Appearance
10.8 — Judging a Person: When To, and When Not To
Two pieces of advice circulate and they contradict each other.
"Do not judge." Which, taken literally, is impossible and would be dangerous. You have to decide who to trust with your money, your children, your evenings and your safety.
"Trust your instincts about people." Which, given 10.1 and 10.6, is advice to rely on a confident and largely inaccurate instrument.
Neither is usable. The distinction that resolves it is between two different operations that share one word.
Two things called judging
Assessment. Working out what someone is likely to do, so you can decide how to act. Is this person reliable? Will they keep this to themselves? Should I lend them money?
Condemnation. Deciding what someone is worth as a person. He is a bad person.
The first is necessary, improvable, and should be done carefully. The second is almost never necessary, is not improvable, and mostly costs the person doing it.
Almost every argument about judging people dissolves once you separate these. You can assess someone as unreliable — and act accordingly — without concluding anything about their worth. The assessment is what protects you. The condemnation is the part that does nothing except keep you angry (5.4).
When assessment is required
Do it deliberately, not by instinct, in these situations.
Anything with your safety in it. Physical safety, and financial safety.
Anything with a long commitment. A business partner, a marriage, a shared lease, a co-founder. The cost of a wrong assessment scales with how hard it is to exit.
Anything involving people who depend on you. Who looks after your children, who your elderly parent relies on.
Anything where you are relying on someone's follow-through.
And anything requiring confidentiality.
Where none of those apply, you frequently do not need the judgement at all, and making it anyway is where most of the error and most of the unpleasantness enters.
How to assess well
Six rules, and they are all about evidence rather than perception.
One: behaviour over words. What someone says about themselves is nearly worthless as evidence, including the self-deprecating parts. What they do, repeatedly, is the data.
Two: behaviour when it costs them something. The strongest single indicator. Anybody is generous, honest and patient when it is free. Watch what happens when honesty is expensive, when helping is inconvenient, when admitting the error has a consequence. One observation of that is worth twenty of the cheap kind.
Three: how they treat people who cannot do anything for them. Waiters, junior staff, cleaners, strangers, the person who made a mistake. This is the most reliable indicator of character available in ordinary life, and it is available constantly and free.
Four: patterns, not instances (10.5). Three occasions, across different situations, over time.
Five: what other people's experience has been — used carefully. Second-hand reports carry the reporter's biases and are frequently the only way to learn about a pattern you will not personally see for years. Weight them by whether the reporter has a stake.
Six: consistency between the accounts they give. Over months. Nobody maintains an invented version perfectly (10.6).
The specific things worth assessing
Not "is this a good person", which is not answerable. Specific, and each one is separately true or false.
Reliability. Do they do what they said, by when they said?
Honesty under cost. Do they admit the error that will be expensive to admit?
Reciprocity. Does the effort run both ways over time, or only towards them?
How they handle being wrong. From 5.7, the single most informative thing about a person you will work with. Concession, or deflection?
How they handle someone else's success. Genuine pleasure, or a subtraction?
And how they behave under pressure. Everybody is fine when nothing is at stake. Character is what remains when the conditions are bad, and you will not know until you have seen it.
Assess each separately. People are inconsistent across domains: someone can be scrupulously honest and completely unreliable, or generous and unable to keep a confidence. A single overall verdict throws away the useful detail.
Why condemnation costs you
Three reasons, and none of them is about being nice.
It stops the assessment. Once someone is written off as a bad person, you stop collecting information, including information that would have been useful.
It keeps you activated. Condemnation requires maintenance — rehearsing the grievance, re-establishing the verdict (5.4). The other person is not paying for this.
And it is frequently wrong about the situation (10.5). The middle layer explains more behaviour than character does.
None of this means excusing anything. You can hold someone fully responsible for what they did, remove them from your life, and decline to conclude that they are a bad person. Those are compatible, and the combination is both more accurate and considerably cheaper to maintain.
Where the line is
Assessment stops being enough in a small number of cases, and it should be said plainly rather than lost in nuance.
Some behaviour tells you what you need to know in one instance. Violence. Cruelty to someone who cannot defend themselves. Deliberate humiliation. Betrayal of a serious confidence for advantage.
The pattern rule does not apply here. You do not need three instances of someone hitting you to know something (13.6).
And there is a class of person for whom the ordinary assumptions fail — where the warmth is instrumental, the apologies are tactical, and the pattern only becomes visible over years. Part 13 is written for that, and the diagnostic there is behavioural and specific rather than intuitive.
Judging yourself, and the double standard
You judge yourself by your intentions and other people by their effects. Universal, and correcting it in both directions is what fairness actually consists of.
Two corrections.
For others: consider intention. Ask what they were trying to do. Frequently it was not what happened.
For yourself: consider effect. I did not mean it that way is true and does not undo what landed. The effect is the thing that requires the apology (4.8), and the intention is the thing that determines whether you need to change something structural.
What to do with this
Separate assessment from condemnation, out loud, when you catch yourself judging. I am deciding whether to rely on this person is a useful operation. I am deciding whether they are a good person is not one you need to complete.
Assess only what you need to. Most people in your life require no verdict at all.
Watch the two high-information rungs: how they treat people who can do nothing for them, and what they do when honesty or decency costs them something. Both are visible weekly and almost nobody is watching.
Assess the specific dimension rather than the whole person. Reliable but indiscreet is a useful thing to know and a global verdict destroys it.
And where safety is involved, do not wait for the pattern. One instance is enough, and everything else on this page is about the ordinary cases.
Next: 10.9 gives the disciplined method that replaces the hunch — how to work out what is actually going on with a person, from evidence, the way an investigator would.