Why your best ideas come after your worst

This article very nearly didn’t exist. For several weeks, it made a conspicuous effort not to. It began, or rather did not begin, when I was invited to pitch a second article for Big Think about virtually any topic in neuroscience. 

Triumph. I had freedom and unlimited time. What could be easier? A lot, it turns out. Weeks went by, and I did not write. My inbox began to fill with cheerful nudges from Stephen, my editor. Still keen to write something?

The first time around had been cleaner. I’d been assigned a slot in Big Think’s consciousness issue on a tight deadline. That forced me to write about the neuroscience behind sleep, a relevant topic I knew well enough to write cold. There was no dithering, just a window of opportunity narrowing by the minute.

This time, the window was wide open. My brain was free to go anywhere, and so, of course, it went nowhere. I emailed Stephen with a counterproposal: Would he mind reducing the scope of “anything”? He sent over a handful of ideas, one of which centered on a strategy to spark creativity, coined by Scott Dikkers, founder of The Onion.

“The Clown and the Editor?” I had no idea what it meant, but it had a nice ring.

I opened a blank document titled Clownnotes_1 and resumed not-writing with renewed discipline. I researched several subjects I would not write about. Ideas were everywhere — some I lobbed at Stephen with excitable incoherence, the rest I dragged into private extinction under Clown_2, clown10, and the self-explanatory trash.docx.

I emailed Stephen: “What if I cough up a filthy draft? Something terrible. We’ll call it a starting point.” He agreed, and that afternoon, I deleted 12 of 13 documents, preserving any threads I thought showed potential. Eventually, I started looking into the Clown-Editor concept I was supposedly writing about. 

The two halves of the creative mind

The Clown, in Dikkers’ model, is the mind’s generator. The Clown rapidly throws all types of ideas against the wall, not caring about the audience, their judgment, or even basic feasibility. Those issues can wait for the Editor — the part of the mind that makes sure you’ve got something defensible to submit. An early interruption from the Editor, however, risks killing your idea before ink hits the page.

Without the Clown, there are no novel ideas; without the Editor, those ideas lack concision, coherence, and logical structure. 

Soon enough, my article began taking shape. The deleted drafts had achieved their purpose; I could see flickers of thought and phrasing that were clear descendants of their chaos. I realized that I’d been using this rhythm of reckless creation and careful revision throughout my writing life — I just hadn’t put a name to it.

The Clown–Editor framework is meant to be a simple, intuitive way to approach the writing process, not a scientific claim. Yet the two characters are mirrored by real, measurable neural networks, each playing a distinct role in the creative process. What sets proficient creatives apart is dexterity in switching between these modes.

The Default Mode Network (DMN) is the Clown who shows up barefoot, eagerly blurting out hypotheses about what a frog might wear if invited to dinner (not unlike those early emails I lobbed at Stephen). The DMN is your brain’s internal daydream factory: It recruits cortical midline structures, like the medial prefrontal cortex and posterior cingulate, to let thoughts drift. 

The Executive Control Network (ECN) is the pragmatic adult in the room, the Editor that marshals frontoparietal systems in pursuit of internal goals and strategic planning. 

These two systems tend to operate in opposition: When one is active, the other typically quiets. Creativity, at the network level, is about passing the baton at the right moment: letting the DMN meander, then handing over to the ECN for focused development. Functional MRI during divergent thinking has allowed us to observe this temporal handoff, typically characterized by a prolonged period of DMN activation, followed by ECN engagement. Highly creative thinkers seem to modulate these transitions more flexibly and fluently.

Because these two networks have incompatible priorities, they tend not to multitask well, and trying to engage both at once can lead to shutdown. In creative writing, we call this “blocking.”

The dorsolateral prefrontal cortex (dlPFC), loyal agent of the ECN, is great at vetoing weak ideas but terrible at offering alternatives. An early appearance from the Editor means premature evaluation — a mind interrupting itself just as the creative thoughts were getting fun.

This is why so many great writers swear by deliberately messy first passes. John McPhee describes his early creative process as “flinging mud at a wall.” If we’ve any hope of completing a first draft, he explains, we’ve got to “blurt out, heave out, babble out something—anything.” This method works precisely for its arbitrariness; you’re tricking the DMN to hand you something — anything — tangible, then protecting that something from early prosecution by delaying the ECN. 

Anne Lamott’s “shitty first draft” strategy, Ernest Hemingway’s write drunk, edit sober mantra, Dikkers’ Clown and Editor. All of them, knowingly or not, have trained for neural timing: DMN first, ECN second. And me too, I suppose, given that “What if I cough up a filthy draft?” ended up being the “open sesame” to this article.

Your first ideas are almost certainly terrible

Premature evaluation aside, there’s another obstacle waiting in the wings: your first ideas. Mine were sent to die in a string of ecstatic emails and since-deleted documents, but could just as easily have trapped me into early resignation. The romance of inspiration would have us believe originality sprints from the starting gun, when actually, it more often catches up — if you keep going. 

Original ideas tend to arrive later, because the mind needs extra beats to exit well-worn associative paths. Ask your brain to dream up possible uses for a spoon, for example, and its immediate response will be things you’ve spent decades training it to associate spoons with. Once it has exhausted all the spoonable foods, it’ll start listing known uses for spoon-like objects, then whatever else is semantically nearby. 

This has a behavioral analog in what’s colloquially known as the streetlight effect, named after the allegory about a drunk searching for his lost keys under a streetlight. A policeman asks him if he’s certain this is where he lost them; the drunk says no, he’d lost them in the park. “Then why are you searching here?” the policeman asks. “Because this is where the light is.”

The brain mirrors this tendency, searching high-salience regions in semantic memory that are brightly lit by years of repeated exposure. It shows up as the order effect, where early ideas are plentiful and predictable, while later ideas are fewer but, on average, more original. A study of divergent thinking in children found that the originality of ideas peaked around the seventh or eighth idea, suggesting that good ideas might only come after most people would’ve settled or given up.

Originality, in that sense, is a function of search persistence. It comes not from inspiration, but from staying in the problem space longer than your neural heuristics want to. Slowing down helps too; longer latency between successive ideas can lead to more creative responses. 

Persistence has its limits, however. Sometimes, what’s needed is complete disengagement. Incubation is the performance boost that comes when you return to a creative task after taking a break from it. There are multiple versions of this: short delays, long ones, breaks filled with unrelated tasks — with sleep, or with nothing at all. But across dozens of studies, the effect holds: Stepping away helps, especially when your initial attempts have seduced you into the wrong corner of the solution space. Incubation gives your mind time to clear out bad ideas that were beginning to take root. 

Slowing down gives originality space to emerge, but while persistence and incubation help you stay in the game, they don’t guarantee progress. You still need the problem to start taking shape in the first place — if only so your brain knows where to send the next idea.

The tyranny of “anything”

In 1960, Theodor Seuss Geisel, better known as Dr. Seuss, was dared by his publisher to write a children’s book using no more than 50 unique words. The result was Green Eggs and Ham, a triumph that went on to sell over 200 million copies. Inspired by this, psychologist Catrinel Tromp formally proposed the “Green Eggs and Ham Hypothesis,” which argues that creative output improves under imposed constraints.

Constraints give creative energy a direction to travel, shaping the search space while also confining it — just enough that exploration becomes deeper, weirder, and more productive. In cognitive terms, constraints funnel the search away from dominant semantic pathways, encouraging more remote associations. They lower the probability that the first answer — the one glowing under the streetlamp — will be the one we accept, forcing the Editor to sharpen their criteria, and the Clown to search less obvious terrain.

Constraints can also be the thing that gets you generating in the first place. Before idea generation can begin, the system has to decide how to search memory for material. When the brief is too open, retrieval policies compete. That’s where I was, given total freedom to write anything. My brain had nowhere specific to look, so it didn’t look at all. With too many possible paths through memory, competing features and priorities interfered, which meant no retrieval strategy gained control — none were selected — so the search couldn’t begin.

Stephen’s stray ideas gave me a starting condition, a concrete frame that specified features, relations, and relevance. A point from which associations could begin to form. In studies of semantic cognition, this is where the left ventrolateral prefrontal cortex (VLPFC) comes in. Under constraint — when the task is at least partially defined — the VLPFC begins to bias memory retrieval toward features that match the problem. It narrows the scope and sets a direction. Only then do other systems get involved.

When the task is internally focused and structured, two major networks begin a tag-teamed collaboration. The DMN, which supports associative and conceptual processing, couples with the ECN, which supplies top-down control and goal-directed planning. Clown-Editor coupling happens when the system has a starting point, which, in open-ended creative tasks, can be a constraint or an initial stimulus.

For me, the ignition cue that sparked creative Clown-Editor coupling was a passing mention of the Clown and the Editor — a concept that I did, eventually, sit down to research. Though only after a sufficient period of Clownly abandon.