But it’s clear that some of what I said was heavily downvoted because I took a stance people didn’t like. Saying things like “Yep, I could have phrased this post in a more epistemically accurate way… but for this post in particular I really don’t care.”
Well, that particular comment had a lot of other stuff going on, and yes I think it’s a kind of comment that doesn’t belong here and no I don’t particularly feel like explaining that.
But also, yeah, I do kinda feel like “downvoting people when they admit they did something bad” is a thing we sometimes do here and that’s not great incentives. If someone wants to avoid that kind of downvote, “stop admitting to the bad thing” seems like an obvious strategy. Oops! And like, I remember times when I asked someone a question and they got downvoted for their answer, and I did think it was a bad answer that in a vacuum deserved downvotes, but I still upvoted as thanks for answering.
I’m not sure it’s so bad though. Some things that mitigate it as a strategy:
“This person strategically fails to answer certain questions” is a thing it’s possible for someone to notice and point out.
Someone might not have realized the thing they did was bad-according-to-LW, and the downvotes help signal that. (Maybe better to instead upvote the admission and downvote the thing they did? But that’s not always a thing that can be downvoted, or downvotes might not be specifically targetable to make it clear “this thing you did was bad”.)
If someone did a bad thing and doesn’t care, maybe we just don’t want them here. Downvotes probably marginally push them away, as well as marginally push them towards not-admitting-things. Notably, I feel like we’re more likely to downvote “I did a bad thing and don’t care” than “I did a bad thing, oops, sorry”.
Sometimes someone might take “not being able to say a thing” as a cost, and prefer the downvotes over the silence.
In general it seems like a hard problem, and it’s not clear to me that downvoting this kind of thing is a mistake.
I’d also really like to see a return of the old LW cultural thing of, if you downvote then you explain why. There are some downvotes on my comments that I’m left scratching my head about and going “Okay, whatever.” It’s hard for downvotes to improve culture if the feedback amounts to “Bad.”
I think there’s currently too many things that deserve downvotes for that to be realistic.
That’s really not a central example of what I meant. I meant more like this one. Or this one.
But also, yeah, I do kinda feel like “downvoting people when they admit they did something bad” is a thing we sometimes do here and that’s not great incentives. If someone wants to avoid that kind of downvote, “stop admitting to the bad thing” seems like an obvious strategy. Oops! And like, I remember times when I asked someone a question and they got downvoted for their answer, and I did think it was a bad answer that in a vacuum deserved downvotes, but I still upvoted as thanks for answering.
Yep. This is messy and unfortunate, I agree.
Someone might not have realized the thing they did was bad-according-to-LW, and the downvotes help signal that.
It’s not possible to take the downvotes as a signal of this if downvotes get used for a wide range of things. If the same signal gets used for
“This was written in bad form, but if you’d written it differently it would have been welcome”
and
“Your attitude doesn’t belong on this website, and you should change it or leave”
and
“I don’t like your vibe, so I’m just gonna downvote”
then the feedback isn’t precise enough to be helpful in shaping behavior.
If someone did a bad thing and doesn’t care, maybe we just don’t want them here.
True.
Although if the person disagrees with whether it was bad, and the answer to that disagreement is to try to silence them… then that seems to me like a pretty anti-epistemic norm. At least locally.
I’d also really like to see a return of the old LW cultural thing of, if you downvote then you explain why. There are some downvotes on my comments that I’m left scratching my head about and going “Okay, whatever.” It’s hard for downvotes to improve culture if the feedback amounts to “Bad.”
I think there’s currently too many things that deserve downvotes for that to be realistic.
I have a hard time believing this claim. It’s not what I see when I look around.
The dynamic would be pretty simple:
After I downvote, I skim the replies to see if someone else already explained what had me do the downvote. If so, I upvote that explanation and agree-vote it too.
If there’s no such explanation, I write one.
Easy peasy. I seriously doubt the number of things needing downvotes on this site is so utterly overwhelming that this approach is untenable. The feedback would be very rich, the culture well-defined and transparent.
I don’t know why LW stopped doing this. Once upon a time it used to cost karma to downvote, so people took downvotes more seriously. I assume there was some careful thought put into changing that system to the current one. I haven’t put more than a sum total of maybe ten minutes of thinking into this. So I’m probably missing something.
But without knowing what that something is, and without a lot of reason for me to invest a ton more time into figuring it out… my tentative but clear impression is that what I’m describing would be way better for culture here by a long shot.
It’s not possible to take the downvotes as a signal of this if downvotes get used for a wide range of things.
Perhaps not in general, but I think it’s often pretty clear. Like you’ve already said “I’m guessing the intention was to punish me for not caring”, and yes, I think you’re right. Seems to me the signal was recieved as intended.
Although if the person disagrees with whether it was bad, and the answer to that disagreement is to try to silence them… then that seems to me like a pretty anti-epistemic norm. At least locally.
Well, if someone comes here arguing for flat-earthism, I’m probably going to downvote without bothering to read their arguments. Is that anti-epistemic? Maybe, I guess? Certainly yes, if it turns out that the earth is flat (and that their arguments are correct). And “this practice isn’t anti-epistemic as long as we only dismiss false ideas” is, um. Nevertheless, I endorse that practice.
If someone comes around here calling people names, and we downvote that rather than checking in “hey are you doing this because you think name calling is good actually? Would you like to dialogue about that?” is that anti-epistemic? Again, maybe yes? But I endorse it anyway.
The dynamic would be pretty simple:
After I downvote, I skim the replies to see if someone else already explained what had me do the downvote. If so, I upvote that explanation and agree-vote it too.
If there’s no such explanation, I write one.
Easy peasy.
I do not consider writing these explanations to be easy.
I seriously doubt the number of things needing downvotes on this site is so utterly overwhelming that this approach is untenable.
I can think of a few places we might disagree here: how many things deserve downvotes, how costly it is to explain them, how realistic it is for people to pay those costs. I’m not super enthusiastic about trying to drill down into this, though.
But I also think I’m less optimistic than you about the benefits of doing it. I can think of multiple conversations I’ve had where I wanted people to change what they’re doing, I explained why I thought they were doing something bad, and they just keep on doing it. You yourself seem to understand what it is that many people dislike in many of your posts and comments, and yet you keep doing the thing. Surely there are cases where it does help, but I think they’re a minority. (It seems plausible to me that the helpful cases actually do get explained more often than others. E.g. if someone explicitly asks why they’re getting downvoted, that’s evidence they’re interested in improving, and also it makes them more likely to get an explanation.)
Another thing worth mentioning is that reacts reduce the cost of explaining downvotes. I dunno how much they’re used, since I mostly use GreaterWrong which doesn’t (yet?) support them. I believe they were only added to this post later, so they wouldn’t have been helpful at the time. But yeah, if a comment gets downvoted a bunch with not even any reacts explaining why, that seems not ideal.
After I downvote, I skim the replies to see if someone else already explained what had me do the downvote. If so, I upvote that explanation and agree-vote it too.
If there’s no such explanation, I write one.
Easy peasy. I seriously doubt the number of things needing downvotes on this site is so utterly overwhelming that this approach is untenable. The feedback would be very rich, the culture well-defined and transparent.
I don’t know why LW stopped doing this. Once upon a time it used to cost karma to downvote, so people took downvotes more seriously. I assume there was some careful thought put into changing that system to the current one. I haven’t put more than a sum total of maybe ten minutes of thinking into this. So I’m probably missing something.
But without knowing what that something is, and without a lot of reason for me to invest a ton more time into figuring it out… my tentative but clear impression is that what I’m describing would be way better for culture here by a long shot.
I agree with you that what you propose would be better for LW’s culture. However, I think I can answer the “why did LW stop doing this” question:
An increased prevalence, in those social circles which influence decisions made by the LW admin team, of people who have a strong aversion to open conflict.
You write a post or a comment. Someone writes a reply explaining why they downvoted—in other words, a critical reply. This is open conflict—confrontation.
You reply to them to dispute their criticism, to question their characterization, to argue—more open conflict. Encouraging downvote explanations is nothing more nor less than encouraging critical comments, after all! More critical comments—more open conflict.
Some people can’t stand open conflict. So, they use their influence to cause to be enacted such policies, and to be built such structures, as will prevent confrontation, explicit disagreement, direct criticism. (This is usually couched in euphemisms, of course, as calling such things by their simple names also invites confrontation.)
Well, that particular comment had a lot of other stuff going on, and yes I think it’s a kind of comment that doesn’t belong here and no I don’t particularly feel like explaining that.
But also, yeah, I do kinda feel like “downvoting people when they admit they did something bad” is a thing we sometimes do here and that’s not great incentives. If someone wants to avoid that kind of downvote, “stop admitting to the bad thing” seems like an obvious strategy. Oops! And like, I remember times when I asked someone a question and they got downvoted for their answer, and I did think it was a bad answer that in a vacuum deserved downvotes, but I still upvoted as thanks for answering.
I’m not sure it’s so bad though. Some things that mitigate it as a strategy:
“This person strategically fails to answer certain questions” is a thing it’s possible for someone to notice and point out.
Someone might not have realized the thing they did was bad-according-to-LW, and the downvotes help signal that. (Maybe better to instead upvote the admission and downvote the thing they did? But that’s not always a thing that can be downvoted, or downvotes might not be specifically targetable to make it clear “this thing you did was bad”.)
If someone did a bad thing and doesn’t care, maybe we just don’t want them here. Downvotes probably marginally push them away, as well as marginally push them towards not-admitting-things. Notably, I feel like we’re more likely to downvote “I did a bad thing and don’t care” than “I did a bad thing, oops, sorry”.
Sometimes someone might take “not being able to say a thing” as a cost, and prefer the downvotes over the silence.
In general it seems like a hard problem, and it’s not clear to me that downvoting this kind of thing is a mistake.
I think there’s currently too many things that deserve downvotes for that to be realistic.
That’s really not a central example of what I meant. I meant more like this one. Or this one.
Yep. This is messy and unfortunate, I agree.
It’s not possible to take the downvotes as a signal of this if downvotes get used for a wide range of things. If the same signal gets used for
“This was written in bad form, but if you’d written it differently it would have been welcome”
and
“Your attitude doesn’t belong on this website, and you should change it or leave”
and
“I don’t like your vibe, so I’m just gonna downvote”
then the feedback isn’t precise enough to be helpful in shaping behavior.
True.
Although if the person disagrees with whether it was bad, and the answer to that disagreement is to try to silence them… then that seems to me like a pretty anti-epistemic norm. At least locally.
I have a hard time believing this claim. It’s not what I see when I look around.
The dynamic would be pretty simple:
After I downvote, I skim the replies to see if someone else already explained what had me do the downvote. If so, I upvote that explanation and agree-vote it too.
If there’s no such explanation, I write one.
Easy peasy. I seriously doubt the number of things needing downvotes on this site is so utterly overwhelming that this approach is untenable. The feedback would be very rich, the culture well-defined and transparent.
I don’t know why LW stopped doing this. Once upon a time it used to cost karma to downvote, so people took downvotes more seriously. I assume there was some careful thought put into changing that system to the current one. I haven’t put more than a sum total of maybe ten minutes of thinking into this. So I’m probably missing something.
But without knowing what that something is, and without a lot of reason for me to invest a ton more time into figuring it out… my tentative but clear impression is that what I’m describing would be way better for culture here by a long shot.
Perhaps not in general, but I think it’s often pretty clear. Like you’ve already said “I’m guessing the intention was to punish me for not caring”, and yes, I think you’re right. Seems to me the signal was recieved as intended.
Well, if someone comes here arguing for flat-earthism, I’m probably going to downvote without bothering to read their arguments. Is that anti-epistemic? Maybe, I guess? Certainly yes, if it turns out that the earth is flat (and that their arguments are correct). And “this practice isn’t anti-epistemic as long as we only dismiss false ideas” is, um. Nevertheless, I endorse that practice.
If someone comes around here calling people names, and we downvote that rather than checking in “hey are you doing this because you think name calling is good actually? Would you like to dialogue about that?” is that anti-epistemic? Again, maybe yes? But I endorse it anyway.
I do not consider writing these explanations to be easy.
I can think of a few places we might disagree here: how many things deserve downvotes, how costly it is to explain them, how realistic it is for people to pay those costs. I’m not super enthusiastic about trying to drill down into this, though.
But I also think I’m less optimistic than you about the benefits of doing it. I can think of multiple conversations I’ve had where I wanted people to change what they’re doing, I explained why I thought they were doing something bad, and they just keep on doing it. You yourself seem to understand what it is that many people dislike in many of your posts and comments, and yet you keep doing the thing. Surely there are cases where it does help, but I think they’re a minority. (It seems plausible to me that the helpful cases actually do get explained more often than others. E.g. if someone explicitly asks why they’re getting downvoted, that’s evidence they’re interested in improving, and also it makes them more likely to get an explanation.)
Another thing worth mentioning is that reacts reduce the cost of explaining downvotes. I dunno how much they’re used, since I mostly use GreaterWrong which doesn’t (yet?) support them. I believe they were only added to this post later, so they wouldn’t have been helpful at the time. But yeah, if a comment gets downvoted a bunch with not even any reacts explaining why, that seems not ideal.
I agree with you that what you propose would be better for LW’s culture. However, I think I can answer the “why did LW stop doing this” question:
An increased prevalence, in those social circles which influence decisions made by the LW admin team, of people who have a strong aversion to open conflict.
You write a post or a comment. Someone writes a reply explaining why they downvoted—in other words, a critical reply. This is open conflict—confrontation.
You reply to them to dispute their criticism, to question their characterization, to argue—more open conflict. Encouraging downvote explanations is nothing more nor less than encouraging critical comments, after all! More critical comments—more open conflict.
Some people can’t stand open conflict. So, they use their influence to cause to be enacted such policies, and to be built such structures, as will prevent confrontation, explicit disagreement, direct criticism. (This is usually couched in euphemisms, of course, as calling such things by their simple names also invites confrontation.)
Hence, the Less Wrong of today.