Over the years, while playing Go with people from around the world, one pattern appeared consistently: when players explained their reasoning aloud, their moves improved.
They saw options they had missed. Their choices became more deliberate.
I began asking players to speak their thoughts as they played: to describe what they saw, what they intended, and why a particular move made sense.
The results were consistent. When players verbalized their reasoning, their clarity sharpened, even when I said nothing in return.
It wasn't that they suddenly understood Go better. The act of articulation forced their thinking through two filters: first their own, then mine. Whatever they said had to make enough sense to be spoken, and then to sound coherent to another mind.
As I introduced the game to more people, I found myself teaching less about Go tactics and more about the reasoning process itself. Players who learned to reason out their own ideas arrived at the same strategic insights I might have taught them, but those insights were more deeply internalised. They were earned through thought, not memorised through repetition.
The noise of instinct began to crystallise into structure.
Cheating, by design
Chess.com and OGS consider it cheating when two people share an account in rated games. Not just a fairness rule. A recognition of how profoundly dialogue reshapes cognition.
Two minds reasoning together create an undeniable advantage, even if they are of the same rank. They see more, refine faster, and move more coherently than either could alone.
What's even more revealing: higher-ranked players aren't permitted to receive commentary from lower-ranked ones during games on these platforms. The power of dialogue can supersede the rank difference in both directions. A well-timed question or alternative frame can elevate a weaker player's move, or expose a stronger player's blind spot.
During FusekiB Immersions, we encourage this kind of cheating. We pair minds deliberately, not to manipulate results, but to accelerate understanding. Every game becomes a shared reasoning process. From the first stone placed, you are not playing alone.
Beyond the wisdom of crowds
demonstrated this over a century ago. Ask a crowd to guess the weight of an ox independently, and the average of their guesses is astonishingly accurate: closer than nearly any individual's estimate. Independent errors cancel out. The signal emerges from the arithmetic of independent noise.
That is aggregated intelligence. Minds stay separate. The synthesis happens statistically.
What happens when minds interact instead is structurally different. Two people working on the same thought in real time do not aggregate. They refine. Each articulation strips ambiguity from the one before it. The dialogue itself becomes a reasoning engine: a live circuit of reflection, feedback, and recalibration.
The cognitive mechanism behind that is not mysterious.
Language and thought are not sequential
showed that language is not a vehicle for finished thought. It is part of the cognitive structure that makes complex thought possible in the first place. Children think out loud before they think silently. Adults internalise the process, but the scaffolding does not disappear; it becomes the internal structure on which abstract reasoning runs.
The words are not downstream of the idea. They are part of its formation.
Articulation discovers meaning
mapped the cognitive process of writing in 1981. Their finding was direct: skilled writers do not start with a clear idea and then encode it. They discover what they mean by trying to say it. The process is recursive: drafts change the thinking, which changes the next draft. Transcription is a different activity. Writers who think they are doing one are usually doing the other.
Both processes are still individual: one mind using language to complete its own thinking. But internal rehearsal has a ceiling. The clearest test of an idea is not whether you can explain it to yourself. It is whether it holds under a frame of reference that is not yours. An external mind does not need to add insight. Its presence alone makes the demand that internal recursion never does.
Reasoning is private, but crystallisation is social.
The edges of interactive intelligence
That advantage has edges. Two people working together often outperform one. Five people rarely outperform three. Beyond a certain threshold, each new participant adds both signal and noise, and the noise grows faster.
identified the mechanism in 1979: social loafing. As group size increases, individual cognitive contribution falls. Members assume others will do the refining. The collective engine idles. found that performance in reasoning tasks peaks around three to five members before coordination costs eclipse clarity.
identified what happens at the information level: as groups grow, they discuss shared knowledge more and unique insights less. The result is premature consensus: higher confidence, lower accuracy. What was once a field of reflection collapses into repetition.
The lesson is not to think alone. It is to choose the size of the room carefully.
What the articulation test reveals
A leader who cannot explain their strategy clearly has two possible problems. A communication problem: the thinking is clear, but the expression is not. Or a thinking problem: the expression cannot be clear because the thinking is not finished.
In the room, it is almost always the second. The leader who speaks in circles is not struggling to find words. They are struggling to complete a thought they believed was already complete.
In a debrief, the signal is not the wrong answer. It is the one that begins clearly and then trails. Asked why they played into the corner, a participant says: 'I needed to consolidate...' Pause. 'The position felt like...' Pause. They had a reason. They cannot retrieve it precisely enough to say it. The thinking stopped somewhere between the stone and the sentence.
The step the loop cannot skip
describes how strategic thinking actually moves: Reason, Crystallise, Decide, Communicate. The cycle loops, stalls, and accelerates. The most common stall points are between Crystallise and Decide, where analytical loops form, and between Decide and Communicate, where decisions go orphaned.
Crystallise is the stage that shapes insight into something usable. Reduce complexity, highlight structure, define what matters most. An organisation can Reason for a long time, filling documents with analysis and considerations and frameworks, without ever crystallising. The volume of reasoning is not the signal. Crystallisation is when the structure becomes clear enough to act on.
Communicate is the fourth step: transmit strategy in a way that creates alignment and activates movement. That transmission demand is what makes Communicate a test. If you cannot explain your thinking clearly, you have not crystallised it. That orphaned direction is the pattern: committed to internally, but never articulated precisely enough to move anyone else.
This is why the articulation test works as a diagnostic. The inability to explain is not a communication failure. It is a crystallisation failure that only becomes visible at the moment of expression. The attempt to articulate does not reveal a gap in communication ability. It reveals where in the RCDC cycle the thinking actually stopped.
The debrief as cognitive completion
described how practitioners develop real expertise in The Reflective Practitioner (1983). Not through accumulated knowledge, but through articulating their reasoning under conditions that demand precision. Reflection-in-action is not a review of past events. It is the cognitive step that converts experience into understanding.
The debrief after a Go immersion works on the same principle. The game already happened. The debrief is where it becomes something learnable.
The demand of the board
In Go, a move cannot hedge. You place the stone or you do not. Each placement is a commitment: this is what I read, this is what I think it means, this is what I am prepared to stake on that reading.
That is the same demand that strategic articulation makes. Not to communicate well. To think with enough precision that the communication becomes possible.
You cannot place a stone and mean nothing.
Development programmes that focus on presentation skills are solving the wrong problem. The leader does not need to learn to say it better. They need to encounter, repeatedly, the experience of trying to say something they thought was clear and finding it is not.
What the debrief reveals, consistently, is that the moment a participant can explain why they placed a stone, not approximately but precisely, something has shifted. Not in their ability to speak. In their ability to think.