They meant well, probably

"It's a pattern with them"

Chapter 7 · 2 min read

This is the honest objection, and it's sometimes right. One interruption is a slip. Ten from the same colleague, always when you're speaking and never when the director is, is a pattern. One forgotten birthday is a Tuesday. A brother who has never once remembered while expecting you to organise his is something else. The book isn't asking you to go blind. It's asking you to know what a pattern is.

A pattern needs three things. It needs the same act more than once, from the same person, with the situation varied. That last part is the one the charge skips. If a colleague always talks over you in the Monday meeting, but the Monday meeting is where everyone talks over everyone because it's twenty minutes for an hour's agenda, you don't have a pattern about them. You have a pattern about the meeting. If a friend has gone quiet for two months, and the two months are the two months since their father got ill, you don't have a pattern about you. Vary the situation before you call it a pattern, and look for the act in a situation that doesn't explain it.

There's a subtler problem with patterns, which is that they're built by the charge itself. Benjamin Karney and Thomas Bradbury followed newlywed couples across eight waves of data and found no evidence that people have a fixed style of explaining their partner's behaviour. The explanations moved with satisfaction. As spouses grew less happy, they explained the same acts less generously, and the harsher explanations in turn predicted falling satisfaction. So the pattern you see in someone is partly the accumulated charges you've made about them. Once you've decided a colleague is dismissive, every short reply gets filed under dismissive, and the file grows, and the thickness of the file feels like evidence when it's mostly filing.

So here's a test. Before calling it a pattern, count the instances you can actually name, with dates, and for each one write the situation next to it. Then look for one instance where the situation doesn't explain it. If you can't find one, you haven't got a pattern. If you can find three, you might, and chapter 18 is about what to do with it, which is ask, not conclude.