Moderation notes re: recent Said/Duncan threads

I generally agree with the above and expect to be fine with most of the specific versions of any of the three bulleted solutions that I can actually imagine being implemented.

I note re:

It'd be cruxy for me if more high-contributing-users actively supported the sort of moderation regime Duncan-in-particular seems to want.

... that (in line with the thesis of my most recent post) I strongly predict that a decent chunk of the high-contributing users who LW has already lost would've been less likely to leave and would be more likely to return with marginal movement in that direction.

I don't know how best to operationalize this, but if anyone on the mod team feels like reaching out to e.g. ~ten past heavy-hitters that LW actively misses, to ask them something like "how would you have felt if we had moved 25% in this direction," I suspect that the trend would be clear. But the LW of today seems to me to be one in which the evaporative cooling has already gone through a couple of rounds, and thus I expect the LW of today to be more "what? No, we're well-adapted to the current environment; we're the ones who've been filtered for."

(If someone on the team does this, and e.g. 5 out of 8 people the LW team misses respond in the other direction, I will in fact take that seriously, and update.)

I think Said and Duncan are clearly channeling this conflict [LW(p) · GW(p)], but the confict is not about them, and doesn't originate with them. So by having them go away or stop channeling the conflict, you leave it unresolved and without its most accomplished voices, shattering the possibility of resolving it in the foreseeable future. The hush-hush strategy of dealing with troubling observations, fixing symptoms instead of researching the underlying issues, however onerous that is proving to be.

(This announcement is also rather hush-hush, it's not a post and so I've only just discovered it, 5 days later. This leaves it with less scrutiny that I think transparency of such an important step requires.)

Just want to note that I'm less happy with a lesswrong without Duncan. I very much value Duncan's pushback against what I see as a slow decline in quality, and so I would prefer him to stay and continue doing what he's doing. The fact that he's being complained about makes sense, but is mostly a function of him doing something valuable. I have had a few times where I have been slapped down by Duncan, albeit in comments on his Facebook page, where it's much clearer that his norms are operative, and I've been annoyed, but each of those times, despite being frustrated, I have found that I'm being pushed in the right direction and corrected for something I'm doing wrong.

I agree that it's bad that his comments are often overly confrontational, but there's no way to deliver constructive feedback that doesn't involve a degree of confrontation, and I don't see many others pushing to raise the sanity waterline. In a world where a dozen people were fighting the good fight, I'd be happy to ask him to take a break. But this isn't that world, and it seems much better to actively promote a norm of people saying they don't have energy or time to engage than telling Duncan (and maybe / hopefully others) not to push back when they see thinking and comments which are bad. 
 

The thing that feels actually bad is getting into a protracted discussion, on a particular (albeit fuzzy) cluster of topics

I think I want to reiterate my position that I would be sad about Said not being able to discuss Circling (which I think is one of the topics in that fuzzy cluster). I would still like to have a written explanation of Circling (for LW) that is intelligible to Said, and him being able to point out which bits are unintelligible and not feel required to pretend that they are intelligible seems like a necessary component of that.

With regards to Said's 'general pattern', I think there's a dynamic around socially recognized gnosis [LW(p) · GW(p)] where sometimes people will say "sorry, my inability/unwillingness to explain this to you is your problem" and have the commons on their side or not, and I would be surprised to see LW take the position that authors decide for that themselves. Alternatively, tech that somehow makes this more discoverable and obvious--like polls or reacts or w/e--does seem good.

I think productive conversations stem from there being some (but not too much) diversity in what gnosis people are willing to recognize, and in the ability for subspaces to have smaller conversations that require participants to recognize some gnosis.

Is there any evidence that either Duncan or Said are actually detrimental to the site in general, or is it mostly in their interactions directly with each other? As far as I can see, 99% of the drama here is in their conflicts directly with each other and heavy moderation team involvement in it.

From my point of view (as an interested reader and commenter), this latest drama appears to have started partly due to site moderation essentially forcing them into direct conflict with each other via a proposal to adopt norms based on Duncan's post while Said and others were and continue to be banned from commenting on it.

From this point of view, I don't see what either of Said or Duncan have done to justify any sort of ban, temporary or not.

I condemn the restrictions on Said Achmiz's speech in the strongest possible terms. I will likely have more to say soon, but I think the outcome will be better if I take some time to choose my words carefully.

Ray writes:

Here are some areas I think Said contributes in a way that seem important:

• Various ops/dev work maintaining sites like readthesequences.com, greaterwrong.com, and gwern.com. 

For the record, I think the value here is "Said is the person independent of MIRI (including Vaniver) and Lightcone who contributes the most counterfactual bits to the sequences and LW still being alive in the world", and I don't think that comes across in this bullet.

I could imagine an admin feature that literally just lets Said comment a few times on a post, but if he gets significantly downvoted, gives him a wordcount-based rate-limit that forces him to wrap up his current points quickly and then call it a day.

I feel like this incentivizes comments to be short, which doesn't make them less aggravating to people. For example, IIRC people have complained about him commenting "Examples?". This is not going to be hit hard by a rate limit.

I was surprised and updated on You Don't Exist, Duncan [LW · GW] getting as heavily upvoted as it did

I'm not sure what you think this is strong evidence of?

If the lifeguard isn't on duty, then it's useful to have the ability to be your own lifeguard.

I wanted to say that I appreciate the moderation style options and authors being able to delete and ban for their posts. While we're talking about what to change and what isn't working, I'd like to weigh in on the side of that being a good set of features that should be kept. Raemon, you've mentioned [LW(p) · GW(p)] those features are there to be used. I've never used the capability and I'm still glad it exists. (I can barely use it actually.) Since site wide moderators aren't going to intervene everywhere quickly (which I don't think they should or even can, moderators are heavily outnumbered) then I think letting people moderate their local piece is good.

If I ran into lots of negative feedback I didn't think was helpful and it wasn't getting moderated by me or the site admins, I'd just move my writing to a blog on a different website where I could control things. Possibly I'd set up crossposting like Zvi or Jefftk and then ignore the LessWrong comment section. If lots of people do that, then we get the diaspora effect from late LessWrong 1.0. Having people at least crossposting to LessWrong seems good to me, since I like tools like the agreement karma and the tag upvotes. Basically, the BATNA for a writer who doesn't like LessWrong's comment section is Wordpress or Substack. Some writers you'd rather go elsewhere obviously, but Said and Duncan's top level posts seem mostly a good fit here. 

I do have a question about norm setting I'm curious about. If Duncan had titled his post "Duncan's Basics of Rationalist Discourse" would that have changed whether it merited the exception around pushing site wide norms? What if lots of people started picking Norm Enforcing for the moderation guidelines and linking to it?

One thing I'd want going forward is a more public comment that, if he's going to keep posting on LessWrong, he's not going to do that (take down all his LW posts) again.

I think Duncan also posts all his articles on his own website, is this correct?

In that case, would it be okay to replace the article on LW with a link to Duncan's website? So that the articles stay there, the comments stay here, the page with comments links the article, but the article does not link the page with comments.

I am not suggesting to do this. I am asking that if Duncan (or anyone else) hypothetically at some moment decided for whatever reason that he is uncomfortable with his articles being on LW, whether doing this (moving the articles elsewhere and replacing them with the links towards the new place) would be acceptable for you? Like, whether this could be a policy "if you decide to move away from LW, this is our preferred way to do it".

Are we entertaining technical solutions at this point? If so, I have some ideas. This feels to me like a problem of balancing the two kinds of content on the site. Balancing babble to prune, artist to critic, builder to breaker. I think Duncan wants an environment that encourages more Babbling/Building. Whereas it seems to me like Said wants an environment that encourages more Pruning/Breaking. 

Both types of content are needed. Writing posts pattern matches with Babbling/Building, whereas writing comments matches closer to Pruning/Breaking. In my mind anyway. (update: prediction market)

Inspired by this post [LW · GW] I propose enforcing some kind of ratio between posts and comments. Say you get 3 comments per post before you get rate-limited?^[1] This way if you have a disagreement or are misunderstanding a post there is room to clarify, but not room for demon threads. If it takes more than a few comments to clarify that is an indication of a deeper model disagreement and you should just go ahead and write your own post explaining your views. ( as an aside I would hope this creates an incentive to write posts in general, to help with the inevitable writer turn-over)

Obviously the exact ratio doesn't have to be 3 comments to 1 post. It could be 10:1 or whatever the mod team wants to start with before adjusting as needed.

1. ^{^}

   I'm not suggesting that you get rate-limited site-wide if you start exceeding 3 comments per post. Just that you are rate-limited on that specific post. 

Said and Duncan are both among the two single-most complained about users since LW2.0 started (probably both in top 5, possibly literally top 2).

Complaints by whom? And why are these complaints significant?

Are you taking the stance that all or most of these complaints are valid, i.e. that the things being complained about are clearly bad (and not merely dispreferred by this or that individual LW member)?

(See also this recent comment [LW(p) · GW(p)], where I argue that at least one particular characterization of my commenting activity is just demonstrably inconsistent with reality.)

Just noting as a "for what it's worth"

(b/c I don't think my personal opinion on this is super important or should be particularly cruxy for very many other people)

that I accept, largely endorse, and overall feel fairly treated by the above (including the week suspension that preceded it).

This sounds drastic enough that it makes me wonder, since the claimed reason was that Said's commenting style was driving high-quality contributors away from the site, do you have a plan to follow up and see if there is any sort of measurable increase in comment quality, site mood or good contributors becoming more active moving forward?

Also, is this thing an experiment with a set duration, or a permanent measure? If it's permanent, it has a very rubber room vibe to it, where you don't outright ban someone but continually humiliate them if they keep coming by and wish they'll eventually get the hint.

A background model I want to put out here: two frames that feel relevant to me here are "harm minimization" and "taxing". I think the behavior Said does has unacceptably large costs in aggregate (and, perhaps to remind/clarify, I think a similar-in-some-ways set of behaviors I've seen Duncan do also would have unacceptably large costs in aggregate).

And the three solutions I'd consider here, at some level of abstraction, are:

1. So-and-so agrees to stop doing the behavior (harder when the behavior is subtle and multifaceted, but, doable in principle)
2. Moderators restrict the user such that they can't do the behavior to unacceptable degrees
3. Moderators tax the behavior such that doing-too-much-of-it is harder overall (but, it's still something of the user's choice if they want to do more of it and pay more tax). 

All three options seem reasonable to me apriori, it's mostly a question of "is there a good way to implement them?". The current rate-limit-proposal for Said is mostly option 2. All else being equal I'd probably prefer option 3, but the options I can think of seem harder to implement and dev-time for this sort of thing is not unlimited.

Quick update for now: @Said Achmiz [LW · GW]'s rate limit has expired, and I don't plan to revisit applying-it-again unless a problem comes up. 

I do feel like there's some important stuff left unresolved here. @Zack_M_Davis [LW · GW]'s comment on this other post [LW(p) · GW(p)] asks some questions that seem worth answering. 

I'd hoped to write up something longer this week but was fairly busy, and it seemed better to explicitly acknowledge it. For the immediate future I think improving on the auto-rate-limits and some other systemic stuff seems more important that arguing or clarifying the particular points here.

A thing that is quite important to me is that users feel comfortable ignoring Said if they don’t think he’s productive to engage with. (See below for more thoughts on this). One reason this is difficult is that it’s hard to establish common knowledge about it among authors. Another reason is that I think Said’s conversational patterns have the effect of making authors and other commenters feel obliged to engage with him (but, this is pretty hard to judge in a clear-cut way)

It seems like the natural solution here would be something that establishes this common knowledge. Something like the twitter "community notes" being attached to relevant comments that says something like "There is no obligation to respond to this comment, please feel comfortable ignoring this user if you don't feel he will productive to engage with. Discussion here [LW(p) · GW(p)]"

Do I understand you correctly as saying that the problem, specifically, is… that people reading my comments might, or do, get a mistaken impression that there exists on Less Wrong some sort of social norm which holds that authors have a social obligation to respond to comments on their posts?

That aside, I have questions about this rate limit:

• Does it apply to all posts of any kind, written by anyone? More specifically:
  • Does it apply to both personal and frontpage posts?
  • Does it apply to posts written by moderators? Posts written about me (or specifically addressing me)? Posts written by moderators about me?
  • Does it apply to this post? (I assume that it must not, since you mention that you’d like me to make a case that so-and-so, you say “I am interested in what Said actually prefers here”, etc., but just want to confirm this) EDIT: See below
  • Does it apply to “open thread” type posts (where the post itself is just a “container”, so to speak, and entirely different conversations may be happening under different top-level comments)?
  • Does it apply to my own posts? (That would be very strange, of course, but it wouldn’t be the strangest edge case that’s ever been left unhandled in a feature implementation, so seems worth checking…)
  • Does it apply retroactively to existing posts (including very old posts), or only new posts going forward?
• Is there any way for a post author to disable this rate limit, or opt out of it?
• Does the rate limit reset at a specific time each week, or is there simply a check for whether 3 posts have been written in the period starting one week before current time?
• Is there any rate limit on editing comments, or only posting new ones? (It is presumably not the intent to have the rate limit triggered by fixing a typo, for instance…)
• Is there a way for me to see the status of the rate limit prior to posting, or do I only find out whether the limit’s active when I try to post a comment and get an error?
• Is there any UI cue to inform readers or other commenters (including a post’s author) that I can’t reply to a comment of theirs, e.g., due to the rate limit?

ETA: After attempting to post this comment last night, I received a message informing me that I would not be able to do so until some hours in the then-future. This answers the crossed-out question above, I suppose. Unfortunately, it also makes the asides about wanting to know what I think on this topic… well, somewhat farcical, quite frankly.

The discussion under this post [LW · GW] is an excellent example of the way that a 3-per-week per-post comment limit makes any kind of useful discussion effectively impossible.

I continue to be disgusted with this arbitrary moderator harrassment of a long-time, well-regarded [LW(p) · GW(p)] user, apparently on the pretext that some people don't like his writing style.

Achmiz is not a spammer or a troll, and has made many highly-upvoted contributions. If someone doesn't like Achmiz's comments, they're free to downvote (just as I am free to upvote). If someone doesn't want to receive comments from Achmiz, they're free to use already-existing site functionality to block him from commenting on their own posts. If someone doesn't like his three-year-old views about an author's responsibility or lack thereof to reply to criticisms, they're free to downvote or offer counterarguments. Why isn't that the end of the matter?

Elsewhere, Raymond Arnold complains that Achmiz isn't "corrigible about actually integrating the spirit-of-our-models into his commenting style" [LW(p) · GW(p)]. Arnold also proposes that [LW(p) · GW(p)] awareness of frame control—a concept that Achmiz has criticized [LW(p) · GW(p)]—become something one is "obligated to learn, as a good LW citizen". I find this attitude shockingly anti-intellectual. Since when is it the job of a website administrator to micromanage how intellectuals think and write, and what concepts they need to accept? (As contrated to removing low-quality, spam, or off-topic comments; breaking up flame wars, &c.)

My first comment on Overcoming Bias was on 15 December 2007 [LW(p) · GW(p)]. I was at the first Overcoming Bias meetup on 21 February 2008 [LW · GW]. Back then, there was no conept of being a "good citizen" of Overcoming Bias. It was a blog. People read the blog, and left comments when they had something to say, speaking in their own voice, accountable to no authority but their own perception of reality, with no obligation to be corrigible to the spirit of someone else's models. Achmiz's first comment on Less Wrong was in May 2010 [LW(p) · GW(p)].

We were here first. This is our garden, too—or it was. Why is the mod team persecuting us? By what right—by what code—by what standard?

Perhaps it will be replied that no one is being silenced—this is just a mere rate-limit, not any kind of persecution or restriction on speech. I don't think Oliver Habryka [LW · GW] is naïve enough to believe that. Citizenship—first-class citizenship—is a Schelling point. When someone tries to take that away from you, it would be foolish to believe that they don't intend you any further harm.

I think Oli Habryka has the integrity [LW · GW] to give me a staight, no-bullshit answer here.

Maybe a middle ground would be to give authors a double-strong downvote power for comments on their posts. A comment with low enough karma is already hidden by default, and repeated strong downvotes without further response would tend chill rather than inflame the ensuing discussion, or at least push the bulk of it away from the author's arena, without silencing critics completely.

To me it seems unlikely that there'd be enough banning to prevent criticism from surfacing. Skimming through https://www.lesswrong.com/moderation, [? · GW] the amount of bans seems to be pretty small. And if there is an important critique to be made I'd expect it to be something that more than the few banned users would think of and decide to post a comment on.

One solution is to limit the number of banned users to a small fraction of overall commentors. I've written 297 posts so far and have banned only 3 users from commenting on them. (I did not ban Duncan or Said.)

My highest-quality criticism comes from users who I have never even considered banning. Their comments are consistently well-reasoned and factually correct.

What exactly does "nobody wanted to criticize it" signal that you don't get from high/low karma votes?

Vaniver privately suggested to me that I may want to offer some commentary on what I could’ve done in this situation in order for it to have gone better, which I thought was a good and reasonable suggestion. I’ll do that in this comment, using Vaniver’s summary of the situation as a springboard of sorts.

In this comment [LW(p) · GW(p)], Said describes as bad “various proposed norms of interaction such as “don’t ask people for examples of their claims” and so on”, without specifically identifying Duncan as proposing that norm (tho I think it’s heavily implied).

Then gjm objects [LW(p) · GW(p)] to that characterization as a straw man.

So, first of all, yes, I was clearly referring to Duncan. (I didn’t expect that to be obscure to anyone who’d bother to read that subthread in the first place, and indeed—so far as I can tell—it was not. If anyone had been confused, they would presumably have asked “what do you mean?”, and then I’d have linked what I mean—which is pretty close to what happened anyway. This part, in any case, is not the problem.)

The obvious problem here is that “don’t ask people for examples of their claims”—taken literally—is, indeed, a strawman.

The question is, whose problem (to solve) is it?

There are a few possible responses to this (which are not mutually exclusive).

On the one hand, if I want people to know what I mean, and instead of saying what I mean, I say something which is only approximately what I mean, and people assume that I meant what I said, and respond to it—well, whose fault is that, but mine?

Certainly one could make protestations along the lines of “haven’t you people ever heard of [ hyperbole / colloquialisms / writing off the cuff and expecting that readers will infer from surrounding context / whatever ]”, but such things are always suspect. (And even if one insists that there’s nothing un-virtuous about any particular instance of any one of those rhetorical or conversational patterns, nevertheless it would be a bit rich to get huffy about people taking words literally on Less Wrong, of all places.)

So, in one sense, the whole problem would’ve been avoided if I’d taken pains to write as precisely as I usually try to do. Since I didn’t do that, and could have, the fault would seem to be mine; case closed.

But that account doesn’t quite work.

For one thing, if someone says something you think is wrong, and you say “seems wrong to me actually”, and they reply “actually I meant this other thing”—well, that seems to me to be a normal and reasonable sort of exchange; this is how understanding is reached. I made a claim; gjm responded that it seemed like a strawman; I responded with a clarification.

Note that here I definitely made a mistake; what I should’ve included in that comment, but left out, was a clear and unambiguous statement along the lines of:

“Yes, taken literally, ‘don’t ask people for examples of their claims’ would of course be a strawman. I thought that the intended reading would be clear, but I definitely see the potential for literal (mis-)reading, sorry. To clarify:”

The rest of that comment would then have proceeded as written. I don’t think that it much needs amendment. In particular, the second paragraph (which, as Vaniver notes, does much of the work) gives a concise and clear statement of the claim which I was originally (and, at first, sloppily) alluding to. I stand by that clarified claim, and have seen nothing that would dissuade me from it.

Importantly, however, we can see that Duncan objects, quite strenuously, even to this clarified and narrowed form of what I said!

(As I note in this comment [LW(p) · GW(p)], it was not until after essentially the whole discussion had already taken place that Duncan edited his reply to my latter comment to explicitly disclaim the view that I ascribed to him. For the duration of that whole long comment exchange, it very much seemed to me that Duncan was not objecting because I was ascribing to him a belief he does not hold, but rather because he had not said outright that he held such a belief… but, of course, I never claimed that he had!)

So even if that clarified comment had come first (having not, therefore, needed any acknowledgment of previous sloppiness), there seems to be little reason to believe that Duncan would not have taken umbrage at it.

Despite that, failing to include that explicit acknowledgement was an error. Regardless of whether it can be said to be responsible for the ensuing heated back-and-forth (I lean toward “probably not”), this omission was very much a failure of “local validity” on my part, and for that there is no one to blame but me.

Of the rest of the discussion thread, there is little that needs to be said. (As Vaniver notes, some of my subsequent comments both clarify my claims further and also provide evidence for them.)

In the response I would have wanted to see, Duncan would have clearly and correctly pointed to that difference. He is in favor of people asking for examples [combined with other efforts to cross the gap], does it himself, gives examples himself, and so on. The unsaid [without anything else] part is load-bearing and thus inappropriate to leave out or merely hint at. [Or, alternatively, using "ask people for examples" to refer to comments that do only that, as opposed to the conversational move which can be included or not in a comment with other moves.]

I agree that the hypothetical comment you describe as better is in fact better. I think something like ... twenty-or-so exchanges with Said ago, I would have written that comment?  I don't quite know how to weigh up [the comment I actually wrote is worse on these axes of prosocial cooperation and revealing cruxes and productively clarifying disagreement and so forth] with [having a justified true belief that putting forth that effort with Said in particular is just rewarded with more branches being created].

(e.g. there was that one time [LW(p) · GW(p)] recently where Said said I'd blocked people due to disagreeing with me/criticizing me, and I said no, I haven't blocked anybody for disagreeing/criticizing, and he responded "I didn’t say anything about 'blocked for disagreeing [or criticizing]'. (Go ahead, check!)" and the actual thing he'd said was that they'd been blocked due to disagreeing/criticizing; that's the level of ... gumming up the works? gish-gallop? ... that I've viscerally come to expect.)

Like, I think there's plausibly a CEV-ish code of conduct in which I "should", at that point, still have put forth the effort, but I think it's also plausible that the correct code of conduct is one in which doing so is a genuine mistake and ... noticing that there's a hypothetical "better" comment is not the same as there being an implication that I should've written it?

Something something, how many turns of the cheek are actually correct, especially given that, the week prior, multiple commenters had been unable, with evidence+argument+personal testimony, to shift Said away from a strikingly uncharitable prior.

(This does not count as two hypotheses [LW(p) · GW(p)] in my culture.)

Mine either, to be clear; I felt by that point that Said had willingly put himself outside of the set of [signatories to the peace treaty], turning down many successive opportunities to remain in compliance with it.  I was treating his statements closer to the way I think it is correct to treat the statements of the literal Donald Trump than the way I think it is correct to treat the statements of an undistinguished random Republican.

(I can go into the reasoning for that in more detail, but it seems sort of conflicty to do so unprompted.)

If Said ate breakfasts of only cereal, and Duncan said that was unhealthy and he shouldn't do it, it is not quite right to say Duncan 'thinks you shouldn't eat cereal', as he might be in favor of cereal as part of a balanced breakfast; but also it is not quite right for Duncan to ignore Said's point that one of the main issues under contention is whether Said can eat cereal by itself (i.e. asking for examples without putting in interpretative labor).

I'm a little lost in this analogy; this is sort of where the privileging-the-hypothesis complaint comes in.

The conversation had, in other places, centered on the question of whether Said can eat cereal by itself; Logan for instance highlighted Said's claim in a reply on FB:

Furthermore, you have mentioned the “inferential gap” several times, and suggested that it is the criticizer’s job, at least in part, to bridge it. I disagree.

There, the larger question of "can you eat only cereal, or must you eat other things in balance?" is front-and-center.

But at that point in the subthread, it was not front-and-center; yes, it was relevant context, but the specific claim being made by Said was clear, and discrete, and not at all dependent-on or changed-by that context.

The history of that chain:

Said includes, in a long comment "In summary, I think that what’s been described as 'aiming for convergence on truth' is some mixture of" ... "contentless" ... "good but basically unrelated to the rest of it" ... "bad (various proposed norms of interaction such as 'don't ask people for examples of their claims' and so on)"

gjm, in another long comment, includes "I don't know where you get 'don't ask people for examples of their claims' from and it sounds like a straw man" and goes on to elaborate "I think the things Duncan has actually said are more like 'Said engages in unproductive modes of discussion where he is constantly demanding more and more rigour and detail from his interlocutors while not providing it himself', and wherever that lands on a scale from '100% truth' to '100% bullshit' it is not helpful to pretend that he said 'it is bad to ask people for examples of their claims'.

There's a bunch of other stuff going on in their back and forth, but that particular thread has been isolated and directly addressed, in other words.  gjm specifically noted the separation between the major issue of whether balance is required, and this other, narrower claim.

Said replied:

If “asking people for examples of their claims” doesn’t fit Duncan’s stated criteria for what constitutes acceptable engagement/criticism, then it is not pretending, but in fact accurate, to describe Duncan as advocating for a norm of “don’t ask people for examples of their claims”.

Which, yes, I straightforwardly agree with the if-then statement; if "asking people for examples of their claims" didn't fit my stated criteria for what constitutes acceptable engagement or criticism, then it would be correct to describe me as advocating for a norm of "don't ask people for examples of their claims."

But like.  The if does not hold.  It really clearly doesn't hold.  It was enough of an out-of-nowhere strawman/non-sequitur that gjm specifically called it out as "???" at which point Said doubled down, saying the above and also

Duncan has, I think, made it very clear that that a comment that just says “what are some examples of this claim?” is, in his view, unacceptable. That’s what I was talking about. I really do not think it’s controversial at all to ascribe this opinion to Duncan.

It seems like, in your interpretation, I "should" (in some sense) be extending a hand of charity and understanding and, I dunno, helping Said to coax out his broader, potentially more valid point—helping him to get past his own strawman and on to something more steel, or at least flesh.  Like, if I am reading you correctly above, you're saying that, by focusing in on the narrow point that had been challenged by gjm and specifically reaffirmed by Said, I myself was making some sort of faux pas.

(Am I in fact reading you correctly?)

I do not think so.  I think that, twenty exchanges prior, I perhaps owed Said something like that degree of care and charity and helping him avoid tying his own shoelaces together.  I certainly feel I would owe it to, I dunno, Eric Rogstad or Julia Galef, and would not be the slightest bit loath to provide it.

But here, Said had just spent several thousand words the week prior, refusing to be budged from a weirdly uncharitable belief about the internals of my mind, despite that belief being incoherent with observable evidence and challenged by multiple non-me people.  I don't think it's wise-in-the-sense-of-wisdom to a) engage with substantial charity in that situation, or b) expect someone else to engage with substantial charity in that situation.

(You can tell that my stated criteria do not rule out asking people for examples of their claims in part because I've written really quite a lot about what I think constitutes acceptable engagement or criticism, and I've just never come anywhere close to a criterion like that, nor have I ever complained about someone asking for examples unless it was after a long, long string of what felt like them repeatedly not sharing in the labor of truthseeking. Like, the closest I can think of is this thread with tailcalled [LW(p) · GW(p)], in which (I think/I hope) it's pretty clear that what's going on is that I was trying to cap the total attention paid to the essay and its discussion, and thus was loath to enter into something like an exchange of examples—not that it was bad in any fundamental sense for someone to want some.  I did in fact provide some, a few comments deeper in the thread, though I headlined that I hadn't spent much time on them.)

So in other words: I don't think it was wrong to focus on the literal, actual claim that Said had made (since he made it, basically, twice in a row, affirming "no, I really mean this" after gjm's objection and even saying that he thinks it is so obvious as to not be controversial.  I don't think I "ought" to have had a broader focus, under the circumstances—Said was making a specific, concrete, and false claim, and his examples utterly fail to back up that specific, concrete, and false claim (though I do agree with you that they back up something like his conception of our broader disagreement).

I dunno, I'm feeling kind of autistic, here, but I feel like if, on Less Wrong dot com, somebody makes a specific, concrete claim about my beliefs or policies, clarifies that yes, they really meant that claim, and furthermore says that such-and-such links are "citations for [me] expressing the sentiment [they've] ascribed to [me]" when they simply are not—

It feels like emphatic and unapologetic rejection should be 100% okay, and not looked at askance.  The fact that they are citations supporting a different claim is (or at least, I claim, should be) immaterial; it's not my job to steelman somebody who spent hours and hours negatively psychologizing me in public (while claiming to have no particular animus, which, boy, a carbon copy of Said sure would have had Words about).

I think there's a thing here of standards unevenly applied; surely whatever standard would've had me address Said's "real" concern would've also had Said behave much differently at many steps prior, possibly never strawmanning me so hard in the first place?

I think this situation is, on some level, pretty symmetric.

1. I think the features of Said's commenting style that people (not just Duncan!) find annoying are things that Said is deliberately optimizing for or the results of principled commitments he's made, so it's not just a simple bug that can be fixed.
2. I think the features of Duncan's conflict resolution methods that people find offputting are similarly things that Duncan is deliberately optimizing for or the results of principled commitments he's made, so it's not just a simple bug that can be fixed.

I think the asymmetry breaks in that, like, a bunch of people have asked Said to stop and he won't; I'm quite eager to stop doing the conflict resolution that people don't like, if there can pretty please be some kind of system in place that obviates it. I much prefer the world where there are competent police to the world where I have to fight off muggers in the alley—that's why I'm trying so hard to get there to be some kind of actually legible standards rather than there always being some plausible reason why maybe we shouldn't just say "no" to the bullshit that Zack or Said or anonymouswhoever is pulling.

Right now, though, it feels like we've gone from "Ben Hoffman will claim Duncan wants to ghettoize people and it'll be left upvoted for nine days with no mod action" to "Ray will expound on why he thinks it's kinda off for Said to be doing what he's doing but there won't be anything to stop Said from doing it" and I take Oli's point about this stuff being hard and there being other priorities but like, it's been years.  And I get a stance of, like, "well, Duncan, you're asking for a lot," but I'm trying pretty hard to earn it, and to ... pave the way?  Help make the ask smaller? ... with things like the old Moderating LessWrong post and the Concentration of Force post and the more recent Basics post.  Like, I can't think of much more that someone with zero authority and zero mantle can do.  My problem is that abuse and strawmanning of me gets hosted on LW and upvoted on LW and people are like, well, maybe if you patiently engaged with and overturned the abuse and strawmanning in detail instead of fighting back—

I dunno.  If mods would show up and be like "false" and "cut it out" I would pretty happily never get into a scrap on LW ever again.  

Locally, my sense is Duncan started it,

:(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((

This, more than anything else, is like "just give up and leave, this is definitely not a garden."

I didn't make it to every point, but hopefully you find this more of the substantive engagement you were hoping for.

I interpret a lot of Duncan’s complaints here thru the lens of imaginary injury that he writes about here.

I just want to highlight this link (to one of Duncan’s essays on his Medium blog), which I think most people are likely to miss otherwise.

That is an excellent post! If it was posted on Less Wrong (I understand why it wasn’t, of course EDIT: I was mistaken about understanding this; see replies), I’d strong-upvote it without reservation. (I disagree with some parts of it, of course, such as one of the examples—but then, that is (a) an excellent reason to provide specific examples, and part of what makes this an excellent post, and (b) the reason why top-level posts quite rightly don’t have agree/disagree voting. On the whole, the post’s thesis is simply correct, and I appreciate and respect Duncan for having written it.)

(Expanding on this comment [LW(p) · GW(p)])

The key thing missing from your account of my views is that while I certainly think that “local validity checking” is important, I also—and, perhaps, more importantly—think that the interactions in question are not only fine, but good, in a “relational” sense.

So, for example, it’s not just that a comment that just says “What are some examples of this?” doesn’t, by itself, break any rules or norms, and is “locally valid”. It’s that it’s a positive contribution to the discussion, which is aimed at (a) helping a post author to get the greatest use out of his post and the process and experience of posting it, and (b) helping the commentariat get the greatest use out of the author’s post. (Of course, (b) is more important than (a)—but they are both important!)

Some points that follow from this, or depend on this:

First, such contributions should be socially rewarded to the degree that they are necessary. By “necessary”, here, I mean that if it is the case that some particular sort of criticism or some particular sort of question is good (i.e., it contributes substantially to how much use can be gotten out of a post), but usually nobody asks that sort of question or makes that sort of criticism, then anyone who does do that, should be seen as making not only a good but a very important contribution. (And it’s a bad sign when this sort of thing is common—it means that at least some sorts of important criticisms, or some sorts of important questions, are not asked nearly often enough!)

Meanwhile, asking a sort of question or making a sort of criticism which is equally good but is usually or often made, such that it is fairly predictable and authors can, with decent probability, expect to get it, then such a question or criticism is still good and praiseworthy, but not individually as important (though of course still virtuous!).

In the limit, an author will know that if they don’t address something in their post, somebody will ask about it, or comment on it. (And note that it’s not always necessary, in such a case, to anticipate a criticism or question in your post, even if you expect it will be made! You can leave it to the comments, being ready to respond to it if it’s brought up—or proactively bringing it up yourself [LW(p) · GW(p)], filling the role of your own devil’s advocate.)

In other words—

I mean, if 10 years from now, when you are doing something quick and dirty, you suddenly visualize that I am looking over your shoulders and say to yourself “Dijkstra would not have liked this”, well, that would be enough immortality for me.

And this is a good thing. If you posit some abstraction in your post, you should think “they’re gonna ask me for examples in the comments”. (It’s a bad sign, again, if what you actually think is “Said Achmiz is gonna ask me for examples in the comments”!) And this should make you think about whether you have examples; and what those examples demonstrate; or, if you don’t have any, what that means; etc.

And the same goes for many other sorts of questions one could ask, or criticisms one could make.

(Relatedly: I, too, want to “build up a context in which people can hold each other accountable”. But what exactly do you think that looks like?)

Second, it is no demerit to a post author, if one commenter asks a question, and another commenter answers it, without the OP’s involvement (or perhaps with merely a quick note saying “endorsed!”). Indeed it’s no demerit to an author, even, if questions are asked, or criticisms made, in the comments, to which the OP has no good answer, but which are answered satisfactorily by others, such that the end result is that knowledge and understanding are constructed by a collective effort that results in even the author of the post, himself, learning something new!

This, by the way, is related to the reasons why I find the “authors can ban people from their posts” thing so frustrating and so thoroughly counterproductive. If I write a comment under someone’s post, about someone’s post, certainly there’s an obvious sense in which it’s addressed to the author of the post—but it’s not just addressed to them! If I wanted to talk to someone one-on-one, I could send a private message… but unless I make a point of noting that I’m soliciting the OP’s response in particular (and even then, what’s to stop anyone else from answering anyway?), or ask for something that only the OP would know… comments / questions are best seen as “put to the whole table”, so to speak. Yes, if the post author has an answer they think is appropriate to provide, they can, and should, do that. But so can and should anyone else!

It’s no surprise that, as others have noted [LW(p) · GW(p)], the comments section of a post is, not infrequently, at least as useful as the post itself. And that is fine! It’s no indictment of a post’s author, when that turns out to be the case!

The upshot of this point and the previous one is that in (what I take to be) a healthy discussion environment, when someone writes a comment under your post that just says, for instance, “What are some examples of this?”, there is no good reason why that should contribute to any “relational” difficulties. It is the sort of thing that helps to make posts useful, not just to the commentariat as a whole but also to those posts’ authors; and the site is better if people regularly make such comments, ask such questions, pose such criticisms.

And, thus: third, if someone finds that they react to such engagement as if it were some sort of attack, annoyance, problem, etc., that is a bug, and one which they should want to fix. Reacting to a good thing as if it were a bad thing is, quite simply, a mistake.

Note, again, that the question isn’t whether some particular comment is “locally valid” in an “atomic” sense while being problematic in a “relational” sense. The question, rather, is whether the comment is simply good (in a “relational” sense or in any other sense), but is being mistakenly reacted to as though it were bad.

Thank you for laying out your reasoning.

I don’t have any strong objections to any of this (various minor ones, but that’s to be expected)…

… except the last paragraph (#5, starting with “I think Said is trying to figure out …”). There I think you importantly mis-characterize my views; or, to be more precise, you leave out a major aspect, which (in addition to being a missing key point), by its absence colors the rest of your characterization. (What is there is not wrong, per se, but, again, the missing aspect makes it importantly misleading.)

I would, of course, normally elaborate here, but I hesitate to end up with this comment thread/section being filled with my comments. Let me know if you want me to give my thoughts on this in detail here, or elsewhere.

(EDIT: Now expanded upon in this comment [LW(p) · GW(p)].)

By the way, I will note that I am both quite surprised and, separately, something like dismayed, at how devastatingly effective has been what I will characterize as "Said's privileging-the-hypothesis [LW · GW] gambit."

Like, Said proposed, essentially, "Duncan holds a position which basically no sane person would advocate, and he has somehow held this position for years without anyone noticing, and he conspicuously left this position out of his very-in-depth statement of his beliefs about discourse norms just a couple of months ago"

and if I had realized that I actually needed to seriously counter this claim, I might have started with "bro do you even Bayes?"

(Surely a reasonable prior on someone holding such a position is very very very low even before taking into account the latter parts of the conjunction.)

Like, that Vaniver would go so far as to take the hypothesis

Duncan has, I think, made it very clear that that a comment that just says 'what are some examples of this claim?' is, in his view, unacceptable"

and then go sifting through the past few comments with an eye toward using them to distinguish between "true" and "false" is startling to me.

"Foolishness," Severus said softly. "Utter foolishness. The Dark Mark has not faded, nor has its master."

"See, that's what I mean by formally insufficient Bayesian evidence. Sure, it sounds all grim and foreboding and stuff, but is it that unlikely for a magical mark to stay around after the maker dies? Suppose the mark is certain to continue while the Dark Lord's sentience lives on, but a priori we'd only have guessed a twenty percent chance of the Dark Mark continuing to exist after the Dark Lord dies. Then the observation, 'The Dark Mark has not faded' is five times as likely to occur in worlds where the Dark Lord is alive as in worlds where the Dark Lord is dead. Is that really commensurate with the prior improbability of immortality? Let's say the prior odds were a hundred-to-one against the Dark Lord surviving. If a hypothesis is a hundred times as likely to be false versus true, and then you see evidence five times more likely if the hypothesis is true versus false, you should update to believing the hypothesis is twenty times as likely to be false as true. Odds of a hundred to one, times a likelihood ratio of one to five, equals odds of twenty to one that the Dark Lord is dead -"

"Where are you getting all these numbers, Potter?"

"That is the admitted weakness of the method," Harry said readily. "But what I'm qualitatively getting at is why the observation, 'The Dark Mark has not faded', is not adequate support for the hypothesis, 'The Dark Lord is immortal.' The evidence isn't as extraordinary as the claim."

The observation "Duncan groused at Said for doing too little interpretive and intellectual labor relative to that which he solicited from others" is not adequate support for "Duncan generally thinks that asking for examples is unacceptable." This is what I meant by the strength of the phrase "blatant falsehood." I suppose if you are starting from "either Mortimer Snodgrass did it, or not," rather than from "I wonder who did the murder," then you can squint at my previous comments—

(including the one that was satirical [LW(p) · GW(p)], which satire, I infer from Vaniver pinging me about my beliefs on that particular phrase offline, was missed)

—and see in them that the murderer has dark hair, and conclude from Mortimer's dark hair that there should be a large update toward his guilt.

But I rather thought we didn't do that around here, and did not expect anyone besides Said to seriously entertain the hypothesis, which is ludicrous.

(I get that Said probably genuinely believed it, but the devout genuinely believe in their gods and we don't give them points for that around here.)

If the mods clearly disagree with Duncan, then what does 'losing the bet' look like? What was staked here?

I plan to just leave/not post essays here anymore if this isn't fixed. LW is a miserable place to be, right now. ¯\_(ツ)_/¯

(I also said the following in a chat with several of the moderators on 4/8: > I spent some time wondering if I would endorse a LW where both Duncan and Said were banned, and my conclusion was "yes, b/c that place sounds like it knows what it's for and is pruning and weeding accordingly.")

I note that this is leaving out recent and relevant background mentioned in this comment [LW(p) · GW(p)].

Meanwhile I find Duncan vaguely fascinating like he is a very weird bug

I don't know^[1] for sure what purpose this analogy is serving in this comment, and without it the comment would have felt much less like it was trying to hijack me into associating Duncan with something viscerally unpleasant.

1. ^{^}

   My guess is that it's meant to convey something like your internal emotional experience, with regards to Duncan, to readers.