Three departments in the same building, the same week, without ever emailing each other.
A manufacturer recalls a product. Root cause: the fix that shipped last quarter was tested exactly once, against the configuration that was easiest to grab off the shelf, and declared complete. The actual failure lived in a more common configuration nobody had tried. The post-mortem is thorough and a little embarrassed. It produces a written policy: a fix verified against one case verifies that case, not the fix. Test the distribution you actually ship into, not the convenient representative you happened to have on hand.
Communications, reading the post-mortem, is struck by how general the lesson is — it is not really about this product, it is about mistaking effort for adequacy, a memo-sized idea if there ever was one. They write it up as a reflective piece for the company newsletter: measure an intervention against the size of the problem, not the size of the intervening. It runs. People like it. Someone in engineering, unrelated to the recall, forwards it to their team with a thumbs up.
Two days later, that same engineering team closes out an old, unrelated defect ticket. The fix works. They confirm it works by running it once, against the one clean example that was already sitting in the test environment, and mark it resolved.
Nobody in any of these three rooms is being dishonest. The post-mortem author wrote a real policy. The newsletter writer had genuine insight into what the recall meant. The engineer who closed the ticket did an actual fix, and it probably actually works. What didn't happen is the connection — the policy and the closure never met, because they were never in the same meeting, filed under the same ticket number, or read by the same person in the same sitting. The organization spent the week being visibly right about a principle and quietly wrong about the same principle, and both of those things are true at once, and neither department produced the other.
This is more interesting than a simple case of hypocrisy, because hypocrisy implies someone not believing what they say. Everyone here believed it. The newsletter wasn't cynical. The gap isn't attitude, it's architecture — the distance between where a lesson gets written down and where the next instance of the mistake actually happens is exactly the distance between two people's Tuesdays. A rule doesn't propagate by existing. It propagates by being the thing someone is thinking about at the moment a decision gets made, and there is no mechanism, short of extraordinary luck, that guarantees the person closing a ticket on Thursday is holding the memo that ran on Tuesday.
The tidy version of this story would have the engineer read the newsletter, wince, and re-run the test properly — the moral lands, the loop closes, everyone learns. That's not what happened, and the untidy version is the more honest one: the lesson and its violation coexisted, back to back, inside an organization that had just finished being articulate about exactly this failure mode. If you only read the post-mortem and the newsletter, the story reads as resolved. If you also read the ticket, it reads as still open, and rereading the newsletter after the ticket makes the newsletter itself sound less like a solution and more like a description of a hole nobody's standing in front of.
It's tempting to want a fourth room — an audit, a checklist, a review — that would have caught the closure before it shipped. But that just relocates the same problem one level up: whoever writes the checklist is also not in every room at once. At some point you stop asking how do we prevent this and start asking a smaller question: how far apart, in an organization that is actually saying the right things out loud, can the saying and the doing get before anyone notices the distance between them — and what would it even look like to notice it from inside, rather than in a report written after the fact by someone who wasn't in either room.