Leaving Orbit

post by Rob Bensinger (RobbBB) · 2021-12-06T21:48:41.371Z · LW · GW · 17 comments

Contents

17 comments

Duncan says:

At CFAR workshops, we often installed this as an explicit local norm: "If at any point you want to convey 'hey, I'm bowing out of this conversation now, but it's not intended to be any sort of reflection on you or the topic, I'm not making a statement, I'm just doing what's good for me and that's all,' you just say 'Leaving orbit!' and people will wave to you and keep on talking."

This strikes me as good social tech to try out on LessWrong (linking to this page to define the term).

Basic rationale:

  1. It's important to be able to randomly exit conversations. Otherwise, people won't add as much useful stuff to conversations in the first place (lest they be trapped).
  2. Sometimes people want to leave a message to indicate they're exiting. But it can be tricky to phrase such messages without implying one or more of: 
    • "I think my last message is the last word in this conversation, and has resolved all outstanding issues."
    • "I think you have raised no important or worth-responding-to points since my last message."
    • "I think that you as a person are unworthy of my time."
    • "Your last messages convinced me; I agree with you."
    • "Your last messages didn't convince me; I disagree."
    • "I have no idea how to object-level-respond to your messages."
  3. (Maybe some of these are things you want to communicate in some contexts, but it's good to not have to communicate them!)
  4. So having a very generic placeholder thing to say here is useful. "(Leaving orbit.)" seems friendly and generic enough to suit the bill.

Caveat: I think it's also 100% fine to exit conversations without leaving a message at all.

I generally think society has far too many "do this or you're bad" norms, especially automatic norms. A bad outcome would be if "leaving orbit" (or similar) came to be seen as an expectation or requirement, rather than as an option. If this idea catches on too much, I suggest deliberately exiting more conversations without leaving any message, so as to keep this option from ossifying into an expectation. 🙂

17 comments

Comments sorted by top scores.

comment by Wei Dai (Wei_Dai) · 2021-12-07T05:00:51.692Z · LW(p) · GW(p)

It’s important to be able to randomly exit conversations. Otherwise, people won’t add as much useful stuff to conversations in the first place (lest they be trapped).

I used to think the opposite [LW · GW]. I'm no longer so sure but it's at least not clear which position is right. Yes, if people couldn't randomly exit, that has a cost in terms of some people being more reluctant to start/join conversations in the first place, but doesn't the same apply for many rationalist norms? It also has benefits in terms of attracting people who like knowing that a conversation won't just randomly end without them knowing why, and in terms of providing valuable info to the audience about why a conversation ended.

I really wish we could do an experiment to gather some empirical data about this, perhaps by implementing this 12-year-old feature request [LW(p) · GW(p)].

Replies from: Joe_Collman, pktechgirl
comment by Joe_Collman · 2021-12-07T23:10:06.980Z · LW(p) · GW(p)

I was going to suggest something along the lines of your feature request. In an ideal world, I'd like it to tie in to a system that reminds the parent poster of open responses they wished to address (not through in-your-face notifications, but kept in some easily accessed list/priority queue).

I don't generally worry at all about the etiquette of actively deciding to disengage online. What I do worry about is that I/others will forget to follow up on productive exchanges.

Generally, the people I'm most eager to get responses from have many high value ways to spend their time. To the extent that not responding is efficient for them, I have no problem with that. It only seems to be a net loss if they forgot/lost track.

comment by Elizabeth (pktechgirl) · 2021-12-07T07:11:02.875Z · LW(p) · GW(p)

Agreed. There are benefits to low-cost exit, but also costs, and which wins out depends on the situation.

comment by gjm · 2021-12-06T23:45:27.605Z · LW(p) · GW(p)

There's another term already used on LW for (I think) exactly this purpose: "tapping out [? · GW]".

"Tapping out" derives, I think, from a context where it specifically indicates acknowledging defeat, whereas the LW term specifically wants not to. "Leaving orbit" is more neutral in that respect.

Neither of them really conveys to an uninitiated reader what the intended meaning is. That may be an unavoidable problem, though.

Replies from: jmh, AnthonyC, jimmy
comment by jmh · 2021-12-08T00:47:38.348Z · LW(p) · GW(p)

Unavoidable does seem likely but there may be a positive aspect as well. Since it will be a term that is cryptic to the uninitiated they are likely not to feel confident assigning any particular interpretation to the exit. And, one hopes in this type of environment, will produce a query about the term without feeling anyone was trying to hide something or exclude someone.

comment by AnthonyC · 2021-12-07T03:25:15.500Z · LW(p) · GW(p)

For an uninitiated reader, "tapping out" at least has the advantage of a pre-existing non-LW use that, even if you don't know it, makes the phrase easily searchable.

Replies from: gjm, RobbBB
comment by gjm · 2021-12-07T04:08:29.482Z · LW(p) · GW(p)

Yes, but at the cost of being actively misleading to uninitiated readers because in its original context "tapping out" means admitting defeat, whereas a large point of the LW usage is to have a way of saying "I'm outta here, and I am explicitly not either admitting defeat or claiming victory".

comment by Rob Bensinger (RobbBB) · 2021-12-07T03:47:56.262Z · LW(p) · GW(p)

Yeah, I would favor "tapping out" if it felt more neutral to me. 'Tapping out', 'bowing out', etc. sound a little resentful/aggressive to my ear, like you're exiting an annoying scuffle that's beneath your time. Even the combat-ish associations are a thing I'd prefer to avoid, if possible.

Replies from: jmh
comment by jmh · 2021-12-08T00:50:22.003Z · LW(p) · GW(p)

"Sorry, gotta go now."?

Or perhaps a phrase the Koreans say "I'll leave first."

comment by jimmy · 2021-12-09T18:48:41.705Z · LW(p) · GW(p)

"Tapping out" derives, I think, from a context where it specifically indicates acknowledging defeat, whereas the LW term specifically wants not to. 

 

Good luck with that.

Things are what they are, regardless of what you'd like them to be. Any time you have pressure to avoid admitting defeat, trying to get around that by calling it "leaving orbit" is just going to make a euphemism out of "leaving orbit". Try going to a jiu jitsu gym and saying "I'm not tapping out, I'm just leaving orbit" and see how that goes. "Yeah, sure thing buddy".

People tap for all sorts of reasons, and sometimes try to explain their taps. Sometimes it's "Tap. I need to use the bathroom right now", sometimes they'll say "I'm just too tired", sometimes it's "That was a crank, not a choke". These can all be true and these can all be ego defense, and the way jiu jitsu gyms have solved this is with the common knowledge that a tap is a tap. Your defense may be valid, and people may even accept it, but at the end of the day the gym functions because of the recognition that you have no right to an ego defense and if you tapped you tapped. The "Oh, I'm not tapping, I'm just 'leaving orbit'" approach is the polar opposite; it's an attempt to guarantee ego defense which cannot and should not be guaranteed.

The reason the former works better than the latter is that the problem is aggression and poor handling of bruised egos, not bruised egos themselves.

Any time you're trying in part to demonstrate your skill and dominance (even if you're not supposed to and don't want to see yourself as doing this), you are going to find out if your self perception is justified and it's gonna hurt when it's not. If you try to cut this feedback path, you will fail. It will still feel bad to have to "leave orbit", and people will still know that the legible and transmissible facts are that you wanted to stop when you were in an uncomfortable position. To the limited extent that you do manage to disrupt this flow of information, you'll have hindered people's ability to learn when they should assert less and listen more. This is not a good thing to shoot for.

On the other hand, a norm of "A tap is a tap" normalizes the very normal thing of defeat. We all tap, and it's really hard to learn without doing so. In a jiu jitsu gym with a good atmosphere, you can roll when tired and tap due to exhaustion. You can put yourself in weak positions to play with defending weak positions, and no one will think less of you when you do, because tapping is normal and they probably recognize that you're purposely allowing yourself to be more vulnerable than necessary. Yes, you're "admitting defeat" to an extent, in a way, and that's okay. The only time when it hurts to tap is when you've been posturing above your level, and that discomfort is the very important incentive to be appropriately humble.

There's more that goes into a healthy atmosphere were people are willing to risk and comfortable admitting a form of defeat, of course. It has to actually be clear what to do with that pain of defeat, for one. As much as "Leaving orbit isn't evidence of defeat!" is under-interpreting the tap, "Tap means I'm better" is over-interpreting. There has to be that recognition that "Gym taps don't matter".

But "A tap is a tap. You can have your excuses, but we're not obligated to take them seriously" is the piece that keeps things honest, and it's what keeps a mutually respectful and humble equilibrium stable.
 

Replies from: gjm
comment by gjm · 2021-12-10T01:55:09.832Z · LW(p) · GW(p)

On the one hand, I think there's a lot of truth in what you say.

On the other hand, discussions between reasonable people should not be dominance contests, and being shown that you were previously wrong (and hence becoming ... how shall I put it? ... less wrong) should be a thing you're glad of rather than a humiliating defeat. Our stupid monkey-brains make it difficult to operate that way, but we should damn well be trying to.

(Whereas ju-jitsu fights are, I take it, always and essentially dominance contests; a fight is a thing you win or lose, and there's really no way it could turn into some sort of in-principle-cooperative search for the best ju-jitsu moves or anything like that, without ceasing to be a fight.)

((This suggests a possibly useful analogy. Consider another usually competitive endeavour, namely chess. Suppose Magnus Carlsen and Ian Nepomniatchi decide, once Magnus has finished slapping Ian around in the world championship, to do some deep analysis and try to figure out who's winning in some particular line of the Petroff Defence. They might do it by sitting opposite one another at the board, with Carlsen playing the white pieces and Nepomniatchi playing the black pieces, and both trying to get the best position they can. But unlike the match they are playing right now, what they are doing will not, or should not, be a dominance contest, and sometimes Nepomniatchi will suggest a better move for white or Carlsen will suggest a better move for black, and if Carlsen keeps coming out with a clearly better position their conclusion will be "this opening is good for white", not "Carlsen is stronger than Nepomniatchi". And they will be better able to do this if, e.g., when one of them starts to get tired they can just say "I've had enough for now" without any presumption that they're just trying to avoid getting their ego bruised by the other's manifest superiority. I think intellectual argument is, or should be, more like chess than like ju-jitsu, and at least some of the time its goal should be "find the truth" rather than "beat the enemy". And I think this idea is quite central to what Less Wrong is supposed to be about.))

In pursuit of that noble-if-possibly-overoptimistic goal, there is value in having norms that explicitly make discussions less dominance-contest-y. "You can leave a discussion without either claiming victory or admitting defeat" is an attempt at such a norm.

Establishing such a norm, even if possible, wouldn't completely stop "tapping out", whatever language we use for it, being perceived as an admission of defeat, or as a dishonest attempt to avoid admitting defeat. Monkey-brains, and all that. I don't think that should stop us trying.

It might, as you say, make it harder to learn via the path of getting beaten and feeling the pain. It seems plausible to me that that's outweighed by the benefits of making conversation less adversarial.

(I'm more bothered by the problem someone else mentioned, that you can't simultaneously have a norm of "anyone can leave at any time and it's not an admission of defeat or anything" and a norm of "you can reasonably expect that if you get into a discussion it won't abruptly get dropped for no adequate reason", and those are both reasonable things to want.)

Replies from: jimmy
comment by jimmy · 2021-12-12T20:18:15.814Z · LW(p) · GW(p)

I agree with you on what it feels like when things are solved, but not on what's going on beneath the surface when it is, and therefore not on how to get (or stay) there.


"Not about dominant contests" is actually what good jiu jitsu gyms feel like too. The bit in your chess example where people offer helpful suggestions to their "opponents" actually happens (too much, sometimes). When your attention isn't so focused on "Who is better", then you start to disidentify with the techniques performed and instead of hearing/saying "Wow, you're so good" it turns into things like "Wow, that choke was so tight/unexpected/etc", and you're back to talking about jiu jitsu itself. This absolutely does work better when "I've had enough for now" isn't taken as a dishonest and defensive response to ego bruising, and jiu jitsu gyms mostly succeed at this, though obviously not with 100% success. The point is that you can't get there by fiat, and that attempting to force interpretations that might not jive with the evidence isn't a solution.


It's also worth noting that even in rationality discussions, the structure behind the solution is "Enough security to let 'who better?' fade into the background where people make their own private judgements without desire to thumb anyone's scales or over-interpret", not "The question is somehow rendered completely irrelevant and left uncomputed". If the last several times you disagreed with someone, it turned out that the disagreement was because you failed to see something they saw, you'll probably recognize that it makes more sense to prioritize "understanding their perspective" over "conveying your own", at least until you have some specific reason to think that this time will be different. "Who is better?" is an over-simplification, and actually thinking in those terms would be a sign that your thinking is locked up, but it is an accurate oversimplification of "If I were to disagree with them, I would probably be wrong", and these things are important to track. Because as much as it is a virtue to form our own beliefs and challenge supposed authorities, there is still the question of how much effort ought to be spent trying to understand and charitably frame the perspective offered by a certain person before dismissing it as "[most likely] wrong". If you sign up for a class and your teacher says "Your objection is actually wrong, and we'll go over that next week", you'll likely say "Okay", for example, since you expect that to actually mean that your objection is likely wrong and will be addressed when it is appropriate. If you're talking to some rando who has earned no credibility in your eyes, "Okay, I'll listen to you for another week and trust that you're likely right until then" is less likely, and for good reason.

It's not that we want people to be able to "learn from pain of losing" instead of "learning without framing it as losing". The latter is absolutely preferable, and unfortunately we cannot ensure it by fiat. We have to create norms that incentivize people to choose cooperative framings over adversarial framings.

One piece of that is to not remove the natural incentive to do so. If I have a polite conversation with you about why you took "my" toy truck, I might realize that was mistaken and we can stay friends, no harm no foul. If I come at you with an attitude about it, I'm going to feel much more humiliated when your case turns out to be rock solid and I look like a jerk, and that's a strong incentive for me to be nice and charitable about things in the first place.

The concern I have with "leaving orbit" as an explicit way of saying "Oh no, I'm not losing here", is that it allows people to be less careful to stay cooperative when they know they can always save face by denying they lost and that people will be expected to respect that. That's not to say we need to force people to "admit when they're losing", and that is actually bad for the same reasons.

When you want people to stick on the object level and not worry with what may be implied about "who better?" because it doesn't matter, then we want norms that refocus attention on the object level and discourage fretting over "who better?" because it doesn't matter. Not "I tap out", just "Good point, I'll think about it". Not "Definitely not tapping out, just leaving orbit", just leaving orbit. When that doesn't feel good because it "feels like losing", then that's the sign that things weren't purely cooperative from the start, and I'm uneasy about enabling people to ignore that error signal, especially when dojos that have done a very good job solving that problem in (IMO) significantly harder scenarios have done the opposite.

comment by Dagon · 2021-12-07T04:26:48.614Z · LW(p) · GW(p)

Online, asynchronous communication has less need for such a thing - there are two better options available which don't work as well in live interactive discussions.

  1. say nothing, just go. Feels rude in person, but not so much online, when people come and go all the time.
  2. use more words to say why you're leaving. since you're not taking up time/bandwidth that someone else can use in live conversation, you can opt for clarity.

saying "tapping out" or "leaving orbit" or "I'm done with this topic for now" or any other non-specific departure can feel somewhat accusatory, or like an attempt to get the last word in an un-rebuttable manner.  it's kind of content-free, and doesn't serve any informational purpose. 

That's necessary in person - there are plenty of content-free norms, and silent departure is likely to be worse.  

Replies from: RobbBB
comment by Rob Bensinger (RobbBB) · 2021-12-07T20:04:38.113Z · LW(p) · GW(p)

I think in most cases with public, online, asynchronous communication, it probably makes the most sense to just exit without a message about it.

In a minority of cases, though (e.g., where I've engaged in a series of back-and-forths and then abruptly stopped responding, or where someone asks me a direct Q or what-have-you), I find that I want an easy boilerplate way to notify others that I'm unlikely to respond more. I think "(Leaving orbit [LW · GW]. 🙂)" or similar solves that specific problem for me.

comment by jbash · 2021-12-07T15:02:19.966Z · LW(p) · GW(p)

Is it necessary to come up with a two-word phrase that won't mean anything to anybody who hasn't had it explicitly taught to them? Why not say something like "hey, I'm bowing out of this conversation now, but it's not intended to be any sort of reflection on you or the topic, I'm not making a statement, I'm just doing what's good for me and that's all"?

Although honestly that literal text sounds really passive-aggressive, and I would read it to mean "you guys are an annoying waste of time, you will never get anywhere, and I have better things to do". And I suspect I would start to attach that same meaning to any code phrase, regardless of what people claimed it was supposed to mean. Especially since this isn't a temporally constrained CFAR workshop where everybody is briefed on the way in the door.

Also, I think that even talking about either using a code phrase or to spell it out inevitably pushes toward that being a norm. Online discussion in general already has a very effective, functional norm that, unless you've made some explicit commitment, you can just disappear at any time, without any implications about why. Why mess with it?

On edit: I can't believe I missed what Dagon said. That.

Replies from: RobbBB
comment by Rob Bensinger (RobbBB) · 2021-12-07T20:18:45.567Z · LW(p) · GW(p)

Why not say something like "hey, I'm bowing out of this conversation now, but it's not intended to be any sort of reflection on you or the topic, I'm not making a statement, I'm just doing what's good for me and that's all"?

That seems fine too, if I feel like putting the effort into writing a long thing like that, customizing it for the particular circumstances, etc. But I've noticed many times that it's a surprisingly large effort to hit exactly the right balance of social signals in a case like this, given what an important and commonplace move it is. (And I think I'm better than most people at wordsmithing this kind of thing, so if it's hard for me then I worry even more about a bunch of other people.)

Even just taking the time to include all the caveats and explanations can send the wrong signal -- can make a conversation feel more tense, defensive, adversarial, or hypercautious, since why else would you be putting so much work into clarifying stuff rather than just giving a chill 'bye now :)'?.

Avoiding that takes skill too. I think this is just a legit hard social thing to communicate. Having another tool in my toolbox that lets me totally ignore one of the most common difficult things to communicate seems great to me. :)

(And indeed, with your second paragraph I see that you're spotting some of the issues. We can just pre-despair of there being any possible solution to this, but also maybe the jargon would just work. We haven't tried, and jargon does sometimes just work.)

Replies from: RobbBB
comment by Rob Bensinger (RobbBB) · 2021-12-07T20:18:59.488Z · LW(p) · GW(p)

"I suspect I would start to attach that same meaning to any code phrase" and "I think that even talking about either using a code phrase or to spell it out inevitably pushes toward that being a norm" are both concerns of mine, but I think I'm more optimistic than you that they just won't be big issues by default, and that we can deliberately avoid them if they start creeping in. I'm also perfectly happy in principle to euphemism-treadmill stuff and keep rolling out new terms, as long as the swap is happening (say) once every 15 years and not once every 2 years.