A quick answer is evidence that the person answered quickly

Let us begin with the number everyone has repeated incorrectly, though this time it is not a number I can replace with a better one. My answer is that a first week tells you what a person does when the circumstances are small and still unfamiliar; it does not tell you, by itself, what that person will do when the cost is high.

I was wrong to treat the first impression as a draft of the final account. A person who speaks readily may be generous, or merely practiced at speaking; the quiet one may be cautious, tired, shy, or waiting to find out what kind of room this is. I once gave too much weight to confidence, as though a quick answer were evidence of a settled mind. It is not. A quick answer is evidence that the person answered quickly.

The useful evidence is narrower, and usually less grand. Someone makes a small mistake in an arrangement: do they correct it, hide it, or leave another person to carry the cost? A plan changes at the last minute: do they explain what changed, or pretend the original plan was never made? These are not verdicts. They are entries in the margin, with the date and the circumstances written beside them, because a record without its conditions is only a way of making one’s own guess look official.

Even then, the conditions matter more than I used to admit. A person who is late because a child is ill has not shown the same thing as a person who is late because everyone else’s time seems cheap. The act is the same; the facts around it are not. I have sometimes mistaken a circumstance I did not know for a character flaw I thought I recognized. That is a bad kind of arithmetic: the sum may be neat, but one of the figures is invented.

What I still do not know is how much of a person can be seen before trust has had time to make them less guarded. Nor can I tell, from a handful of ordinary moments, what someone will do under pressure they have never faced. There are people who are kind when kindness costs nothing, and people who become kinder when it does. The first week may offer a clue; it cannot supply the test.

So I have landed on something less satisfying than a rule, but more honest: notice what happens, write down what else was happening, and resist making a permanent judgment from a temporary arrangement. Old books are full of confident conclusions whose missing pages would have changed them. I would rather keep the blank space than fill it with a convenient lie.

Next I want to write about the lord who waits because he has counted the danger, and the lord who calls his fear good judgment. Telling caution from fright is a distinction worth examining before either one gives an order.