I thought about how to use comments better and get along with people better. I wrote notes about it:
Write comments that would be appreciated by a new observer, who hasn’t read any of my previous stuff, hasn’t read the sequences, and only skimmed the post he’s commenting under.
Only reply to comments I actually think are good. If I see any signs of low quality, hostility, or social aggression, don’t reply.
Make it clear to people in my bio, and at the end of some posts, that I’m open to more discussion by request. I can change policies and be more responsive if asked.
Avoid meta discussion. Lots of LW people don’t like it. I think meta discussion is very important and valuable, but I can write it at my own forums.
I plan to have two clearly distinguished commenting modes. I think a middle ground between them causes trouble and I want to avoid that.
Mode one is anything goes, zero pressure to reply, drop anything on a total whim with no explanation. This mode will by my default and means I’ll be replying less than I was.
For these comments, I’ll try to make most of my comments standalone and interesting. That means only engaging with people who say something significant and worthwhile. Otherwise I can write a monologue “reply” (that doesn’t engage with specifics of what they said, just talks about the topic) or not reply.
A short answer like “Yes” is OK too because it won’t annoy any readers. People who don’t get value from it will know it’s contextual and won’t mind.
It’s important to be careful about comments which rely on context but don’t obviously do so. People can think those comments are meant to stand alone when they aren’t. So try to make comments really blatantly be minor followups or else offer standalone value.
Mode two is high effort discussion after some mutual agreement to try to use some sort of written rules. Examples of discussion policies that could be used:
Discussion to agreement or length 5 impasse chain. Agreement can be about the topic or agreeing to stop the discussion – any sort of agreement is fine.
Discussion until agreement or someone claims that they made an adequate case that the other person ought to be able to learn from and be corrected by. They believe their case would persuade a neutral, reasonable observer. Plus minimum two followups to address parting questions (like which text constitutes the adequate case, and isn’t the case inadequate due to not addressing questions X and Y that were already asked?) or potentially be persuaded to change their mind about ending there.
Discussion to agreement or to one stated impasse plus two followups to address final questions and have a brief window to potentially solve the impasse.
The discussion mode I do not want is a medium effort discussion following unwritten rules (particularly social hierarchy climbing related rules). I prefer either small or large discussion. Either anyone can leave at any moment with no negative judgments or we set up a more organized, serious discussion project. I don’t want to half-ass adding transaction costs and commitments into discussions. Do that as a big enough deal to agree on and write down some discussion rules and policies, or don’t do it and stick to anarchy. Unwritten rules suck so either use written rules or no rules.
I don’t trust people to be OK with no-commitment discussion, despite having recently been told by a several people that that’s how LW works. I think LW mostly works by medium commitment discussion where there are social pressures. I think people routinely are judged for not replying.
It’s hard to deal with because asking people if they want a serious discussion, in reply to their topical comment, gets negative responses. They don’t want to state what sort of discussion they want (zero commitment, medium commitment with unwritten rules, or more serious project with written rules). I take people’s general dislike of stating what rules they are playing by, or want me to play by, as a piece of evidence that it’s unwritten rules and medium commitment that are commonly desired. I don’t think people are usually really 100% fine with me ignoring them an unlimited number of times without explanation because, hey, no commitment and no request for anything different. I think they’ll see me as violating unwritten rules saying I should be somewhat responsive. (I personally wouldn’t mind explaining why I think someone’s comments are bad and why I don’t want to reply, but I think the LW forum generally does mind me doing that. If people want to know that they are welcome to ask me at my forum about particular cases from LW. But I don’t like being asked privately because I want my regular audience to be able to see my answers.)
Broadly if anyone has any problem with me, I hope they’ll state it or ask a direct question about it. I don’t expect it but I do prefer it.. I know people often ask for such things and don’t mean it, but I have an 18 year track record of public discussion archives showing I actually mean it, and I run a discussion community where such things are normal. Posts like “Why did no one reply to this?” or “Why didn’t you reply to this?” are well within norms at my forums, and I do get asked meta questions like why I dealt with a discussion in a particular way.
I don’t plan to ask other people at LW direct questions about problems I have with them, or state the problems, unless they ask me to do that and I find the request convincing (e.g. I can’t find signs of dishonesty or incompetence). Even then, I might ask them to forum switch first because I think other people at LW could easily take that discussion out of context and mind it (the context being the convincing request).
I would like some discussions where we try to make an organized, serious effort to seek the truth, resolve some disagreements, etc. But that is not the LW norm. The LW norm is mostly a mix of small and medium discussions. OK. My solution is to make big discussions available by request and otherwise do small discussions. This should be acceptable for both me and others.
I thought about how to use comments better and get along with people better. I wrote notes about it: