate, Reverse | Eberle |
| 10 | Morphological analysis | Systematically combine attributes of the problem space | Zwicky |
| 11 | Five Whys (for creativity) | Drive down to underlying need, then generate solutions at that level | Ohno |
| 12 | Constraint removal | Ask what the solution would look like with a key constraint removed | General |
The single most important rule of creative thinking: do not evaluate while generating. Evaluation is a different cognitive mode that shuts down the generation process. Generate first, evaluate later.
Standard procedure:
Violating this sequence — evaluating during generation — is the most common failure mode of group brainstorming. One skeptical voice early in the process can kill the divergent phase entirely.
Pattern: Assign each thinker (or each round of thinking) one of six colored "hats," each representing a different thinking mode. Rotate through them systematically so that every mode gets its turn.
Why it works. Normal discussion lets each person hold multiple modes simultaneously, which usually defaults to critical evaluation. The hats enforce a single mode at a time, making creative generation possible as its own stage.
Worked example. A team discussing a new product launch runs Six Hats:
Pattern: Deliberately step sideways from the dominant line of thinking rather than forward along it. Ask "what is the next adjacent idea?" rather than "what is the deeper version of this idea?"
Worked example. The problem: city traffic is too congested during rush hour.
Vertical (linear) thinking: more lanes, better traffic lights, highway expansion, rapid transit.
Lateral movement: make rush hour irrelevant (remote work), price the road (congestion charges), change the pattern (stagger work hours), change the geometry (bikes and scooters), change the concept (city redesign so commutes shrink).
Lateral movement does not abandon the linear ideas — it asks what other directions exist.
Pattern: Deliberately state something absurd, impossible, or clearly wrong, then use it as a starting point. Extract whatever value the provocation generates before returning to the real problem.
Syntax. "PO: [absurd statement]." The PO marker signals that the statement is a deliberate provocation, not a real claim.
Worked example. Problem: grocery stores lose money on expired produce.
PO: Customers should pay stores to take produce they do not want.
Analyze the provocation. What if customers literally paid stores? That would make stores sell produce faster. What if there were a "last-day discount auction" where customers bid down the price? That's a real idea — dynamic discounting. The provocation unlocked the real idea even though the provocation itself was absurd.
Discipline. Treat the provocation as a stepping stone, not a target. The goal is what the provocation reveals, not the provocation itself.
Pattern: Generate a random word, image, or object and force a connection between it and the problem. The arbitrariness of the stimulation forces the mind to build new paths.
Worked example. Problem: how to improve library checkouts. Random word: "butterfly."
Connections: light touch, brief visit, migration patterns, transformation from one form to another, fragility, colorful display. Ideas: a "touchpoint" checkout that takes two seconds instead of thirty; a book that transforms its appearance based on who's read it; migration patterns mapping which books move through which branches.
Why it works. Random words are uncorrelated with the problem, which means they force creative connections rather than reinforcing habitual thoughts.
Pattern: List every assumption underlying your current view of the problem. For each, ask "what if this were false?"
Worked example. Problem: reducing restaurant wait times.
Assumptions:
Negations:
Each negation opens a new design space. Surfacing and negating assumptions is the single highest-leverage creative technique because assumptions are invisible until they are named.
Pattern: Find a well-solved problem in a distant domain that shares structural features with yours. Map its solution onto your problem.
Worked example. Problem: hospital emergency room triage is inefficient.
Analogy: airline boarding groups. Airlines solved the problem of boarding hundreds of people through one door by creating priority groups and sequencing them. What if the ER had formal "boarding groups" for common presentations, with parallel processing streams?
Discipline. Analogies must share the right features. The analogical source should match the target on the dimensions that matter, not on surface features. Airlines and ERs both manage serialization of many people through limited capacity — that's a structural match.
Pattern: Restate the problem in different words. The new framing often reveals solutions invisible in the original.
Worked example. Original frame: "How do we get customers to wait patiently?"
Reframings:
Each reframing points toward different solution spaces. Reframing is especially powerful when the original problem statement contains a hidden assumption about where the solution must come from.
Pattern: Apply each of seven operators to an existing product, idea, or process to generate variants:
Worked example. Existing product: a coffee mug.
SCAMPER works because it provides seven specific generation prompts rather than the vague "be creative."
Pattern: Identify the key attributes of the problem space. For each attribute, list the possible values. Combine across attributes to generate the full option space.
Worked example. Designing a new transportation service.
| Attribute | Values |
|---|---|
| Speed | Walking, cycling, driving, flying |
| Capacity | 1 person, 2-5, 6-50, 50+ |
| Energy source | Human, electric, gas, hybrid |
| Ownership | Private, shared, rented, subscription |
| Autonomy | Manual, assisted, self-driving |
The Cartesian product (4 × 4 × 4 × 4 × 3 = 768 combinations) defines the option space. Most are unworkable, but the systematic exploration surfaces combinations nobody would have thought of otherwise.
Pattern: Drive down to the underlying need by asking "why" five times, then generate solutions at the root level rather than the surface.
Worked example. Customer asks for a faster horse.
Now generate at the real level. "Time with family" opens up many solutions besides transportation: remote work, family visits at work, flexible hours. The original "faster horse" request was framed too narrowly.
Pattern: Identify a key constraint and imagine the solution if that constraint were relaxed. Often the useful answer is a partial relaxation.
Worked example. Constraint: budget of $10,000. What would we do with a $100,000 budget? What about $1,000? $10?
The $100,000 version reveals what's actually valuable that we're skipping for cost. The $10 version forces radical minimalism and often reveals that most of the $10,000 budget is spent on unnecessary overhead.
| Mistake | Why it fails | Fix |
|---|---|---|
| Evaluating during generation | Shuts down the divergent phase | Enforce a strict generate-then-evaluate sequence |
| Stopping at the first good idea | Misses better options | Generate at least 20 before evaluating |
| Ignoring wild ideas | Wild ideas often contain the seed of real ones | Treat absurd ideas as provocations, not targets |
| Reframing into problems you prefer | Not reframing, just dodging | Verify the reframe addresses the real concern |
| Using creativity techniques ritualistically | Mechanics without mindset | Remember the goal — escape the dominant pattern, not check boxes |