“I had a rough day at work yesterday…”
“So basically, your boss is bad at communicating. Here's what you should do.”
Congratulations. You've solved a conversation that hasn't even happened yet.
Fast thinkers can look impressive. But Toshio Okada's idea of three thinking gears raises a better question: is being quick to reach a conclusion the same as being good at understanding people? And is this simply Daniel Kahneman's System 1 and System 2 in different packaging?
The three gears: quick answers, ordinary thinking, and listening
Okada uses the term “combat thinking” for forming responses while a conversation is still moving. His proposed drills include giving yourself three or five seconds to answer, announcing that you have three reasons and organizing them as you speak, and sketching out an explanation before polishing it—like an artist drawing rough lines.
That last idea is useful: sometimes you discover what you mean by trying to say it.
But there's a catch. Declaring “I have three reasons” doesn't magically create a third reason. A clever presentation trick is not permission to invent facts.
The low gear is different. Instead of racing to summarize another person's story, you let it unfold, recall a relevant experience and respond to what was actually said. Slowing your urge to conclude can matter more than slowing your voice.
System 1 and System 2 aren't simply a speedometer
The popular shorthand is that System 1 is fast and intuitive, while System 2 is slow and deliberate. Useful, but incomplete.
A practiced speaker might construct a demanding argument very quickly. A friend might instantly notice sadness in your face. Neither example fits neatly into “fast equals System 1, slow equals System 2.”
In a 2013 review, Evans and Stanovich emphasize automatic processing on one side and more deliberate, working-memory-dependent reasoning on the other. Okada's gears are a practical metaphor for behavior, not a scientific relabeling of the two systems.
What the research actually backs up
A 2014 review by Heitz details a robust finding: speed and accuracy often trade off. Faster decisions typically use less evidence. Demanding both maximum speed and flawless judgment is like ordering a race car that never needs brakes.
Self-explanation has better evidence than the “three reasons” trick. A 2018 meta-analysis covering 64 reports found that prompting learners to explain material to themselves improved learning on average, with a moderate effect (g = 0.55). That doesn't prove that endless improvised talking makes you smarter.
Nor does drilling three-second answers establish a general intelligence boost. A 2016 meta-analysis of working-memory training found little convincing evidence that gains reliably transferred to unrelated abilities.
Plot twist: good listeners may answer faster
A large 2023 review found that feeling listened to was associated with better work relationships (correlation 0.51) and job performance (0.36). These are associations, not proof of cause and effect.
More surprisingly, two studies published in 2025, involving 646 people in total, linked observable listening behaviors—such as acknowledging a speaker and asking follow-up questions—to faster, not slower, conversational responses. The precise pattern varied by setting.
So no, empathy isn't a contest to see who can maintain the longest dramatic silence. A buffering video call would win.
The useful skill is knowing when not to jump in
For brainstorming, offer a tentative idea quickly. For important decisions, check the evidence. For personal conversations, resist writing the other person's ending before they've reached it.
Okada also discusses shifting a talk between addressing a whole group, speaking to one person and sounding as if you're thinking aloud. That's a presentation technique, not a scientifically established three-step formula.
The real lesson is simple: quick thinking and careful listening are different skills. You can have both. And “so basically” sounds a lot better after the other person has finished their sentence.
Research: Evans & Stanovich (2013) · Heitz (2014) · Bisra et al. (2018) · Melby-Lervåg et al. (2016) · Kluger et al. (2023) · West et al. (2025)
