Month: January 2016

In Which a Facebook Comment Becomes A Post

Epistemic Status: Interesting model I’ve been tossing about for a while and too lazy to write down. Brienne was looking for blog ideas, and I realized that the comment I left was, while not well-written, complete enough to share.

When someone talks about a person being “good”there seems to be two separate ideas they could be referring to, which often get conflated. Let’s call them System 1 and System 2 morality. System 2 morality is when you are given a large formalized issue and have time to think about it. Examples include Effective Altruism or political issues (eg abortion).

System 1 morality is the day to day interactions that generally consist either of following politeness heuristics or running models of other people in your head. By “politeness” I don’t necessarily mean saying “please” and “thank you”, but more important and basic things. For example, if you’re dropping off an item at your friend’s house and are running two hours late, you should text to let them know.

In Parfit’s Hitchhiker, System 1 moralists pay the driver because that’s the not-asshole thing to do. System 2 moralists invent Timeless Decision Theory. You want lots of System 2 moralists to exist in the world. You want lots of System 1 moralists to exist in your social circles.

Rationalists are often very good at System 2, but not so great at System 1. A problem arises IMO when they decide to drop politeness heuristics for “tell culture” WITHOUT particular skill or care for running models of other people. How this plays out for me is that I’m constantly needing to either set new boundaries (which should be obvious to anyone running even a shitty model), or be okay with my boundaries always being broken. Usually the transgressions are so obvious that I don’t think I have to say anything until it’s already been transgressed. (Like, I shouldn’t have to say “btw, if you happen to be running two hours late you should text me so that I know that.”)

The level of explicitness for each possible boundary is draining. For example, let’s say I have a fancy coffee table and someone puts a cold drink on it without a coaster. No big deal. I ask them to use a coaster to protect the table and they agree. But if one minute later they put their muddy boots on that same table, this is a failure on their part. Likewise, no one wants to list out every potential failure mode. “Don’t wake me if I’m sleeping, especially if I have work the next day. Don’t publicly share my private information. Leave a respectable tip, even if you’re never coming back to the restaurant” And so on.