Communication

Being understood is not transmission. It is reconstruction. Your sentence does not carry meaning across to another person; it is an instruction set that runs on their model of the world, and their model is not yours. What they end up with is something they built, using your words plus their assumptions, their context, and their guess about your intent. Every communication failure lives in the gap between the model you encoded with and the model they decoded with.

The mechanism

When you speak, you compress. You omit everything you consider obvious, because to you it is obvious. The listener decompresses using their own defaults for whatever you left out. If the defaults match, it works and both parties conclude they are good communicators. If they do not, both parties conclude the other one is careless.

Two well-established ideas describe most of this.

The curse of knowledge: once you know something, you cannot accurately model not knowing it. Experts systematically overestimate how much of their meaning survived the compression, and the effect grows with expertise, which is why the most knowledgeable person in the room is often the worst explainer in it.

Grice's cooperative principle: listeners assume you are being relevant, truthful, and appropriately informative, and they fill gaps with inferences generated by that assumption. This is why what you did not say is often louder than what you did. Omit the reason for a decision and the listener will supply one, and their version will usually be less flattering than the truth.

Get the reconstruction back

The only reliable way to know whether the rebuild succeeded is to inspect it. Ask the other person to say back what they are going to do, in their own words. Not "does that make sense" — that question tests politeness. The read-back tests the model.

Communication is an error-correcting protocol, not a broadcast. Design for the confirmation, not the send. Aviation and emergency medicine both figured this out and made read-back mandatory, because in those domains the cost of a silent decoding error is immediate and visible. In office work the same errors occur at the same rate and simply take three weeks to surface as someone building the wrong thing.

The corollary is that written artifacts beat conversations for anything consequential. A document can be re-read, disagreed with in the margin, and checked against later. A meeting produces N private reconstructions and no record of which one was right.

Where the common understanding is wrong

Clarity is treated as a property of a message. It is a property of a pair — a message and a receiver. The same paragraph is clear to one person and unreadable to another, so "I explained it clearly" is not a defensible statement without naming who it was clear to.

"I told them" is the most common false proof of communication in organizations. Telling is the send. It carries no evidence of receipt, and email and Slack are worse for this than speech, because they produce a timestamped record of the send that feels like a record of the understanding.

The last error is fixing misunderstanding with volume. Repeating yourself at length re-runs the same failed encoding. What actually moves someone from confused to aligned is a change of representation: a worked example, a concrete case, a drawing, a number, the specific bad outcome you are trying to avoid. If two attempts have failed, stop adding words and change the format.