Takeaways: Lessons learned during the session

The Difference Between a Real Problem and a Feeling

     We began with a familiar scene: as one person raises their voice, the other goes quiet. The silence is perceived as agreement — most people recognise some version of this. The relevant standard is closer to Popper's than to the usual call for more communication: a claim only counts once it is precise enough that it could be proven wrong. "We need to communicate better" fails that test instantly. It stays true no matter what happens next, which means it explains nothing. "I go quiet when your voice rises, not when I disagree with you" passes. The other person can check it against what actually happened and say no. That is the difference between a real problem and a feeling being described as a problem.

     Musk's design sequence supplied the discipline for our session: question the requirement, delete, simplify, accelerate, automate — in that order, never inverted. The instinctive fix, such as a weekly check-in or a resolution to be more open, automates the steps for requirements that were never questioned. Once the real requirement is identified, the fix for the named situation is small: one person names it out loud when they have gone quiet under pressure, the other lowers their voice and asks again.

When Specificity Runs Out

     The above-described vague version of resolution is common: one vague sentence distracting from a specific truth. A slightly different, but just as common, failure mode is harder to see and cannot be solved by specificity alone. It is also far more expensive. Teams routinely spend full quarters in workshops built around a single shared word and leave with three unrelated problems half-solved. It is a careful, well-intentioned effort aimed at nothing in particular.

     The session ran directly into that second mode. One participant proposed climate change as the next problem to scope. The instinct was to do what had just worked with the communication example: cut it down, find the one specific claim underneath, identify the actual issue. It did not work, and it took several failed attempts to see why.

Climate change was not hiding one truth. It was several different concepts using the same word.

     A physical measurement, a policy argument, a marker of identity, a marketing category, a personal ethical dilemma — and no amount of digging into any single one produced the others. We did address the definition on time, just to discover three more layers of the same issue hiding underneath, which kept us in an extended loop that appeared to resist the narrowing altogether.

How Many Watches

     Herbert Simon's 1962 essay "The Architecture of Complexity" opens with a parable about two watchmakers, Hora and Tempus, both building watches of a thousand parts. Tempus assembles his as one continuous process. If interrupted, the unfinished watch falls apart, and he starts over. Hora builds in small, stable groups that combine into larger groups and then the whole. If interrupted, he loses only the small group he was working on. Simon calculated that with a one-percent chance of interruption, Tempus takes on average four thousand times longer. The lesson is that a system tolerates being cut only if it has real seams. Cut across a seam that is not there, and you do not get two smaller problems. You get one broken problem.

     Climate change is not a classic Gallie case of one deeply value-laden concept whose proper criteria are endlessly disputed. Unlike "art" or "social justice," where the competing values genuinely never resolve, the concepts here are separable once named. It is a polysemantic term: several distinct problems that have come to share one name. The practical consequence is the same as Simon's parable taken one step earlier: before you look for seams, you have to establish how many watches you are actually holding.

One Real Seam, Not the Whole Word

     The participant who raised it found his own specific problem within seconds: he wants to act sustainably and wants to stay frugal, and those two goals collide daily in ordinary purchasing decisions. That is checkable, specific, and his to solve. Finding one real seam did not collapse the rest of the word into agreement. It just proved that the small personal version and the large contested version were never the same problem to begin with.

     He offered a large-scale answer next: living off the grid. The room caught the mismatch immediately. Off-grid living sounds like the personal-scale answer, and to an extent it solves the daily purchasing dilemma; however, depending on the specifics, it can produce a worse outcome for the climate than staying connected. A solution built for one of the concepts under the word can look plausible while doing nothing for, or actively working against, a different concept sharing the same name.

     Three more people proved the same point from three directions, and the disagreement was worth observing rather than resolving. One argued the real issue was overconsumption, framing it as restraint. Another argued demand is not the problem; delivery is, framing it as innovation. A third argued for education and fewer top-down restrictions, framing it as an individual method. Neither was wrong; they were simply addressing a different version of the concept, with a contradictory underlying issue, while calling it by the same name.

     Call this the multiplicity check. The operational test is simple: if two competent people can both be right, agree on every available fact, and still be talking past each other, you are probably holding more than one watch. The usual advice — just get more specific — assumes there is one system with seams waiting to be found. That advice worked for the communication example. It failed here, and it fails wherever a single name is carrying several independent problems.

The Same Failure, Institutionally

     The multiplicity check already supplies the base. Simon’s test makes it structural: a seam exists where a piece can be removed and the remainder continues to function exactly as before. Applied to the three positions that emerged under “climate change,” each can be subtracted without the other two losing coherence or force. That is the signature of separate systems, not of competing sides inside one argument.

     The same logic has a wider relevance. Whenever a professional term is carrying more than one concept, remove one interpretation and check whether the remaining disagreement still stands on its own. If it does, the word was never naming a single object of work. It was naming several that had been allowed to share a label.

When the Narrative Wasn't There

     Something felt off, and it took most of the session to understand why. I hadn't built a real narrative for this session, just a rough outline and a set of theories I liked. Without an approximate guide, the facilitation decisions were slower and less even than usual.

     I've noticed a version of this before: staying with one contribution past the point where it's still serving the group. It's noticeable when someone simply contributing is suddenly put in the spotlight. Edgar Schein's distinction between genuine and diagnostic inquiry describes this behaviour: a facilitator stops building shared thinking and starts extracting an answer to a fixed agenda. The motivation toward it is simple: a good, specific answer feels like progress, and without a structure telling when a point has been made, the instinct is to keep pulling. The same test the session ran on "we need to communicate better" applies here, turned on myself. Was I still exchanging, or interviewing?

Decide in advance what would count as enough — a specific point reached — or you will keep pulling past the moment it serves the group.

What the Chaos Revealed

     The looseness wasn't only a facilitation flaw. Only in reflection did its actual value become visible: the improvisation that felt like weak preparation was also the condition that let the session's real finding surface. Karl Weick's work on sensemaking helps understanding this. Meaning, in his account, is often made after the fact, built from action taken without a clear plan, rather than the plan producing the action. You don't know what you think until you see what was said, and in this case, what several people said, once examined afterwards, revealed a pattern a more scripted session would likely have closed off before it appeared.

     This isn't a license to keep preparing loosely. The same absence of structure that produced a moment of discomfort also produced the session's central insight, and there's no evidence yet that one requires the other. An earlier piece, Designing Emergence, made a version of this argument: that over-specifying a system removes the slack that lets the right thing grow inside it. This session is a concrete instance of it. The open question for next time isn't whether to prepare more or less. It's whether the specific pattern, staying too long with one contribution, can be corrected without closing off the same looseness that let the unexpected insight occur.