What it does to the row

There's one more cost, and it's to the row itself. The fresh case doesn't just fail to end the row. It trains it.

Each instance handled as a new case makes the next one more likely and more automatic. The path is worn a little deeper. Two people who've walked the same line across a lawn a hundred times don't decide to walk it the hundred-and-first time. Their feet know it.

Look at what each fresh case adds. It adds an entry to the record, so the next instance has more to open. It confirms the pattern from chapters 8 and 9, so the pusher pushes a little sooner and the quiet one goes quiet a little sooner, because both of them know what's coming. It raises predictability, which Johnson and Roloff found went with hopelessness about the row, so the next instance arrives with more dread attached. And it leaves the underneath untouched, so the underneath is exactly where it was, and nearer the surface, because the last instance didn't drain it.

That's what I mean by the row getting better at itself. After enough fresh cases, an instance can run from the first sentence to the thing from March in under two minutes, with neither of you having decided anything, because every step has been rehearsed. That speed feels like the row being deep-rooted. It's the row being practised.

The good news is buried in that. A row that's been trained can be retrained, and it doesn't take fifty instances, because the training in the other direction is deliberate rather than accidental. The name from chapter 20 is a new first step, and it's put in on purpose, at the point where the old path starts. The first few times you say it, it'll feel odd, and your feet will want the old way. By the fifth or sixth instance, the name arrives on its own. Chapter 23 says what to expect, and this is my expectation rather than a finding, but the mechanism runs the same way in both directions. Whatever you do at the first sentence, fifty times, becomes what the row does.