Listening is not agreeing

The first: listening to the end would mean conceding the point. Whoever fears that interrupts, corrects, objects, because they believe silence would amount to assent.

The second: speaking would mean requesting an action. Whoever sets out a difficulty is answered with a solution, when they only wanted to be followed. The two misunderstandings often meet in the same conversation, which quickly makes it painful.

Restating consists in giving back what you understood, in your own terms, before answering. The operation has been taught in interview training since Carl Rogers's work in the mid-twentieth century, and it is often caricatured as mechanical repetition.

Its real function is to make an understanding verifiable, and therefore correctable. It amounts to approval of no kind, and it has the advantage of surfacing gaps before they turn into substantive disagreement.

Proposing a solution is a costly answer for whoever receives it: it presupposes that the problem is simple, that they had not thought of it, and it shifts the conversation towards action at the moment when they were still trying to describe.

It is not that advice is useless. It is that it almost always comes too early, and that a question as banal as "do you want help working out what to do, or do you just want to tell me?" defuses most tensions of this kind.

A person who finely perceives another's signals does not thereby identify their cause any better. Work on empathic accuracy, developed by William Ickes, shows that confidence in one's own reading is not a good indicator of its accuracy.

The most effective listening is therefore not the most intuitive: it is the kind that checks. Asking a closed question at the end of an account, of the kind "was it more the workload or more the way it was said to you?", brings in more than ten minutes of silent inference.

A disagreement expressed after full listening is not the same object as an objection raised along the way. The first bears on an identified position, the second on a supposed one, and one then spends the rest of the exchange correcting the misunderstanding rather than handling the question.

It is the only rule that emerges clearly from the literature on difficult conversations: the order of operations counts for more than the content. Understanding then contesting costs less than contesting and then discovering you had misunderstood.