A 20-Minute Cacao Listening Exercise for Two
This one is best explained backwards. Here is the exercise first, then why each rule is in it — because every rule below exists to defeat something specific about how conversation actually works, and the reasons are more interesting than the instructions.
The exercise
Two people, two cups of cacao, twenty minutes, no screens in the room.
- One person talks for eight minutes.
- The listener may only ask questions or say nothing.
- Swap for eight minutes.
- Four minutes of ordinary conversation at the end.
Three rules for whoever is listening. No stories of your own. Count to four before you speak. Hold one question in mind throughout.
Twelve to fifteen grams of cacao each, 240 ml. Most people break the first rule inside ninety seconds, usually while trying to be kind.
Rule one: no stories of your own
Sociology gives us a clean distinction here. A shift response moves the topic to yourself. "I had a terrible commute." — "Mine was awful too, there was a diversion at..." A support response keeps it with the speaker: "What happened?"
From the inside, both feel like engagement. A shift response often feels like the warmer of the two — you are relating, offering common ground, demonstrating that you understand. But it takes the floor, and done repeatedly the other person never develops anything past their opening sentence.
The sociologist who named this pattern called it conversational narcissism, which is harsher than most instances deserve. Usually it is enthusiasm, not self-absorption. The point is not to eliminate shift responses — a conversation made entirely of support responses becomes an interview — but to notice your ratio, which almost nobody has ever observed in themselves.
For these eight minutes, the ratio is zero. It is harder than it reads.
Rule two: count to four
Most people cannot tolerate more than a second or two of silence before filling it.
There is a reason that feels so uncomfortable, and it is measurable. Studies of turn-taking across ten unrelated languages found a median gap between speakers of roughly two hundred milliseconds — about the length of one short syllable — and the figure was remarkably consistent whether the language was Danish, Japanese, or spoken by a few hundred people in Papua New Guinea.
Against that baseline, a deliberate four-second pause is enormous. It will feel like a failure of the conversation. It is not.
The material worth having almost always arrives after the pause. The first answer to any question is the prepared one; the second, which only appears if nobody fills the gap, is usually the real one. Roughly a third of the time the person will simply start speaking again, and what comes next is more considered than what came before.
Holding a cup helps enormously here, for an unglamorous reason: taking a sip is a socially legible way to leave four seconds of silence. Without something in your hands, the same pause reads as coldness or distraction.
Rule three: give your spare capacity a job
People speak at roughly 125 to 150 words per minute. You can process speech considerably faster than that.
The surplus does not sit idle. It goes to planning your response, evaluating what is being said, drifting to something unrelated, or rehearsing the anecdote you are waiting to deploy. This is the listening gap, and willpower alone does not close it — telling yourself to concentrate harder gives the spare capacity nothing to do.
So assign it. For the whole eight minutes, hold one question: what is the single thing this person most wants me to understand? Not their argument, not the sequence of events — the point underneath. Holding that occupies the surplus and aims it somewhere useful.
It also generates better questions, because you end up asking about the thing that matters rather than about details you happened to notice.
Why the gaps will feel wrong
Worth knowing in advance, because it explains the strangeness of doing this properly.
If the normal gap between speakers is two hundred milliseconds, and planning even a simple spoken reply takes six hundred milliseconds or more, the arithmetic does not work — unless you are building your response while the other person is still talking. Which is exactly what people do. Fluent conversation requires predicting where a sentence is going and preparing against that prediction.
Meaning that ordinary, comfortable, fluent conversation structurally requires you to stop fully listening some way before the other person finishes.
That is not a character flaw. It is a design constraint of real-time speech. But it is why genuine listening feels effortful and slightly unnatural: you are suppressing an automatic process and accepting that you will not have a reply ready on time.
The diagnostic is simple. If your pauses have not got noticeably longer, you have not changed anything.
What this does to the person talking
The most interesting research here is not about the listener.
Experiments assigning speakers either high-quality or ordinary listeners found effects on the speaker's own thinking. Those who felt well listened to became less defensive about their views, described their attitudes in more nuanced and less extreme terms, and showed more awareness of why they held them.
Being listened to well appears to make people think better, not merely feel better. The proposed explanation is that defensiveness consumes attention — when you are not braced for judgement, more capacity is available to actually examine what you think.
This reframes listening from a courtesy into something with an effect. If you want someone to reconsider a position, listening is a more plausible route than arguing, and the mechanism is not manipulation but the removal of threat.
Doing it alone
Most of this transfers with a partner unavailable.
Listen to something spoken rather than read — an interview, a recording, a voice message someone sent you — and practise identifying the one thing the speaker most wants understood. Then notice what you were composing in reply while they were talking, because you will have been composing something. That noticing is most of the skill.
The four-second rule can also be practised in ordinary conversation without announcing it, and it is the single change most likely to alter how conversations go.
One thing that should make this easier
A finding worth carrying, because it lowers the stakes considerably.
Research on first conversations has repeatedly shown that people underestimate how much their partner liked them and enjoyed the exchange. The gap is consistent and reasonably large, and it has been named the liking gap.
Most people leave a conversation privately convinced they talked too much or came across badly, while the other person thought it went perfectly well. If that is a familiar feeling, the evidence suggests it is systematically wrong — and if you spend the eight minutes worrying about your performance as a listener, you are spending capacity that rule three has a better use for.
The hardest rule is the four-second pause, and almost everything else follows from it. For a group version rather than a pair, our facilitator guide covers holding a circle.
A practical note on the two-cup version: identical drinks matter more than good ones. Two 1 oz packets remove the small asymmetry of one person having made the better cup.
