Self-consciousness wants to make everything about itself

I agree with most of this post, but do want to provide a stronger defense of at least a subtype of tone arguments that I think are important (I think those are not the most frequently deployed type of tone arguments, but they are of a type that I think are important to allow people to make).

Tone arguments are practically always in bad faith. They aren't made by people trying to help an idea be transmitted to and internalized by more others.

To give a very extreme example of the type I am thinking about: Recently I was at an event where I was talking with two of my friends about some big picture stuff in a pretty high context conversation. We were interrupted by another person at the event who walked into the middle of our conversation, said "Fuck You" to me, then stared at me aggressively for about 10 seconds, then insulted me and then said "look, the fact that you can't even defend yourself with a straight face makes it obvious that what I am saying is right", then turned to my friends and explained loudly how I was a horrible person, and left a final "fuck you" before walking away.

The whole interaction left me quite terrified. I managed to keep my composure, but I felt pretty close to tears afterwards, had my adrenaline pumping and it took me about half an hour to calm down and be able to properly think clearly again. I think my friends had an even worse experience.

I actually think some of the way the person insulted me were relatively on point and I think reflected some valid criticisms of at least some of my past actions. I nevertheless was completely unable to respond and would have much preferred this interaction to not have taken place. I think it is really important that I can somehow avoid being in that kind of situation.

I usually really don't care about tone, but I do think there are certain choices of tone that imply threats or that almost directly hook into parts of my brain that make it very hard to think (the statement of "the fact that you can't even defend yourself with a straight face makes it obvious that I am right" is one of those for me). I think a lot of this is mitigated by text-based conversations, but far from all of it. In particular, if someone threatens some form of serious violence, or insinuates violence, which is often dependent on tone, then I do think it's reasonable to use tone as real evidence about the other person's intention, or (more rarely) intervene on tone if someone tries intentionally use tone to invoke a sense of threat.

Obviously the ability for people to threaten each other is much greater in person, and it's much easier to directly inflict costs on the other person by interrupting or screaming or hitting someone, but I think many forms of violent threat are still available in text-based communication, and often communicated via tone. I think long-lasting, mostly anonymous online communities tend to learn that most forms of threatening-sounding language don't tend to correspond to actual threats (any long-term user of 4chan will be quite robust to threatening sounding language), but I think for most people using insinuated threats is a quite reliable way to manipulate and control them, making it an actually powerful weapon that I have definitely seen used intentionally.

I think there is a subtype of tone arguments that is something like "you are threatening to hurt me without saying so directly, because saying so directly would be an obvious norm violation", which I think is actually important. I don't know whether it's ever really a good choice to make that argument explicitly (since that argument can easily be abused), or whether it's usually just better to disengage, but I don't want to rule out that it's never a good idea to make it explicitly.

I think the most common form of violence that I see being threatened in online forums is "I will get the audience to hate and attack you by tearing you apart in public". I think in an environment where the audience is full of careful thinkers this can be a fine generalizable strategy, if the audience has enough discernment to make this a highly truth-asymmetric weapon. But in most online environments there are quite predictable and abusable way to cause the audience to get angry at someone, which mostly makes this a weapon that can be wielded without much relation to accuracy of the participant's opinions (not fully uncorrelated, but not correlated enough to avoid abuse).

Wielding the audience this way is often done via the choice of a specific type of tone and rhetorical style, and so arguing against that kind of tone in public environments also strikes me as the correct choice under certain circumstances.

I think on LessWrong, we are much closer to the "the audience is full of careful thinkers so your ability to cause your enemy to be torn apart is actually correlated to the truth" side of the spectrum, which I think is why I tend to have a much higher tolerance for certain tones on LessWrong than in other places. But I think it's still not perfect. LessWrong is still a public forum, and I think there are still rhetorical styles that I expect to at least sometimes (but still with predictable direction) cause parts of the audience to attack the person against whom the rhetoric is wielded, without good justification.

People on LessWrong do however tend to have the patience to listen to good meta-level arguments about threats and against the use of violence in general, but that requires the ability for tone to be a valid subject of discussion at all, which is why I want to make sure that this post doesn't move tone completely out of the overton window.

This line of thought caused me to think that it might be quite valuable to have some kind of "conversation escrow" that allows people to have conversation in private that still reliably get published. As an example, you could imagine a feature on LessWrong to invite someone to a private comment-thread. The whole thread happens in private, and is scheduled to be published a week after the last comment in the thread was made, unless any of the participants presses a veto button if they feel that the eventual-public nature of the conversation was wielded against them during the conversation. I don't know how many conversations would end up being published, but it feels like potentially worth a try that avoids many of the error modes I am worried about with more direct public conversation.

The problem with Calvinism is that it does not allow for improvement. We are (Calvin and Calvinists say) utterly depraved, and powerless to do anything to raise ourselves up from the abyss of sin by so much as the thickness of a hair. We can never be less wrong. Only by the external bestowal of divine grace can we be saved, grace which we are utterly undeserving of and are powerless to earn by any effort of our own. And this divine grace is not bestowed on all, only upon some, the elect, predetermined from the very beginning of Creation.

Calvinism resembles abusive parenting more than any sort of ethical principle.

Eliezer has written of a similar concept in Judaism [LW · GW]:

Each year on Yom Kippur, an Orthodox Jew recites a litany which begins Ashamnu, bagadnu, gazalnu, dibarnu dofi, and goes on through the entire Hebrew alphabet: We have acted shamefully, we have betrayed, we have stolen, we have slandered . . .

As you pronounce each word, you strike yourself over the heart in penitence. There’s no exemption whereby, if you manage to go without stealing all year long, you can skip the word gazalnu and strike yourself one less time. That would violate the community spirit of Yom Kippur, which is about confessing sins—not avoiding sins so that you have less to confess.

By the same token, the Ashamnu does not end, “But that was this year, and next year I will do better.”

The Ashamnu bears a remarkable resemblance to the notion that the way of rationality is to beat your fist against your heart and say, “We are all biased, we are all irrational, we are not fully informed, we are overconfident, we are poorly calibrated . . .”

Fine. Now tell me how you plan to become less biased, less irrational, more informed, lessoverconfident, better calibrated.

When all are damned from the very beginning, when "everything is problematic", then who in fact gets condemned, what gets problematised, and who are the elect, are determined by political struggle for the seat of judgement. At least in Calvinism, that seat was reserved to God, who does not exist (or as Calvinists would say, whose divine will is unknowable), leaving people to deal with each others' flaws on a level standing.

What if it isn’t actually possible to be “not racist” or otherwise “a good person”, at least on short timescales? What if almost every person’s behavior is morally depraved a lot of the time (according to their standards of what behavior makes someone a “good person”)? What if there are bad things that are your fault? What would be the right thing to do, then?

This paragraph (and the entire concept described in this last part of the post) is—if the words therein are taken to have their usual meanings—inherently self-contradictory.

If it’s not possible for you to do something, then it cannot possibly be your fault that you don’t do that thing. (“Ought implies can” is the classic formulation of this idea.)

Rejecting “ought implies can” turns the concepts of fault, good, etc., into parodies of themselves, and morality into nonsense. If morality isn’t about how to make the right choices, then what could it be about? Whatever that thing is, it’s not morality. If you say that according to your “morality”, I can’t be “good”, on account of there being no “right” choice available to me, my answer is: clearly, whatever you mean by “good” and “right” and “morality” isn’t what I mean by those words, and I am probably uninterested in discussing these things that you mean by those words. In other words: by my lights, you have not said anything about good or right or morality, only about some different, unrelated things (for which you have, regrettably, chosen to use these already-existing words).

“Integers,” it was once said [LW(p) · GW(p)], “are slippery … If you say something unconventional about integers, you cease to talk about them.”

And so with morality.

Considering this more widely, here's a diagram I came up with. (Thanks to Raemon for advice on embedding images.)

(Please let me know if you do not see an image above. There might be a setting on my web site that blocks embedding.) (ETA: minor changes to image.)

"Feeling guilty/ashamed, then lashing out in order to feel better" is a good example of an exile/protector dynamic [LW · GW]. It may be possible to reduce the strength of such reactions through an intellectual argument, as you seem to suggest yourself to have done. But as Ruby suggested [LW(p) · GW(p)], the underlying insecurity causing the protector response is likely to be related to some fear around social rejection, in which case working on the level of intellectual concepts may be insufficient for changing the responses.

I suspect that the "nobody is good" move, when it works, acts on the level of self-concept editing [? · GW], changing your concepts so as to redefine the "evil" act to no longer be evidence about you being evil. But based on my own experience, this move only works to heal the underlying insecurity when your brain considers the update to be true - which is to say, when it perceives the "evil act" to be something which your social environment will no longer judge you harshly for, allowing the behavior to update. If this seems false, or the underlying insecurity is based on a too strong of a trauma, then one may need to do more emotional kind of work in order to address it.

That said, one may adjust the problematic response pattern even without healing the original insecurity: e.g. sufficiently strong external pressure saying that the response is bad, may set up new mechanisms which suppress the-responses-which-have-been-judged-as-bad. But this may involve building a higher tower of protectors [? · GW], meaning that healing the original insecurity would allow for more a flexible range of behavior.

Incidentally, I don't think that tone arguments are always made in bad faith. E.g. within EA, there have been debates over "should we make EA arguments which are guilt/obligation-based", which is basically an argument over the kind of a tone to use. Now, I agree that tone arguments often are themselves a protector reaction, aimed at just eliminating the unpleasant tone. But there also exists a genuine disagreement over what kind of a tone is more effective, and there exist lots of people who are genuinely worried that groups using too harsh of a tone to create social change are shooting themselves in the foot.

My guess is that people arguing for a "nice" tone are doing so because of an intuitive understanding that a harsh tone is more likely to trigger protector responses and thus be counterproductive. People may even be trying to communicate that it would be easier for them personally to listen if their protector responses weren't being triggered. On the other hand, people arguing for a harsh tone are doing so because of an intuitive understanding that the "nice" tone is often too easily ignored, and that making people feel guilty over the defensive reaction in the first place is also a strategy which can be made to work. This seems to be an area where different people will react differently to the same strategies, causing people to be inclined to over-generalize from their own experience [LW · GW].

It might be worth separating self-consciousness (awareness of how your self looks from within) from face-consciousness (awareness of how your self looks from outside). Self-consciousness is clearly useful as a cheap proxy for face-consciousness, and so we develop a strong drive to be able to see ourselves as good in order for others to do so as well. We see the difference between this separation and being a good person being only a social concept (suggested by Ruby [LW(p) · GW(p)]) by considering something like the events in "Self-consciousness in social justice" with only two participants: then there is no need to defend face against others, but people will still strive for a self-serving narrative.

Correct me if I'm wrong: you seem more worried about self-consciousness and the way it pushes people to not only act performatively, but also limits their ability to see their performance as a performance causing real damage to their epistemics.

I've been paying more attention to this reaction over the last few days since reading this post, and realize that its' often a bit different than "You're calling this bad, I've done it, I don't want to be a bad person, so I want to defend this as good."

Oftentimes the actual reaction is something like "You're calling this bad, and I've done it and don't believe its' bad. If I let this slide, I'm letting you set a precedent for calling my acceptable behavior bad, and I can't do that."

An example: Having lunch with my sister and her friend. They start making fun of Tom Cruise for dating a younger (but still adult) actress and getting her gifts on set. I don't believe there's anything wrong with age gaps and expressing the love language of giving gifts between consenting adults. If I don't say anything, there's a sense that I'm implicitly giving them license to call ME bad if I date someone with an age gap or give them gifts.

So I defend the practice, not because its' suddenly about a part of myself I don't want to look at, but because its' attacking things I would or have done that I totally endorse.

I imagine that Susan's position is complicated, because in the social justice framework, in most interactions she is considered the less-privileged one, and then suddenly in a few of them she becomes the more-privileged one. And in different positions, different behavior is expected. Which, I suppose, is emotionally difficult, even if intellectually the person accepts the idea of intersectionality.

If in most situations, using tears is the winning strategy, it will be difficult to avoid crying when it suddenly becomes inappropriate for reasons invisible to her System 1. (Ironically, the less racist she is, the less likely she will notice "oops, this is a non-white person talking to me, I need to react differently".)

Here a white man would have the situation easier, because his expected reaction on Monday is the same as his expected reaction on Tuesday, so he can use one behavior consistently.

however, tone arguments aren't about epistemic norms, they're about people's feelings

One interesting idea here is that people's feelings are actually based on their beliefs (this is for instance one of the main ideas behind CBT, Focusing, and Internal Double Crux).

If I make people feel bad about being able to express certain parts of themselves, I'm creating an epistemic environment where there are certain beliefs they feel less comfortable sharing. The steelmanned tone argument is "By not being polite, you're creating an environment where it's harder to share opposing beliefs, because you're painting the people who hold those beliefs as bad people"

In other words, in an environment where people feel comfortable not being judged about their self-image, you're more able to have frank discussions about peoples' actual beliefs.

The only thing left to do is to steer in the right direction: make things around you better instead of worse, based on your intrinsically motivating discernment of what is better/worse. Don't try to be a good person, just try to make nicer things happen. And get more foresight, perspective, and cooperation as you go, so you can participate in steering bigger things on longer timescales using more information.

FWIW, this is roughly my viewpoint.

I don't think being a good person is a concept that is really that meaningful. Meanwhile, many of the systems I participate in seem quite terrible along many dimensions, and not-being-complicit in them doesn't seem very tractable. There were points in the past where I stressed out about this through the 'good person' lens, but for the past few years I've mostly just looked through the frame of 'try to steer towards nicer things.'

I realize this may seem in tension with some of my recent comments but I don't see it that way.

I do feel like there's still some kind of substantive disagreement here, but that disagreement mostly doesn't live in the frame you've laid out. (Some of it does, but I'm not sure whether those particular disagreements are a crux of mine, still mulling it over)

I think individual people should do their best not to see things in the lens of 'am I a good person' or 'are they a good person?', but engineers building systems (social or technological) need to account for people's tendency to do that. (Whether by routing around the problem, or limiting the system to people with common knowledge that they don't have that problem, or building feedback loops into the system that improve people's ability to look at critiques outside the lens of 'bad person-ness')

The phenomenon of people changing the subject to "Who is, or is not, Bad" is a very serious problem that prevents us from making intellectual progress on topics other than the question of who is Bad. (And actually, that topic as well.) It can be a serious problem even if the people who do it are not Bad!

> It's clear why tone arguments are epistemically invalid. If someone says X, then X's truth value is independent of their tone, so talking about their tone is changing the subject.

If tone is independent of truth then it should be possible to make a truth-compliant and tone-compliant comment. That is the thing should be bad even if you sugarcoat it. you don't need ot be anti-tone to be pro-truth althought it is harder to be compliant on both and people typically need to spread their skills over those goods. There is the kind of problematic thing where any negative implication is forbidden. But then tone and truth are mingled.

Sometimes I value safety a lot and I have pretty general principle that discussions should be safe. I would totally lie to prevent my face from being punched. But even if I did that I don't think I would be acting in bad faith. If I am going to be talking about a touchy subject I need to be certain that I won't be touched in anger even if the other participant becomes furious. Bad faith can come to play if I exploit the safety boundaries to "be winning". But I don't need to be able to withstand all truths to be able to have a conversation. Some truths might be devastating and you as the messenger might not be committed to build me back up to working condition (ie building myself back to psychological working condition is my personal problem). So in a touchy subject conversation I might have a edge condition that I not be devastated and need at each step be sure that I am not left devastated. Sure it's tricky if I can't communicate that my participating is limited because of psychological self-preservance and I might not be always be aware that this is going on.

If someone doesn't share the commitment that everything that could be destoyred by the truth should it doesn't mean they are the champion of falsehood. That you trip over comfort safe guards in the name of truth doesn't mean you want to optimise for most tortorous discussion. If one person is truth ambivalent and the other person is comfort ambivalent it means a comfort-truth solution might exist and trying to enforce comfort-ambivalence over comfort-seekment is likely to be ineffective. If one ends up playing prisoners dilemma between a truth seeker and a comfort seeker then the quadrant to shoot for might not be the highest in truth value.

I definitely notice this pattern of "feeling like the thing someone is pointing at is them pointing at me. then having to defend it to defend myself."

I move that's similar to the Calvanism concept that I've done (but makes me feel less like shit) is something like forgiving myself for making mistakes, and realizing that my mistakes don't define me. Thus when someone points at something that feels like one of my mistakes, I can see that as them pointing at the mistake, instead of me.

This is all a framing game of course, the Calvinism frame is implicitly "My mistakes do define me, and I'm evil, so lets' just try to be a little better." This framing is "My mistakes don't define me, and I'm usually good, so I'll just try to be a little better." When I can seperate my self-concept from my mistakes, I can work to make less mistakes over time, but from a frame that feels I'm inherently good and feels much better to me personally.

In "Transforming Yourself", Steve Andreas talks about the importance of including "counter examples" in your self-concept, to make them more robust and be able to notice your mistakes instead of having motivated reasoning to maintain your self concept. I think this is the technical mental movement that underlies the intuition above.

I think I see a motte and bailey around what it means to be a good person. Notice at the beginning of the post, we've got statements like

Anita reassured Susan that her comments were not directed at her personally

...

they spent the duration of the meeting consoling Susan, reassuring her that she was not at fault

And by the end, we've got statements like

it's quite hard to actually stop participating in racism... In societies with structural racism, ethical behavior requires skillfully and consciously reducing harm

...

almost every person's behavior is morally depraved a lot of the time

...

What if there are bad things that are your fault?

...

accept that you are irredeemably evil

Maybe Susan knows on some level that her colleagues aren't being completely honest when they claim to think she's not at fault. Maybe she correctly reads conversational subtext suggesting she is morally depraved, bad things are her fault, and she is irredeemably evil. This could explain why she reacts so negatively.

The parallel you draw to Calvinist doctrine is interesting. Presumably most of us would not take a Christian preacher very seriously if they told us we were morally depraved. As an atheist, when a preacher on the street tells me this, I see it as an unwelcome attempt to impose their values on me. I don't tell the preacher that I accept the fact that I'm irredeemably evil, because I don't want to let the preacher browbeat me into changing the way I live my life.

Now suppose you were accosted by such a preacher, and when you responded negatively, they proclaimed that your choice to defend yourself (by telling about times when you worked to make the world a better place, say) was further evidence of your depravity. The preacher brings out their bible and points to a verse which they interpret to mean "it is a sin to defend yourself against street preachers". How do you react?

Seems like a bit of a Catch-22 eh? The preacher has created a situation where if I accept their conversational frame, I'm considered a terrible person if I don't do whatever they say. See numbers 13, 18 and 21 on this list.

steer in the right direction: make things around you better instead of worse, based on your intrinsically motivating discernment ... try to make nicer things happen. And get more foresight, perspective, and cooperation as you go, so you can participate in steering bigger things on longer timescales using more information.

This seems kind of like... be a good person, or possibly be good, or do good. And I can't square it with:

accept that you are irredeemably evil

If you're irredeemably evil, your discernment is not to be trusted and your efforts are surely futile!

Is jessicata writing as if being a good person implies being thoroughly good in all respects, incapable of evil, perhaps incapable of serious error, perhaps single-handedly capable of lifting an entire society's ethical standing? That's a very tall order. I don't think that's what "good person" means. I don't think that's a reasonable standard to hold anyone to.