IX Coach

Divergent Thinking: Generating Many Ideas

Generating quantity, separating ideation from judgment, and why group brainstorming disappoints

What is divergent thinking and how do you get more ideas?

Divergent thinking is the mental process of generating many varied ideas from a single starting point without judging them initially — the opposite of convergent thinking, which narrows toward one correct answer. The psychologist J. P. Guilford introduced the distinction in the 1950s as part of his model of intelligence, and it is usually described along four dimensions: fluency (how many ideas), flexibility (how many different categories), originality (how unusual), and elaboration (how developed). Withholding judgment during the generating phase is the definitional core, not a stylistic preference: an idea evaluated the moment it appears rarely survives long enough to lead anywhere. The most reliable practical principle that follows is to separate idea generation from evaluation into two deliberate passes. An honest caveat: traditional group brainstorming usually underperforms the same people working alone and pooling ideas, due to a documented productivity loss.

Divergent thinking is the generative half of creativity: producing many, varied, original ideas before narrowing. J. P. Guilford named it in the 1950s to distinguish it from convergent thinking, the search for the single right answer that most testing rewards; divergence is measured instead by how many ideas you produce, how varied they are, how original, and how far you develop them. The durable, well-supported lever is separating generation from evaluation — judging too early kills volume. The honest complication is that the classic group brainstorm, despite its popularity, tends to produce fewer and worse ideas than the same people generating alone, a finding replicated for decades. Below are practices that respect both the lever and the caveat.

Practices

How these practices connect

Every one of these practices works by delaying or removing a specific form of premature narrowing so that generation can run longer before selection cuts it off. Self-censorship, production blocking in groups, clustering around the first obvious ideas, and converging on a single option too soon are all the same failure arriving through different channels — judgment, social dynamics, habit, or timing. Separating generation from evaluation, individual output from group pooling, and divergence from convergence all buy the same thing: more of the idea space gets explored before anything gets cut.

if you're not supposed to judge your ideas while you're coming up with them, when are you actually supposed to start picking the good ones?

Deferring judgment during generation only works if it's paired with a separate evaluation pass afterward — it isn't a rule to never judge, it's a rule about sequencing. The two-phase model treats creativity as diverge first, then converge: you open up and generate without criticism, and only once that phase is exhausted do you switch into a distinct evaluative mode to select and refine. Judging early shuts down the fragile, riskier ideas before they've had a chance to form, which is exactly what the deferred-judgment phase is protecting. Converging before the divergent phase has run its course is the common failure mode on the other side — it locks in a mediocre option instead of finding a better one.

does having everyone write down their own ideas alone first basically do the same job as just telling people to come up with a lot of ideas?

Not quite — they solve different bottlenecks. Going for quantity is about pushing any one person past their first, most obvious ideas to reach the more original ones further down the list. Brainwriting solves a separate problem specific to groups: in a live verbal brainstorm, only one person can talk at a time, so others wait, forget what they were going to say, and self-censor — a loss called production blocking. Having everyone generate silently and individually before pooling lets each person still hit their own quantity target in parallel, instead of a few voices dominating the room while everyone else's ideas evaporate.

if writing ideas down alone already avoids the group pressure problem, why do you still need a separate rule about not judging your own ideas?

Because brainwriting fixes an interpersonal problem, not an internal one. Production blocking and a few voices dominating happen between people in a live group setting; switching to silent, individual writing removes that specific loss. But a person working alone can still shut down their own idea the instant it appears — that's self-censorship, and it happens entirely inside one person's head regardless of whether anyone else is in the room. Deferring judgment is the separate discipline that protects against that internal shutdown, which is why it's still needed even when brainwriting has already solved the group dynamics.

why would you need structured prompts like SCAMPER if you're already trying to generate as many ideas as possible?

Because quantity and variety aren't the same thing, and pushing for one doesn't guarantee the other. Left alone, idea generation tends to cluster around one region — you can produce a long list that's still mostly variations on your first few thoughts. Structured prompts like substitute, combine, or adapt each force your thinking down a different transformation, systematically sampling regions you wouldn't reach through free association alone. They're most useful specifically when a high volume of ideas has stalled into repetition, not as a formality applied to ideas you're already satisfied with.

what's the actual difference between using a framework like SCAMPER and just looking at how a completely different industry solved a similar problem?

Both widen the search beyond what free association reaches, but the source of the novelty is different. SCAMPER prompts push your existing problem through a fixed set of transformations — substitute, combine, adapt — that stay within your own frame of reference. Cross-pollinating from a distant domain instead imports a structure that doesn't exist in your field at all, and the novelty comes specifically from the conceptual distance being crossed. The tradeoff is that distant analogies also produce more ideas that don't actually transfer, so they raise the ceiling on originality at the cost of needing more filtering afterward.

if you pull in a wild idea from a completely unrelated field, does that count against the goal of just generating a lot of ideas, or does it help it?

It helps it, but indirectly — cross-pollination is a source of variety, and generating for quantity is what turns that variety into a usable set of options. Ideas from a distant domain carry structures your own field doesn't have, so each analogy you pull in and adapt becomes one more entry toward your quantity target, not a detour from it. The two practices point in the same direction: variety of inputs is treated as the upstream cause of variety of ideas, and pushing for a high count is what forces you to keep drawing on those distant inputs instead of stopping at the first comfortable one.

How do you choose between these?

Start with the setting: if more than one person is involved, default to brainwriting over a live verbal session, since talking together specifically loses quantity and quality through production blocking — groups are still useful for building on ideas and getting buy-in, just not for the generation step itself. Whether you're alone or in that brainwriting phase, defer judgment and keep pushing for quantity past the first comfortable answers; that combination is the baseline regardless of context. Reach for SCAMPER prompts or cross-pollination specifically when the ideas on the page have started repeating themselves — SCAMPER when you want a fast, structured nudge within the problem as stated, cross-pollination when you have time to search further afield and filter what comes back. Whatever combination you used to generate, treat switching to evaluation as a distinct, later step rather than something that happens automatically once you feel you have enough.

What these approaches do not settle

The clearest tension in this set sits between generate-for-quantity and cross-pollinate-domains: one treats sheer volume as the route to originality, the other treats the diversity of your inputs as the actual cause, and generate-for-quantity's own caveat concedes that without varied inputs, more ideas can just mean more of the same — so quantity's payoff depends on the very thing cross-pollination is supplying. Scamper-prompts rests on the thinnest evidence here: it's flagged mechanistic rather than tested, meaning directed prompts in general are supported for widening search, but the specific SCAMPER framework has never been shown to outperform any other set of structured prompts. Brainwriting-not-brainstorming is the only practice in this group backed by randomized controlled trial evidence; the rest are observational or mechanistic, so their mechanisms should be read as well-reasoned and consistent with the evidence rather than as established causal fact.

Practice this with IX Coach

Practice this with IX Coach

Frequently asked questions

What is divergent thinking and how do you get more ideas?

Divergent thinking is the mental process of generating many varied ideas from a single starting point without judging them initially — the opposite of convergent thinking, which narrows toward one correct answer. The psychologist J. P. Guilford introduced the distinction in the 1950s as part of his model of intelligence, and it is usually described along four dimensions: fluency (how many ideas), flexibility (how many different categories), originality (how unusual), and elaboration (how developed). Withholding judgment during the generating phase is the definitional core, not a stylistic preference: an idea evaluated the moment it appears rarely survives long enough to lead anywhere. The most reliable practical principle that follows is to separate idea generation from evaluation into two deliberate passes. An honest caveat: traditional group brainstorming usually underperforms the same people working alone and pooling ideas, due to a documented productivity loss.

How do you defer judgment during generation?

Evaluation and generation compete: judging an idea the moment it appears activates self-censorship that suppresses the next, riskier idea before it forms. Separating the two phases protects the fragile early ideas long enough to develop, and the wild ones often seed the best final answer. The split is what keeps the pipeline open.

How do you go for quantity first?

Your first ideas are the most obvious and accessible; the original ones live further down, after the easy defaults are exhausted. Pushing for sheer quantity forces you past the obvious into the novel region, which is why a quota outperforms a vague “think of some ideas.” Volume is the path to originality, not a substitute for it.

How do you brainwrite instead of group brainstorm?

Verbal group brainstorming suffers “production blocking”: only one person can talk at a time, so others wait, forget, and self-censor, and a few voices dominate. Brainwriting removes the blocking by letting everyone generate in parallel and in private first, then combining — capturing the diversity of a group without the loss that talking together imposes.

How do you use SCAMPER and other prompts to vary ideas?

Left alone, idea generation clusters around one region. Structured prompts each push your thinking down a different transformation, so they systematically sample regions you wouldn’t reach by free association. The prompts increase the variety, not just the volume, of ideas — and variety is what distinguishes truly divergent thinking from listing.

What is Balance divergent with convergent thinking?

Divergent thinking alone produces a pile of options but no decision; convergent thinking alone optimizes too early on a poor starting set. Creativity needs both, in order: diverge to widen the space, then converge to choose and refine. Keeping them as distinct phases prevents each from sabotaging the other — premature convergence is the usual killer.

How do you cross-pollinate from distant domains?

Ideas from a distant domain carry structures absent in your own field, and transferring one across the gap produces a solution local thinking couldn’t reach. The conceptual distance is the source of novelty: a remote analogy bridges two regions of knowledge that don’t normally touch. Variety of inputs is the upstream cause of variety of ideas.

if you're not supposed to judge your ideas while you're coming up with them, when are you actually supposed to start picking the good ones?

Deferring judgment during generation only works if it's paired with a separate evaluation pass afterward — it isn't a rule to never judge, it's a rule about sequencing. The two-phase model treats creativity as diverge first, then converge: you open up and generate without criticism, and only once that phase is exhausted do you switch into a distinct evaluative mode to select and refine. Judging early shuts down the fragile, riskier ideas before they've had a chance to form, which is exactly what the deferred-judgment phase is protecting. Converging before the divergent phase has run its course is the common failure mode on the other side — it locks in a mediocre option instead of finding a better one.

does having everyone write down their own ideas alone first basically do the same job as just telling people to come up with a lot of ideas?

Not quite — they solve different bottlenecks. Going for quantity is about pushing any one person past their first, most obvious ideas to reach the more original ones further down the list. Brainwriting solves a separate problem specific to groups: in a live verbal brainstorm, only one person can talk at a time, so others wait, forget what they were going to say, and self-censor — a loss called production blocking. Having everyone generate silently and individually before pooling lets each person still hit their own quantity target in parallel, instead of a few voices dominating the room while everyone else's ideas evaporate.

if writing ideas down alone already avoids the group pressure problem, why do you still need a separate rule about not judging your own ideas?

Because brainwriting fixes an interpersonal problem, not an internal one. Production blocking and a few voices dominating happen between people in a live group setting; switching to silent, individual writing removes that specific loss. But a person working alone can still shut down their own idea the instant it appears — that's self-censorship, and it happens entirely inside one person's head regardless of whether anyone else is in the room. Deferring judgment is the separate discipline that protects against that internal shutdown, which is why it's still needed even when brainwriting has already solved the group dynamics.

why would you need structured prompts like SCAMPER if you're already trying to generate as many ideas as possible?

Because quantity and variety aren't the same thing, and pushing for one doesn't guarantee the other. Left alone, idea generation tends to cluster around one region — you can produce a long list that's still mostly variations on your first few thoughts. Structured prompts like substitute, combine, or adapt each force your thinking down a different transformation, systematically sampling regions you wouldn't reach through free association alone. They're most useful specifically when a high volume of ideas has stalled into repetition, not as a formality applied to ideas you're already satisfied with.

what's the actual difference between using a framework like SCAMPER and just looking at how a completely different industry solved a similar problem?

Both widen the search beyond what free association reaches, but the source of the novelty is different. SCAMPER prompts push your existing problem through a fixed set of transformations — substitute, combine, adapt — that stay within your own frame of reference. Cross-pollinating from a distant domain instead imports a structure that doesn't exist in your field at all, and the novelty comes specifically from the conceptual distance being crossed. The tradeoff is that distant analogies also produce more ideas that don't actually transfer, so they raise the ceiling on originality at the cost of needing more filtering afterward.

if you pull in a wild idea from a completely unrelated field, does that count against the goal of just generating a lot of ideas, or does it help it?

It helps it, but indirectly — cross-pollination is a source of variety, and generating for quantity is what turns that variety into a usable set of options. Ideas from a distant domain carry structures your own field doesn't have, so each analogy you pull in and adapt becomes one more entry toward your quantity target, not a detour from it. The two practices point in the same direction: variety of inputs is treated as the upstream cause of variety of ideas, and pushing for a high count is what forces you to keep drawing on those distant inputs instead of stopping at the first comfortable one.

Research

Articles that cover this