It seems like social norms everywhere are structured to protect people from negative feedback, because we’re wired to take it really personally. See “compliment sandwiches”. It also feels that our sensitivity isn’t learned, it’s innate. (This is debatable but I strongly believe it is, so I’ll take it as a premise and move on.)
So we only get a small fraction of negative info, react to negative info really strongly, and have to couch it very gently if they decide to say it to others1. This seems odd when I look at it directly. Why are we in the hypersensitive equilibrium in the first place? It introduces so much inefficiency for the group. As a friend put it:
the expectation is that you downgrade your negative feelings by X% before transmitting them, and the receiver knows this and will try to inflate them by Y% to get a sense of how you really feel, but X and Y are disjoint and the incentives do not encourage sharp alignment between them
A different friend thought that “caring a lot about what your tribe thinks of you, since popularity was a matter of life and death” is a good-enough basic explanation, but I don’t think it explains the degree to which the modal person is driven insane by clearly seeing how much they irritate or inconvenience others. That degree seems actually bad for the individual as well as the group!2
My low-confidence best guess is that if you’re a monkey who has just learned to talk, much of the negative social feedback is bad faith (“grug too selfish with fruit”), so it makes sense to be quick to react aggressively to that. The frequent aggressive response raises the threshold for saying “you’re selfish”. 1000 generations later, when you hear “you’re selfish”, it’s likely that either the hominid is trying to start shit with you, or you’re so selfish that someone just had to say it, both of which makes sense to have a strong emotional reaction to. Once people enter the high-threshold-for-saying-something equilibrium, it’s hard to leave it, even after they exit the high-scarcity, high-bad-faith conditions that produced the equilibrium in the first place.
Higher-confidence secondary factor is the asymmetric risk: if you’re like 1SD less neurotic than optimal-on-average, that’s probably worse than being 1SD more neurotic than optimal, so it’s better to aim for 0.5 SD more neurotic than optimal than exactly optimal.
Footnotes
-
A curse upon event hosts who have an attendee who is ruining the vibe, but not so badly that they feel up for the risk of communicating directly with the person about their behavior.) ↩
-
People who are chill when told to cut out an annoying behavior are much nicer to be around than people who freak out, and seem likelier to be pulled onto the metaphorical raft in winnowing situations. ↩