But it becomes problematic when it inevitably stops being about helping orphans and becomes a backscratching club that has little to do with the original mission. Which is why people who actually care about orphans become disillusioned about the causes they were once passionate about. It's when they see that actually helping orphans has been replaced with a witch-hunting campaign against those using the word orphan in any kind of a metaphor.
I'm honestly not sure. I used to think all organizations existed to do their mission – that's probably due to my being on the spectrum – but then over and over I saw evidence that this model does not match reality. Then I think at some point my model may have been overadjusted towards cynicism; I'm still not sure to what degree I'm overcompensating for the previous naivety, if at all.
I think many organizations are *started* mostly with the goal of pursuing stated mission. I think they then continue for a while working primarily towards the mission, but I do also think that most of them remain small and thus not very effective (eventually dissolving entirely), and the ones that do grow large and influential become victims of their own success in that they become less and less about the mission and more and more about politics and backscratching. I think an organization's best work towards the mission is often done during a pretty brief period: after it has grown large enough to be able to actually do something but before it's grown so established that it's crippled by its own success and overtaken by politics.
I don't have any hard data to support this hypothesis, and as I said earlier, I'm not even sure to what extent it is accurate. This is just the model that I currently estimate as slightly more likely to be accurate than competing models based on anecdotal evidence and general observation. But I'm highly uncertain about my own estimate of its likelihood.
That said, I still need a model to guide my own behaviour wrt various organizations, so I have to pick one even if I am not certain about its accuracy.
Well, I suppose everyone's heard about last week's incident by now, and you probably have a pretty low opinion of us survivors. And, all things considered, perhaps we deserve it. Perhaps we panicked and resorted to cannibalism a bit early. But you weren't there. You don't know what it was like. I just want you to hear our side of the story before you go judging us.
When the six of us got into the elevator on that fateful day, we had no idea what was going to happen. We thought we were just going to take a little ride from the 12th floor to the lobby, just like every other day. Do you think we knew that elevator was going to get stuck between floors? Do you think we got into the elevator saying, "Hey, you know, we should eat our good old pal Jerry Weinhoff from Accounts Payable"? Of course not.
During those first few minutes after the elevator car lurched to a stop somewhere between the seventh and eighth floors, we were still civilized human beings. Everyone kept his cool. We tried pushing the emergency button. ...
One word enters your mind: survive. Survive!
I have no idea how long we'd been marooned when we started edging toward Jerry. Twenty, thirty minutes, time has little meaning when you're in a situation like that. It wasn't a spoken decision, either. We just all looked at each other and knew something had to be done.
It might have been an animal act, but it had a certain logic. Jerry lived alone and had nobody special in his life—no kids, no wife or girlfriend, and his parents had died a long time ago. And, most important, he was the biggest. We figured there was enough meat on him to keep the rest of us alive for days, maybe weeks.
... Maybe it was savage. Maybe it was an animal act. But human teeth are pointed and sharp in front for a reason. Besides, we had no way of knowing that, at that very moment, an Otis Elevator repairman was working to free us.
When a different party takes office, the new party "appoints" 3,000 partisans to engage their political apparatus in government. The Appointed (Anointed?) are expected to further the party's ideology.
The 3,000 in the current clown car will be looking for work if Joe can't carry the torch.
I guess we were all guilty, in a way. We all shot him, we all skinned him, and we all got a complimentary bumper sticker that read, "I helped skin Bob."
> When I finally got home from work that day, some 50 minutes late, my youngest daughter Kellie ran up to me and gave me a big hug. She said, "Daddy, I'm glad you're home." Daddy, I'm glad you're home. It was at that moment I knew I'd done the right thing.
This sounds horribly familiar. People focus on the small good consequence of their actions, and utterly ignore the cost. I swear that "Compared to what?" is a fully complete counterargument to 90% of the drivel on the internet.
regarding it being surprising that there is a more stable solution than the random fair one, you might be interested in Ram Rachum's work on emergent dominance hierarchies in reinforcement learning agents: https://arxiv.org/abs/2401.12258
Rachum studies hawk/dove games and shows that in an iterated setting, agents that start off playing mixed strategies will tend towards always playing hawk or always playing dove, with the agent that always plays hawk randomly decided by the first game. This generalises to multiagent settings like the lifeboat game: the stable solution is to randomly decide for each pair of agents which will be hawk and which will be dove, and this usually results in a dominance hierarchy.
> They’ve heard this story before, so they decide to turn to cannibalism sooner rather than later. They agree to draw lots to determine the victim. Just as the first person is reaching for the lots, Albert shouts out “WAIT LET’S ALL KILL AND EAT BOB!”
As a Bob I can confidently say that any Bob worth their salt would shout: "Albert is trying to defect from a fair decision procedure! Let's eat him!"
Indeed, given social norms and culture, it is not completely clear whether there is a first-mover advantage or a first-mover disadvantage in this hypothetical scenario. When Albert is trying to coordinate people by shouting out “WAIT LET’S ALL KILL AND EAT BOB!”, he also creates a justification for the others to say "ALBERT IS EVIL, LET US HIT HIM!"
For the second time they need to eat someone, they all shout out LET'S EAT ALBERT because Albert has already proven he will shout out someone else's name. By eating him, you're protecting yourself; maybe the lifeboat will be rescued before there is the necessity to eat a third person.
Also you get to feel a little bit better if you survive. Sure, we ate Bob, and I genuinely feel bad about it, but when you think about it, it was really Albert's fault, and when we ate him we were justly punishing him for his selfish defection.
Or someone confident with Albert will say, “Let’s listen to Albert’s wise words on whom to eat next”. Would you object, thereby signalling your low confidence in your chances and making yourself a prime Schelling point?
I think it’s a disadvantage. If we imagine this as a vote to see who is eaten, then Albert has voted for Bob. But if Bob votes next, he will certainly vote for Albert, and then we have two people both in the same situation in terms of votes. I think in that case morality favours Bob
If I wasn't sure how many times we were going to have to iterate through this particular problem I probably wouldn't demonize the technique that I may need to resort to at some point.
Bob could also point out that, if they all go for him, afterwards they'll feel guilty and if they're not rescued, they will pick Albert next time for sure, to assuage the feeling.
Even it didn't work, you have a feeling of payback against Albert.
Yeah, could work. Though probably that guy would be more in the frame if it goes to R2. The others jump on him.
There actually is an old group therapy structure called The Boat Structure that models this process. Though you push people out rather than eat them to stop the lifeboat sinking. İt gets pretty crazy.
It's been a couple of decades since I did it last. But from what I recall...
* mark off a small area of the group room using mattresses
* 10-20 people sit inside it. The scenario is that you've escaped from a sinking ship and are in a lifeboat. However the boat has a small leak and someone will have to be thrown overboard (to certain death) so the others can live. Hopefully you're all rescued up soon.
* everyone has a turn to say why they shouldn't be the one, or to volunteer. Then there's a vote and the loser goes out.
* Continue. I can't remember how you vote on the last person, when it's just one v one. But you don't have to fight and there is just one winner.
Best I can do. PS - the process can bring up a lot of deep dark stuff so it's not a good idea to do it without trained facilitation.
Why waste such a good opportunity to have two guys to eat for the price of bringing one down? Since Bob won’t be much of a threat, let him live till you’re done eating the big dude, so he doesn’t rot.
However, do people coördinate against big dudes like that? After all, we have this notion that if you’ve got a problem with the big dude, the manly thing to do is to tell him to his face and fight him one on one, with no dirty tricks that might help you against his superior strength.
I guess this wouldn’t apply if the rest of the passengers were women—which would make him a glaring Schelling point—and the boat were roomy enough for them to surround him and attack him all at once, though it may still be easier to convince him to sacrifice himself out of chivalry.
When we get to the third level, where it’s about picking a Schelling point, and there’s too many possible Schelling points, we see what the advantage was in the drawing lots - it’s very symmetry makes it a more stable Schelling point.
one idea is that there's an anonymous note left that says, "let's all just eat bob". the next day, there's anonymous notes from everyone that lists everyone. and so on and so forth. this is one situation where i'm not sure a Land Value Tax would solve anything, unfortunately.
Exactly. And I think most groups would then vote to eat Albert. If not in the first round then in the next as he has three big marks going against him:
- he effectively killed bob
- he broke convention and isn't to be trusted
- he made each person's odds slightly less and resentment would surely impact future decisions
The first mover advantage for Albert far from gives him 100% of avoiding death. I'd say it highlight increases his chances of death.
> As a Bob I can confidently say that any Bob worth their salt would shout: "Albert is trying to defect from a fair decision procedure! Let's eat him!"
Getting to finish the first of those two sentences doesn’t sound very Bob-like to me. In my experience, you’re cut off and either ignored or punished for acting above your station; perhaps they’ll be offended by your big words they can’t be bothered to engage, or by your nerve in accusing anyone else of anything; perhaps they’ll just command you to calm down or to stop yelling.
In this case, they’ll probably cut you off with their knives or with their eager teeth.
When Albert shouts "hey let's eat Bob", shouldn't Bob shout "let's eat Albert instead"? Which would bring it to a vote, which Albert should lose*. Even if that fails, Albert should be the next Schelling point for the person to be eaten for having suggested Bob, not the blonde.
*Why should the majority prefer to eat Albert? Albert's effectively trying to move the distribution of victims from uniformly random to favoring whoever can think up a good strategy. But we know Albert can think up good strategies, which means this distribution is worse for the majority of non-alberts.
It's just a hypothetical scenario with a lot of simplification. IRL Albert wouldn't shout 'let's eat Bob' in the first place, they would probably talk with the non-Bobs one-on-one about how Bob is planning to coordinate eating them and how they should deal with this problem.
This simplified scenario plays an important role, namely, presenting the dilemma in the starkest terms possible. Thus eliminating such extraneous machinations you have posited.
No one is looking to pitch this as a vehicle for a Netflix mini series!
yes, that was my point with my comment. it's an abstract game not an actual situation hence options which would be available in the real world are not available in the scenario. Do note that were it an actual real world scenario, one-on-one talk would easily be possible when stuck on a life-boat, because they would take turns staying up looking for help in the night and when Albert is the one awake, he could wake up the people he wants to speak with.
I think the idea is that saying "Let's eat Bob" is a clear Schelling point, because it's the first thing anyone said and satisfies everyone (except Bob) completely, whereas letting Bob say "let's eat Albert" lets Albert say "Well let's eat Carol" and so on.
I know this is more complex than the metaphorical/analogous situations you're drawing here, but I really see *Albert* as the clear Schelling point. He's the only one, in fact, I would be comfortable killing and eating, due to his actions and willingness to defect for personal gain.
But if Albert then tries to move on to "let's eat Carol - no, let's eat Heather - no, let's eat Iolanthe" then everyone will agree even more strongly "let's eat Albert". Bob hasn't said "let's eat Carol", so Carol should support Bob in the "let's eat Albert" coalition.
They would want to disincentivize someone breaking from existing agreements in a way that counterfactually could have harmed them. (blah blah, Rawls veil).
Albert only proposed this because he predicted that the other people in the group would go along with it. However, this 1/9 chance to die — before Albert chooses — is worse than a 1/10 chance (because Albert doesn't pick himself), and thus you should refuse the deal even if there's technically a fact of the matter about who was chosen.
Seems like a classic prediction problem where it would have been better for everybody to pre-commit. Of course, this assumes that everyone believes Albert would choose randomly from the 9 others.
I can see Bob being a good Schelling point among actors who aren't rational enough and so they agree to Albert's plan.
However, I do think Albert is still also significantly more of a schelling point than Carol. He's the one who tried moving things so he would always win at the expense of others, giving everyone else a 1/9 chance of dying rather than a 1/10. Sure, if you think purely causally he's given you a 0% chance (except for Bob), but he chose randomly before his announcement.
Then he becomes a decent point of agreement for trying to manipulate others. (Which is probably part of why humans have the intuition of get rid of the schemer first)
Exactly this. Albert speaking first also asserts his dominance and position of leadership.
If you consider Albert is Donald Trump it becomes interesting when you ask “Well why didn’t everyone just eat Albert”. People aren’t equal even in theoretical scenarios and certain personalities will dominate. There’s a reason Trump uses projection on his enemies because it becomes almost impossible to be on the receiving end and argue that it’s actually the other way around.
Because at that instant, when he was the first and only to shout, you have 2 choices:
- Agree and have a 0% chance of dying
- Try to coordinate and have a >0% chance of dying
This doesn't work in an iterated game, but unless everyone pre-committed not to do this (e.g. by coming from a culture that vilifies this) then _at that point_ the clear best action is to eat Bob
Some people are willing to punish defectors even at a cost to themselves. Granted, the number who will do so when the cost approaches a 100% chance of dying may be small, but surely it goes up quite a bit as the chance of dying goes down. In any case, a quick decisive action from the one person on the lifeboat who has the courage of their convictions even unto death could convince the other seven people (who have to eat *somebody,* after all) and Bob (who likely doesn't need much convincing) to get behind the would-be hero and eat Albert.
But whoever suggest that they eat Albert would stand out, and be the next obvious choice. So people have an incentive to not be the first to nominate Albert.
Maybe I'm missing something but as per the comment you're responding to it would be Bob who suggests Albert and he really has nothing to lose here. So isn't it irrelevant that nobody else would suggest Albert, as Bob is already doing the suggesting for them?
Right. What Bob should do is spread the rumour that Albert has been trying to get everyone to eat Charlotte. Now, whether the group decides to align with the defector or eat the defector, Bob is safe.
This thread has taught me that if I ever do get into a lifeboat then it's important to start strategising immediately. While everyone else is still thinking about "I still can't believe our ship sunk" and "where's my luggage" and "did everyone get off okay", I'll be busy thinking about how to ensure I'm the guy who gets to eat everyone else.
(Ignoring all the metaphor for a second) Drawing lots is actually a stable equilibrium. Shouting out someone’s name incentivizes doing so as fast as possible (because otherwise someone else might shout out yours first). Drawing lots encourages waiting as long as possible because everyone’s life is on the line. Rationally you’d rather endorse the system that drags out the killing even if you could have a safe situation in round 1. Also, maritime law typically only protects you if you draw lots. (Edit: I’m wrong on this last point. See below)
You’re right, I’m several hundred years out of date on the law there (the Dudley case ruled even drawing lots as illegal). Good thing you replied before my next sailing trip.
I thought the Dudley case was at least a little ambiguous on that point. The defendants had killed the most junior member of the crew (the cabin boy) as a Schelling Point rather than drawing lots as customary, and the decision specifically called them out for choosing their victim consciously (arrogating authority that didn't belong to them, and setting a precedent that if upheld could act as a cover for killing out of animis) as well as for abusing their seniority to prey on their subordinate to their own benefit rather than sacrificing themselves to ensure the safety of those under their care.
The decision does also generally reject necessity as a defense for murder, but taken together it seems like the logic of the ruling would probably support killing a volunteer and would have a decent chance of supporting a unanimously voluntarily decision to draw lots.
Thanks for the link, so the victim in Dudley was called Richard Parker! Yann Martel must surely have chosen that name for Life of Pi with the case in mind.
Drawing lots is not a stable (Nash) equilibrium? If the possible actions are (a) support drawing lots, and (b) shout out the name of someone else as fast as possible; then given the provided assumptions any participant is better off switching to (b) if everyone plays (a).
There are rational arguments for why I dislike backscratcher movements, which I might or might not get to writing down someday.
But right here and right now, I just want to shout and to cry and to thunder against the concept! I abhor the notion, I hate coordination capture, I despise backscratchers and all they stand for.
If there was an eldritch ritual that would remove backscratching from the world in all its forms and replace it with acausal trades and updateless decision-theoretical coordination, I would gladly bleed myself to death and spend a year learning latin to perform it.
As I write this, I feel a shroud of anger, real red-violent anger, not just pretend writing-anger fogging my mind and tears welling in my eyes. DOWN WITH BACKSCRATCHING! FUCK THOSE BASTARDS!
Any movement that is, indirectly or not about backscratching incurs an ideological discount in my mind to compensate for it. If the OSC came to my town, I would care no more no less about orphans than before (except for optimizing effectiveness of my current orphan-caring activities based on accurate OSC sensibilization and advice) and any more effort that's asked on pain of social ostracization can DIE IN A FIRE WITH ME.
Before you finalise that ritual, please take some time to discuss the implementation details with the demon you're about to sell your soul to.
Is a family a backscratching club? If I am more willing to help my brother when he is in need than a random stranger, is that backscratching? Is that good or bad? Will that still be allowed after you perform your ritual?
How about friendship? One of my friends is looking for a new job, so I pull some strings at my employer to get him an interview -- he'll still have to pass the tests himself but he gets bumped to the head of the resume-filtering queue. We didn't *start* the friendship as a mutual aid pact, we started it because of similar interests and hobbies and stuff, but now that the friendship exists, being able to call in favors like that is one of the side benefits. (My employer benefits too, since although I can't say from direct experience how competent he is as an employee, I can at least testify that he seems smart, isn't lying about his current employment status, seems like a reasonably sane and reliable person in general, etc.)
Will your eldritch ritual only prevent me from helping my friend bypass the resume queue, or will it eliminate friendship completely? Next time I'm sitting at home bored and want to call a friend and ask if they want to go do something, is that still allowed or do I need to go through some app to give all available people in my area an equal chance to respond?
The unstated but obvious joke in Scott's description of orphan supporter clubs, is that they usually aren't started by some Machiavellian schemer who's deliberately trying to create a backscratcher club. Or at least that's not how they feel on the inside. They genuinely care about orphans and want to connect like-minded people together. Creating a social network of people who can then call in non-orphan-related favors with each other, is a nice side benefit. Will your ritual somehow eliminate the favor-calling opportunities while leaving the orphan support intact, or will it no longer be possible to start any kind of club or organisation or movement or any other structure of people which would inevitably have backscratching as a side effect? Will the net effect be more or fewer orphans getting helped than before?
Jorge Louis Borges wrote a short story called 'the Bribe'.
A young person at a North American university is competing for some position. He knows one of the older professors, whose vote can decide his future, considers him less qualified than his competitor. But he also realizes that the older professor is bound by North American puritan morals.
So he finds a way to insult the older professor publicly. Since the professor now knows that he has a personal motive to vote against the main character, so he votes in favor of him and he gets the position. To quote the story:
"... I realized, my dear Winthrop, that you are ruled by that curious American passion for impartiality. You wish above all else to be 'fairminded.' Precisely because you are from the North, you tried to understand and defend the South's cause. The moment I discovered that my trip to Wisconsin depended upon your recommendation to Rosenthal, I decided to take advantage of my little discovery. I realized that calling into question the methodology that you always use in your classes was the most effective way of winning your support."
It could be that creating a backscratching club sometimes *is* the most reasonable solution to a coordination problem, in which case there's no reason for anyone to get too pissy about it when they notice one.
One of the reasons that I dislike backscratching is that it arises organically from optimization processes with perverse incentives that spawn deviant children organizations.
I want to promote orphan-supporters insofar as it helps support orphans more than it detracts from other causes, which I remain aware of when deciding to promote OSC content based on how I expect the repercussions on society at large to be beneficial. This allows me to promote OSC's orphan-supporting content without promoting OSC's word-banning content.
Before the OSC appears, if people would coordinate using eldritch means like acausal trades, they wouldn't spawn word-banning movements as a byproduct. Of course, I want to discount how much I prefer acausal trades to backscratching as a coordination tool insofar as it has similar negative byproducts; I'm not currently aware of any significant ones (except overhead) nor do I expect there are.
The problem is that, by definition, there is no mechanism for acausal trade to get established. Causation in relevant because it’s how the world works.
I guess a follow-up question would be, what property of acausal trade as a coordination mechanism makes it not spawn word-banning movements?
My intuition here is that the word-banning here is not a natural byproduct of causal coordination, it's a byproduct of Heather being Heather. In the first place, if your goal is actually to help orphans, you can't really _coordinate_ with Heather per se, that word implies having common goals. The most you can do is get Heather to press "cooperate" in some specific situations dictated by decision theory. Which is what I meant by "backscratching with epicycles".
It's even explicit in the article that Orphanism was not optimized to be about orphans in the first place. It's not the word-banning that's a byproduct, it's actually the orphan stuff. Maybe if it worked via acausal trade, they wouldn't need the fig leaf of caring about orphans in the first place, and that's, I dunno, probably good?
(Sorry for the swearing; I generally dislike social signalling, but I didn't really know how to express that I feel strongly negatively about this topic, and stating it plainly doesn't quite convey the same message.)
I mean your reaction is a moral one: such a thing is bad and makes society unsafe. You seem strangely (to me) reluctant to cast it as a moral stance explicitly so instead you're railing against them at a personal level. But the reason you dislike them is that they're obviously horrible: they do not solve for everyone's mutual safety which is what institutions are supposed to do. Even self - serving institutions like corporations are expected, in modernity, to pay *some* attention to the safety of everyone else (that's we have a bunch of laws about them and why we shame the bad ones), and the clubs in this article seemingly do not, hence they are repugnant.
I think the whole language of “acausal trade” sprung up because some people who had convinced themselves of a Humean form of rationality that is purely decision theoretic and leaves no room for morality as a separate norm then re-discovered Kantian arguments, and wanted to cast them in the language of trade and rationality rather than morality.
I just wanted to chime in that I found your comment very insightful. I never made the connection that it was easy for me to accept acausal trade because I was already a believer in the categorical imperative. Thanks!
To the extent I understand it, I believe the mechanism of acausal trade getting established is a reasonable empirical belief that one's mind-state is mirrored by one's trade partner. The idea being that if you can reliably assume that your counterparty will behave as you do (say, because you're both fresh out of the box serially produced AIs of the same model, or identical twins or something, or you just know the person really well and they think like you do), then defection *stops being the dominant solution* because you can assume that any preference for defection on your part will be symmetrically engaged in by your trade partner, and if only symmetrical outcomes are allowed -- i.e. cooperate-cooperate or defect-defect, then cooperate-cooperate is the dominant solution because all of a sudden you're playing stag hunt where the *individually welfare-maximizing choice* of both parties is to cooperate.
This differs from the Categorical Imperative in that the latter acknowledges that cooperate-defect is part of probability-space, it just says that you shouldn't do it as a normative matter.
I don’t think of acausal trade, at least the way it is normally talked about, as requiring such a degree of similarity that it makes asymmetric outcomes literally impossible. I think of it just as requiring that the parties are in some sense recognizing the similarity of their reasoning process in a way that puts the same rational requirements on their decision, such that asymmetric outcomes would necessarily be irrational, though not impossible. But this is basically Kant’s idea - all rational beings are under the same requirements, such that if they were all rational they would do the same thing, and thus doing something that wouldn’t be effective if universalized is irrational (though not impossible).
I appreciate the clarification re: the Categorical Imperative. However, while I confess that I do sometimes think this way and this it may be a decent account of virtue ethics, it seems self-evident that absent robust symmetry guarantees it’s not an account of *individual welfare maximization* and thus not of “rationality” at least as definitionally limited to such maximization (absent additional constraints like reputation effects or non-bounded iterated games) because it’s not robust to defection. Two-boxers really do get epsilon more cash. Defecting really does improve one’s individual position whether or not it’s reciprocal.
The virtue-theoretic / internal cognitive properties are part of why I think of “being a defector” as being something that one is rather than defection being something that one does, and thus there’s at least conceivably potential societal upside in, say, summarily executing everyone who uses a license plate cover to avoid payng tolls or other forms of traffic enforcement or doesn’t return their shopping cart. But that’s conceptually just extending the idea that ex post punishment is the correct way to minimize defection to a world in which defection is an incorrigible and inferable character trait, (basically trying to make group selection actually work). The fact that I’m habitually conformant to cooperative norms dorsn’t mean that I’m operating under a belief that I wouldn’t personally save a few minutes here or there if I ran red lights more often or routinely turned into the offramp to force my way into traffic at the last possible second.
Just so! If the backscratchers coordinate and we do not we will always be at a disadvantage and the backscratchers will take over. We must work together in solidarity to promote fair competition and individual decision making!
Unfortunately (or not), these kinds of organizations do actually serve a purpose beyond just giving members some kind of unfair advantage. While it is omitted from Scott's description, in the real world there is usually some more or less extensive vetting process before you can join one of these societies. Not so much as in ages past, I think (the remaining societies are hard up for members), but the 'costs' are less of the 'do silly rituals' sort than they are 'you actually have to be self sufficient and convincingly honest and trustworthy'.
This allows the organization to function as a reputation service. In a world before credit scores, before you could Google someone's name or check out their LinkedIn, you could go somewhere and demonstrate that you were a member of this organization (to the local chapter, by way of secret handshake or what have you) and so demonstrate some level of trustworthiness. Which, depending on the reputation of the organization in question, might well extend even to outsiders - ah, he's one of the Knights of Columbus, well they're strange Catholics but at least he's not a con man or a serial killer...
I believe that most backscratching clubs of the "fraternal mutual aid society" sort, such as the modern Freemasons, prefer that you join when you don't need aid. And in practice, I have a feeling that your ability to draw aid in the future will depend a lot on whether you have been perceived as willing to give aid when asked.
Yep. And these organizations need to do some checks, because one way they can fail is from the inside, the people who accept favors but don't help other club members.
Surely this is why Abrahamic religions have spread so far. "Ours is the one true God", "Other religions follow demons and false gods", "All other paths lead to damnation". And straight from the Quran "“whoever changes his religion, kill him”".
And as far as I can tell, this isn't a universal feature of religion - just the most successful ones.
It was useful unto necessity in some particularly competitive contexts, so there's lots of relevant instructions still in the scripture, but key portions have been commented out of leading builds ever since industrialization made winning a war unprofitable.
I'm not sure this is a good example, as Abrahamic religions have a strong (though certainly not universal) tendency to tolerate each other, even when "each other" is the dominant outgroup who they encounter. See the Muslim attitude towards "people of the book", or the Jewish attitude towards Noahides.
But also the long-term successful idelogies that take over large parts of the world enact norms that limit the ability of new memetic competitors from challenging them. The dangers tend to come when you have shifts in information technology that render those protections less effective.
For instance, consider the transition from paganism to monotheism where you have a rule blocking other religions from getting off the ground. And the Catholic church was able to effectively combat alternative interpretations by controlling the written word until the printing press. And notice that even seemingly nice messages about not excluding people or good Samaritans also has the secondary effect of detering organization along different fault lines.
I wonder if that's not something we are seeing today as the mechanisms that existing organizations used to limit the ability of people to spontaneously organize in different ways has been reduced by new infotech.
I guess the most interesting thing that stood out to me is that this is a "barberpole model" a la https://slatestarcodex.com/2014/04/22/right-is-the-new-left/. Whenever a Backscratchers' Club gets too successful like Daniel's, its like it never even existed at all, so the Heathers of the world try to form a new Backscratcher's Club within the old Backscratcher's Club. But if they're too successful, they get "couped" in turn by *another* new Backscratcher's Club sprouting within them. It's an endless cycle, because victory doesn't actually solve anything -- just make the ground fertile for the next conflict.
And this applies just as well to ideologies as it does to Backscratcher's Clubs! In the past, the pioneering Christian evangelists like Saint Paul must have believed that, once everyone agreed on Christianity, surely there would be peace right? No more wars between pagans & Christians, just Christians being Christian to each other. All alike, with nothing to fight over. All supporting each other in the Backscratcher's Club of God/universal brotherhood.
But all that really happened was that the Christians started fighting each other over things within Christianity! Arianism, Nestorianism, Pelangiasm, Catharism, more than I can name; if the wars with Paganism seemed like a wildfire that would burn everything to ashes & eventually bring a sort of peace... then what actually happened was the ashes immediately turning into a new forest, that burned in a new wildfire, precisely *because* they were so alike that ideas could spread easily between them & create mass movements!
i.e. Something like Catharism would have made no sense to someone who's never even *heard* of Christianity -- but the moment everyone did, "Catharism" became something that was possible, something that could spread beyond the minds of a few religious scholars into the mass population & trigger an uprising. The monoculture was in fact *more* vulnerable to war, because it was more vulnerable to the explosive spread of ideas -- *any* ideas, not just the approved ones. The peace set the stage for war, *inherently by being the peace*. By connecting people together and making them one. All victory could ever do was set the stage for its own defeat.
And this keeps repeating throughout history: the periodic rebellions & explosions of Imperial China would be very interesting to try out through this lens, but a more familar example for most readers will probably be the effects of the printing press on European history. It printed the Gutenberg Bible, which perhaps led rather more directly to the Protestant Heresy/Protestant Reformation then the Church could have foreseen at the time. For the more ideas could spread -- and the more they were used to spread a monoculture where, if 1 "domino" starts to fall, all the rest will "catch" him -- the more peaceful it paradoxically looks: the field of dominos getting closer & closer & *closer* together. Right up until it all explodes. If everything linked together, none can flip without all flipping; all the rest will catch him.
But if one does come down with something & flip, all must flip with him: all the rest will catch *it*.
The applicability to the modern day & our connected world, social media & modern movements, is best left as an exercise to the reader.
/It was a glorious peace while it lasted, the 90s post-Cold War era. Too bad that it couldn't work; a Backscracther's Club that works for everyone is a Backscratcher's Club that works for no one. All it *can* do is connect people together & make it easier for them to talk about the need for a new Backscratcher's Club -- precisely because the old club brought everyone together. Precisely because it had won.
It's also common to see early Christianity as a backscratcher's club - its early spread was in urban contexts among people with a weak social support network, who could now enter into a club for mutual support and a feeling of moral superiority to outsiders.
> In the past, the pioneering Christian evangelists like Saint Paul must have believed that, once everyone agreed on Christianity, surely there would be peace right? No more wars between pagans & Christians, just Christians being Christian to each other.
Democratic peace also shows that this sort of peace mechanism can, in fact, work; not as a way to get a benefit at the expense of others, but as a way to preserve peace yes. The democratic peace theory (that democracies don't go to war with each other) holds up pretty well. And I don't think it's just because democracies have common rivals in dictatorships; from the 90s to the 00s democracies lacked rivals they considered a serious threat, but that didn't make them want to go to war with each other.
There's a trend of fewer interstate wars since WW2, which doesn't appear to be quite reducible to nations not going to war with each other because they're both democracies. North Korea is as dictatorial as it gets, but nobody expects it to attempt to invade the South again. China & Vietnam are both one-party Communist states, and the former briefly invaded the latter in 1979, but they've also been able to have peace since then.
There are many dictatorships that don't go to war of course, but there is a remarkable scarcity of democracies that do go to war with each other. The American Civil War is the only serious example I know of, and even there I'm not sure how democratic the South was, even if we only consider whites. Plus technical examples like Finland and the UK being at war in WWII because they were allied to dictatorships on opposite sides, but they didn't really fight one another.
Israel went to war mainly with the PLO and later the Hezbollah in Lebanon, which the Lebanese government couldn't control; Lebanese participation in the 1948 war was minimal. Germany wasn't a democracy, the emperor controlled foreign policy. At the time of the siege of Syracuse (is that the one you refer to?), Syracuse was ruled by a tyrant. Anyway, as I said in another comment, I'm not saying there have been no wars between democracies, but there have been few compared to wars involving autocracies, even in the recent era when base rates don't explain it.
It was the first World War that convinced men like Whitaker Chambers that socialism (rather than democracy) was necessary to prevent war. https://entitledtoanopinion.wordpress.com/2024/05/10/witness/ The Kaiser's Germany had elections, and the Social Democrats even obtained a majority, despite how much Bismarck would have hated that. It wasn't sufficient to prevent them from warring with autocracies like Russia, constitutional monarchies like the UK, and republics like France & the US.
Specifically, I think it was that national forms of socialism (trade unions, not actually a reference to the Nazis) had failed to prevent war, but international forms of socialism (communism) might work, right?
The Democratic Peace Theory can be salvaged by limiting it to *liberal* democracies and defining "liberal" restrictively. The most restrictive version is to require >50% of the adult population to be eligible to vote, which is a stronger requirement than it sounds at first because women are generally a little bit over 50% of the adult population and most large democracies only adopted Women's Suffrage during the interwar years.
Further epicycles can be added by excluding "young" democracies that have had democratic constitutions and elections for less than three-ish years, which defines away a bunch of the remaining counterexamples. But then we're left with "No wars between democracies in the past century or so, except for a bunch of wars that happened in the wake of decolonization or the fall of the Soviet Union, and don't pay too much attention to Latin America", which is a much weaker claim than "Democracies don't go to war against each other" is generally understood to mean.
I wouldn't require women's suffrage, but I would require it to be actually a democracy at least among free men: that (almost) all (non-slave where applicable) adult men can vote, elections are actually free, and elected officials actually control most policy, including foreign policy and the military. I wouldn't exclude young democracies, but I'd require that at least one free general election has actually been held; a secessionist militia that promises to establish a democracy, or perhaps holds an unfree election under a state of war, doesn't count. Under these conditions, not many count, and even fewer actually had serious fighting between democracies. (Compare the more recent lists at https://en.wikipedia.org/wiki/Template:Lists_of_wars_by_date )
Anyway, I'm not saying that war between democracies absolutely never happens, but that it's much rarer than either war between autocracies, or between democracies and autocracies. Democracy is among the most reliable things that prevent war. Other ones ones are alliance systems against external enemies, especially if policed by a superpower (but these risk a major war between the alliances, and often have localized but serious proxy wars), and nuclear weapons (but the risks of giving every country nukes to prevent war are obvious).
I'd actually bring Latin America as a positive example: its countries aren't part of tight alliance systems like most Northern democracies are (perhaps because they're too far from potential external threats to be too worried about them), yet it hasn't had a serious war since the fall of the Cold War military dictatorships.
(I don't like the term "liberal democracy" because it's ambiguous between norms that were liberal in the 19th century but really are conditions of actually being a democracy at all—not persecuting opposition politicians, sufficient freedom of speech that they have a chance to convince voters—, and democracies whose policies are liberal in a modern sense.)
The first sovereign democracy of the modern era was the United States of America, in 1776(*). The second was the French Republic in 1789. In 1798, the United States and France went to war. A limited war, fortunately, and one both sides would rather forget, but still.
* The Icelandic althing goes back to the 10th century, but was under the sovereignty of the King of Norway and/or Denmark until 1903; there were a few other proto-democracies under similar circumstances.
There is a stable solution though when two separate backscratchers clubs get established and are comparable in strength. This is how American political parties have worked for most of their history, though in recent decades they started taking on more substantive ideological differences, so they stop being as pure backscratchers clubs. (Van Buren and Jackson had almost no substantive policy interests in common - they just realized their orthogonal interests could both be achieved if they coordinated together.)
Agreed. I think this is why countries seem to naturally slide towards being Lebannon, Brazil, Argentina, Venezuela, Malaysia, and so on -- naturally slide towards being split between different backscratchers clubs. What social scientists call "Clientelism", "Patronage Politics", or just "Corruption". Why is corruption so endemic, so hard to fight? Because you can't have a Backscratchers' Club that works for everyone, at least not in the long term. In the long term, everything slides towards splitting up into multiple Backscratchers' Clubs, each trying to benefit themselves by stomping on the others, in a long-lasting stable equilibrium. Sometimes this looks like the historical US. Mostly, it looks like Lebannon or Malaysia.
(The process of *becoming* Lebannon or Malaysia looks something like the recent FAA hiring scandal [https://www.tracingwoodgrains.com/p/the-faas-hiring-scandal-a-quick-overview], where you invent a test with arbitrary answers, hand out the answers to your Backscratchers' Club, and act shocked when only your club aces the "Your lowest grade in high school should be Science, but your lowest grade in college should be History, and your most common grade in high school can be anything but a C, which is *verboten* for some reason." question set. Damn, they must really know their stuff huh?
A similar thing was key to Erdoğan taking power in Turkey [https://www.astralcodexten.com/p/book-review-the-new-sultan], incidentally; once Gulenists got into power at the Ministry of Education, they could steal or just write the answer sheet to the national examinations, then pass the answer sheet to their own students so the civil service would be increasingly made up of top-scoring Gulenists.)
(You could also bring in some Selectorate Theory [https://en.wikipedia.org/wiki/Selectorate_theory] and note that this helps explain why it was so hard for the US to build democracy in Iraq & Afghanistan. A democracy is, in Selectorate Theory, just a system where the "winning coalition" is really, really big. So big they can't be satisfied by just handing them money, it's more cost-efficient to actually do a good job of running the country & invest the money into things like public works, universal education, and other public goods that benefit everyone and exclude no one.
If the "winning coalition" is small instead though, none of that happens because it's more cost-efficient to just hand your supporters money directly, rather than incidentally benefitting them by building a new road system that works for everyone or whatever. And, as we have just learned, a Backscratchers' Club that works for everyone works for no one. That means existing "We work for everyone!" Backscratcher's Clubs tend to get supplanted from within by new "We work for *us*." clubs -- but it also means a new "We work for everyone!" club is hard to get off the ground, it can't compete with the established "We work for *us*." clubs. Even when the new club is being aggressively promoted by the US and called "Democracy". The dynamics are no different.
i.e. Democracy is not just hard to maintain, but hard to establish. It just doesn't have to beat Clientelism/Corruption when it's old & decaying, but new & trying to get off the ground. The temptation of "But what if we only benefitted ourselves?" is always there, waiting.)
(If you want to bring in more social science, it could be worth looking at Public Choice Theory & the example of things like farmer's subsidies. The famed paradoxical result that the farmers *gain* political power as they *lose* membership in a democracy -- contrary to what you'd expect where the big groups have power & the little groups don't -- precisely *because* they're small. Precisely because if there are only a few farmers, and they take a lot of money per person, that's still only a small amount of money lost per taxpayer, so no taxpayer opposes them while 100% of farmers support it -- so the farmers actually have more power when they're small. When they naturally form a Heather-like Backscratchers' Club rather than a Daniel-like club. In countries like India where the farmers make up a majority, by & large the farmers are actually taxed, not subsidized, simply because they're on the losing end of the Backscratchers' Club wars. Simply because they're large, and a club for them would be a club for everyone.)
Tangential thought: how do you keep together a backscratchers' club that includes everyone? Simple: *Don't tell them that it includes everyone*. You tell them that it excludes someone -- no, a lot of people, the more people the better. The more people you point to as being excluded, the more exclusive & desirable the club becomes, so you better find or invent a lot of people to exclude. If people can only be united in hate, well then invent someone to hate! If people can only be united by an enemy, well then invent an enemy! If the club isn't a persecuted minority, well then just pretend that it is!
I think a lot of historical examples of the persecution of minorities, ranging from ancient pogroms to historical Fascism, are an example of doing that but for real (instead of inventing people to exclude). Even things like, say, picking fights with other countries just to stir up nationalism in your own popuation & unite them together, could be an example. And no doubt there are plenty of modern examples, both of picking fights with a real someone & inventing someone to be scared of & exclude. I'm sure you can think of a few. (Perhaps this is why it was necessary to invent the concept of Satan, for example, to go historical again...)
i.e. So perhaps what the 90s dreams of unity & the UN really needed, was to pretend they *weren't* tolerant & inclusive. No, maybe it needed to pretend it was still excluding the aliens of Tau Ceti or something, like the Star Trek dream but inverted. If people cannot accept a genuinely inclusive world, simply tell them it isn't. Indeed, the more inclusive it becomes, the more exclusive & discriminatory you'll have to make it look (lest it *be*).
Another tangential thought: it's common for these self-promoting "Logrolling"/Backscratcher's Clubs to be made up of mediocrities (who would never rise very far on their own), or have hazing rituals where you submit blackmail material to the Club, or both -- because a club made up of successful people is good, but a club made up of successful people *who would be nothing without the club* is even better. The Varangian Guard/Machiavelli point: you want people who are loyal, because they have nowhere else to go. Because they would lose *everything* if you lost power -- and they know it.
Consider for example all those dictatorships that go around promoting incompetent sycophants who pose no threat to the dictator, and firing/executing anyone who proves themselves competent. Because competence means you could be something without the dictator; it means you have *options*. You could be a threat. The more incompetent you are meanwhile, the more completely & utterly dependent you are on the dictator for your exalted position. Again, Machiavelli in action.
Or look at the various Royal Guards, Presidential Guards, Republican Guards, and so on those dictatorships cultivate. Like the Varangian Guard, the most important thing is to not build the Praetorian Guard; you want people who are completely & utterly dependent on you, who are most emphatically *not* able to go outside you & plan your demise like the Praetorians did, who would be nothing without you. The more hated & foreign the minority, the more sense it makes to exalt them as your guards; it's just business really.
As Scott quoted Krugman as putting it at https://slatestarcodex.com/2015/07/28/non-dual-awareness/ ,“Being incompetent and displaying it, conveys the message I will not run away, for I have no strong legs to run anywhere else...". It's why, for example, people suggested making an all women army for the Afghan National Army, and generally putting women in charge in the Afghan government -- you want people who would lose *everything* if the Taliban took over. People who will be loyal no matter what, because the alternative is death -- a slow & painful death at Taliban hands. It'd be a complete & utter inversion of Afghanistan's native norms & traditions? It'd piss a lot of people off? Exactly the point. Exactly why it'd work. At least you wouldn't have Ashraf Ghani fleeing the country with over a hundred million in cash -- enough to set up a cushy life for himself after the fall of the country.
Agreed. Just imagine if he co-wrote with Robin Hanson. After the amazing allegory would be some counterintuitive solution for backscratching, possibly involving prediction markets.
The initial lifeboat involves a negative consequence: being killed/killing someone else. How would it work with a positive consequence?
10 people meet up because they all heard about a bird-watcher meet-up at a certain location. While watching for birds, they all simultaneously see an envelope fly out of the sky. Inside is a lottery ticket worth $5 million (must be a scratch-off). Only one person can redeem the ticket, and of course that person will also be responsible for income taxes and such. Who gets the ticket?
Someone initially screaming their name before lots can be drawn won't work, and maybe would get their name removed from drawing lots.
Maybe they would decide to split the thing 10 ways, after all distributions are made. That isn't straightforward, either, not even counting that the person selected may be dishonest and have a "change of heart" when given possession of the proceeds.
Maybe the person with the smallest income should get it, since that would result in the smallest of income taxes, maximizing the amount to be divided.
Hand in hand, everyone marches down to the nearest lawyer's office to draw up a contract for the fair division of proceeds.
The problem in this thought experiment is that division of the windfall is legitimate, so it can occur in a low-context matter that exploits the existing framework of law and contract. Both the cannibalism and back-scratching clubs are informal or clandestine arrangements.
Perhaps I could change it, then, to a family of 10 siblings, whose only remaining parent has passed away, leaving a priceless artifact which they all want to possess, not sell, as it represents the family legacy in some irreplaceable way. Only one person can possess it.
Nothing is more destructive than inheritance battles, between family members, over wills and it's even worse where there are no wills. "Sally took Granny's rings that I wanted and Granny promised them to me but Sally says she is the eldest daughter so they should come to her". Absolutely tearing each other apart over small sums of money and rubbish items.
I've heard the claim that it is not that rare for more money to be spent in total legal fees in these disputes than the actual inheritance. Three people each willing to spend up to $X-1 to win the $X inheritance, and sometimes willing to spend $X+T because it's worth another $T to them to make sure those other bastards don't get a dime.
The spending ratchets up not just due to spite ('make sure those other bastards don't get a dime') but also from sunk cost issues, like in a dollar auction (https://en.wikipedia.org/wiki/Dollar_auction). Once you've spent the money on a lawyer the money is spent and gone, so considering the next legal spending should just be about how useful it is on the margin - it makes sense to spend an additional $5k to get $10k of inheritance no matter how much you've already spent.
There is an auction game of, say, auctioning off a $20 bill, where only the top bidder wins, but the top two bidders both have to pay their bid. The idea is that, even if the top bid is $20 and you have bid $15, you will lose less money if you now bid $21. And then the same principle applies to the now 2nd highest bidder.
The only winning move is not to play. Or, perhaps, to make a separate agreement with the 2nd highest bidder, to both split the proceeds.
Dickens satirised it in the case of Jarndyce versus Jarndyce where an inheritance dispute had been going on for generations and eventually only stopped when the entire estate had been eaten up by legal fees so there was nothing left to inherit, but unhappily it went on just long enough to destroy many lives.
I've personally seen incredible bitterness over "so-and-so got that and it should have been mine" in disputes, and the thing in question was not worth wrecking family relationships over. People go crazy when a will is involved, even if there isn't a hefty sum of money or property to inherit. I have to agree that the spirit of "I'm not going to let those bastards win" takes over and quashes any good sense.
Whoever sees the coolest bird first, obviously, else this bird-watching meetup isn't worth their binoculars. Cue arguing over which is the coolest bird.
It's essential to the lifeboat version that the outcome is bad for one person and (relatively) good for the other 9, not the other way around. A version with positive consequences would have to have a positive consequence for all but one person, not just once. Let's say there are 9 remaining spots for a last amusement ride before the park closes. (Except the reward there is too low to engage in unfair scheming and jeopardize friendships.)
That version is still a negative version, just a less-than-what-you-would-get-otherwise version. Assuming all strangers, all of them expect to ride once more, but one of them will have expectations dashed. And in that situation, I think the race belongs to the swift, like in musical chairs.
But what about 10 people with but a single seat on the ride remaining? I think each person would try to paint themselves as most deserving: I came from the furthest away to ride, or I've wanted to ride this for 26 years, or it's my dying wish to ride this one last time, etc.
If by "positive consequence" we mean it's (relatively) good for one and (compared to that) bad for the others, that works completely differently from Scott's version, as everyone has only one preferred outcome, while in the original version one has multiple preferred outcomes, with no preference between them. In your version there is no motivation for everyone to agree on any Schelling point that's not him, so tricks like what Albert tried won't work.
It's the fraction of people who can get the better outcome that matters, not whether you you consider getting an extra ride in to be a positive or negative (how you view it is arbitrary and relative to your expectation). The equivalent of your example with the lottery ticket is would be if the lifeboat could only support one person, and they have to choose 9 people to die instead of just one. Then you can no longer get majority support (from rational self-interested actors) just by choosing a name. If the lifeboat can hold at least 6 of 10 people, you can form an ad-hoc coalition and overpower the other 4 people.
There's a British game show called "Golden Balls", where the final round has the last two contestants doing a Prisoner's Dilemma over the accumulated pot from previous rounds. After discussing, rach privately selects whether they try to "split" or "steal" the pot. If both split, the pot is divided 50/50. If both steal, both go home empty-handed. And if one splits while the other steals, the stealer keeps the whole pot.
Usually, the contestants spend the discussion phase swearing up and down that they'll split and urging the other to do the same, and then about half the time they turn out to be lying. But there's one notorious episode where one of the contestants opened the discussion by promising he would steal, but if the other contestant chose "split", then he would pay the other contestant half the pot afterwards.
Short version: contestant 2 threatened to steal if contestant 1 didn't promise to back down, but backed down himself and put in "split" at the last minute. Turns out contestant 1, despite his claims, also put in "split".
That was fascinating, and I would never have thought about it. But in hindsight, it makes sense, from a Monty Haul standpoint. In Monty Haul, if you switch doors, you KNOW the door they revealed isn't the door you want, so you're switching from a 1/3 chance to 1/2 chance to get the right door.
For this, his opponent has two possibilities for him to think: he will take split, or take steal. But he changed it to more possibilities: 1) he says he will take steal but promises to split the money later, 2) he says he will take steal but will NOT split the money later (he's lying), he will take split even though he says he will take steal. For his opponent, then, two of the three possibilities lead to a split, and one leads to him getting all the money.
The odds aren't as clear-cut as in Monty Haul, as one must still decide whether the opponent is honest, but it kind of looks like a 2/3 chance of winning this way.
This reminds me of a deal offered in a webcomic, where one side had two enemy sides who were both after its treasury. They signed an unbreakable deal with both sides, paying them 2/3 of the treasury, contingent on the OTHER side attacking first. This prevented BOTH sides attacking, until they realized they could sign another unbreakable deal with each other, splitting any treasury proceeds 50/50 no matter who attacked first. In this Golden Balls scenario, of course, after-game deals weren't necessarily binding.
This writing is amongst some of your best work, Scott. (I realize this comment doesn't contribute to the overall discussion, but I'm risking being spammy with a comment just to emphasize how great it is.)
Except...
I just went to share it on Facebook and then hesitated when FB previewed the terrible AI art accompanying the post. Seriously, it's so bad, I'm actually embarrassed to be recommending something associated with it.
I settled for putting a disclaimer on my recommendation of the link to ignore the AI art, but...man. No art would be so much better than the AI pictures, seriously.
So for the last year or so I've been describing the pervasive sense of plasticy homogenized sameness in Scott's AI images and everywhere else as "Lisa Frank." It doesn't matter what the subject of the art is, it always looks like it came out of the Lisa Frank Studio to be emblazoned on an 8 year old girl's trapper keeper in 1989. Scott's AI art for this article looks like the Lisa Frank Studio spun off a goth division or something.
Is it possible all generative AI started with Lisa Frank? That seems like the only explanation.
So you can just imagine my reaction when I clicked through. It might be the first AI image I've ever unironically enjoyed.
The unironic answer to your at-least-half-joking answer is that generative image AIs work by taking an image classification AI and running it backwards so it generates the image that it would most strongly map to the prompt you gave it. Google's 2015 DeepDream demo was the first high-profile implementation, which was pretty bare-bones and worked by perturbing a starting image to make it classify more strongly to the prompt.
As it happens, this makes deeply surreal images that look like Lisa Frank took too much LSD. I'm guessing the "Lisa Frank" part is because her art style is surreal in color but with very bold and striking subjects in the images. And the LSD part because the process of running a classifier backwards is fairly closely analogous to how drug-induced hallucinations in humans seem to work. I wouldn't be surprised if Frank's art style was influenced by 70s-era psychedelic art, since she came of age in the 70s and started doing her iconic art right around then.
The current crop of generative AIs have a bunch of extra stuff added in designed to constrain the output to be less psychedelic and more suitable to substitute for stock photos. Plasticy sameness is what you get when you run the constraints too strongly and Lisa Frank knockoffs are what you get with relatively loose constraints.
Is the objection the quality of the art or that it was made by AI? I'll admit to being fairly tasteless as far as art goes, but quality-wise it seems fine. Not a great work of art, but serviceable.
As for being AI made... I don't really get the objection here. I know people are worried about AI art displacing human artists, but I don't think there's any world in which Scott would have paid a human artist to commission an image to go with this blogpost: this is the sort of thing where AI has enabled art in a place there probably just wouldn't have been (properly paid and compensated) art before.
It doesn't seem worthy of a disclaimer... or is posting "disclaimers" about AI art a qualification of being in the "anti-AI-art backscratchers" club?
100% objecting to the quality of the art. See my comment above to @Eremolalos.
As for the disclaimer, I was actually switching horses midstream there. I took a break writing the comment to post the link to Facebook and was then confronted with goth Lisa Frank.
For the record, I'm totally fine with AI taking artists' jobs (if it does it well). But realistically, there are ways in which paid-for human art would likely be here if it weren't for AI art. Scott (or Substack) might pay for a stock image subscription service that would then go pay individual artists for their work.
Why are people so split on AI art? What is it that you don't like about it?
It has that common AI style: digital painting style, dramatic lighting, dark ArtStation type of feel. If you zoom in, the details are often nonsensical. But overall, it's a nice looking image.
It looks nothing like Lisa Frank to me. It's nearly monochromatic apart from the life jackets. Dark and moody.
I would describe it as the type of images you would get if you did an image search for "striking digital art" or "stunning digital art", back before AI art existed. It always tries to turn up that "wow" factor to an 11.
Stable Diffusion does other styles no problem if you want it to. I don't think it's inherent to AI (unlike the nonsensical details, which all AIs struggle with). Midjourney seems to have purposely tuned their model to have more of an ArtStation vibe by default. I'm pretty sure this style is driven by human preferences. It's just strange how divisive it is. People love it or hate it.
In order for a party or movement to get to power it has to have a great coalition of backscratchers, but once you take power you want to have a narrow coalition so each person can get a larger slice of the pie. The solution is to have a great purge.
They're similar, but the premise is different. Ponzi schemes are actually pyramid schemes, but the con men tell people they make money by some investment strategy, but any returns actually end up coming from new investors.
Well, I know I'm being pedantic here, but Ponzi schemes are not pyramid schemes. They're similar, but the former is not a subset of the latter. It's more like Ponzis and pyramids are subsets of the set "frauds where any returns are coming from new investors."
In a Ponzi scheme, the schemer acts as the hub of the scheme; all new investors interact directly with the schemer and make their payments to him or his employees, even if they were recruited by earlier investors*. The investors invest with the schemer in the first place because, as you say, they've been told a lie about the yield coming from postal reply coupon arbitrage or complicated options straddle trades or something. There's no pyramid; investors don't have to go out and find new investors.
Anyway, the reason I brought it up is precisely because in a pyramid scheme, the participants are perfectly well aware that the yield is coming from new investors, and they naturally want to be as close to the top of the pyramid as possible.
That is an extremely important book, and though it says that, it says a lot more as well. Esp re why some types of govts provide public goods and some don't.
It is not true that Daniel's backscratcher club is meaningless. It has the effect that the local newspapers no longer print anything negative about anyone local, which presumably was not previously the case. (If everyone including non-locals joins, then the newspapers only print puff pieces.)
Thinking about this a bit more, it seems like some people benefit more than others from Daniel's club. For example, people who are better at parsing subtle, implicit social signals might do better, relative to others anyway, in an environment where most explicit negative communication has been banned.
You could tell a story that this is how we get social norms around politeness - they are imposed as a backscratcher's club designed by people who are good at communicating with subtlety. Not sure I believe it though
This also explains why it can be economically rational to remain a racist in a racist society even though classical economic theory says it shouldn't be. The cost of being excluded from the white Backscratcher's Club following this kind of defection can be higher than the gains from open and rational business activity.
Perfect competition doesn't predict that profits get competed down to zero anyway (in the accounting profit sense, where the cost of capital, that of the investors deferring consumption and taking risk, is included in the profit, as opposed to the economic profit sense).
Economic theory also doesn't actually predict that no company will discriminate, or that all companies that discriminate go bankrupt. What it does predict is that, if the market is competitive, and race doesn't provide information about the quality of workers beyond what can be easily discerned from other factors, then equally good workers of different races will have approximately equally good job opportunities. If, say, 13% of the population is black, and half the employers and/or workers are irrationally racist against them, and the rest of the population doesn't care about race, then on a perfect market, half the companies will employ whites only, and at the other half 26% of the employees will be black on average, and everyone will make a similar salary at similar jobs. On a perfect market, black people only start to suffer if, in some profession, a bigger percentage of employers (weighted by the number of positions) are racist than the percentage of non-black workers in that profession. Of course on an imperfect market, problems start sooner because of the non-perfect fungibility of employers and workers.
This does predict that if a generation doesn't experience discrimination as young people, then equally performing members will have equal shots at achieving high positions once they're old and experienced. And perfect competition does predict that young people won't experience discrimination (again assuming race doesn't provide information about the quality of workers beyond what can be easily discerned from other factors).
I wouldn't say the perfect competition model assumes worker quality is solely a fixed inherent quality like IQ; it doesn't make assumptions like that. I'd rather say that when we say the perfect competition model predicts that workers of equal quality will have equal opportunities, quality is to be understood as current quality, however that came about, even if it's affected by earlier discrimination. But if a generation participates on a labor market with perfect competition throughout their lives, then worker quality in this sense won't be affected by discrimination.
David Bernstein made the best argument I'm aware of for the 1964 CRA, based on the idea that uncoordinated defection from the equilibrium would result in extra-legal violence :
Cato Unbound hosts debates between people, and usually those people are not employees of Cato. Bernstein is a libertarian, and he's arguing against other libertarians who criticize the CRA.
This is a good place to discuss if Backscratching Clubs even make sense as a way to benefit their members in the first place.
The example that's most relevant here was: "If a member runs a company, they should preferentially hire other members for good positions". But if you own a company, and you hire based on club membership rather than merit and salary demand, the club member you hire benefits, but you lose as much as he gains, possibly more! (Proof: if you didn't hire based on club membership, maybe he could offer to do the job for sufficiently less money that you'd still hire him. If he'd still be willing to do the job for that money, he gains as much as you lose by paying him more. If he could get a better job at that point, he gains less than you lose. And if you wouldn't employ him even if he'd do it for free, as a non-member is so much more competent, that means you lose more than the entire salary you pay him through your nepotism.) The club, as whole, doesn't benefit.
It's another matter if you're a hiring manager of the company, but not the owner, so you benefit a fellow member, but you aren't paying with your own money, and the shareholders aren't members. So in economic matters it seems to me that there is a point in a club only if either
- it's a small club, so they have e.g. a management position here and there, but they don't also make up most of the shareholders. I.e. they have an agent, but aren't the principal. It works better if membership or their behavior isn't well-known, otherwise people don't put them in positions of trust.
- or the situation is asymmetric, members tend to be on one side of a trade (e.g. they make and sell some product), and the other side of the trade are non-members, the members can collude to (say) raise the price of their product, and non-members can't easily enter the market (perhaps the members lobby to have them barred).
----
About racial Backscratchers' Clubs: Assume almost all managers and investors are white, while whites are a smaller majority of workers. A club where white-led companies are expected to prefer white employees is in the interest of white workers, but against the interest of white businessmen (managers and owners). Once the club contains all whites, it may be risky for a white business to defect from it alone; but rich whites have no reason to join in the first place, and they have an interest to coordinate to leave the club, at which point white workers have no choice but work for non-members. And even if one or a few white business defect on their own, white workers have little individual incentive to shun them, as workers or customers, even if they have a collective interest in maintaining the club. The club would only be stable if the workers are somehow able to coordinate, but the businesses aren't, which is unlikely. In a free market that is: white workers have an interest in voting for governments that enforce segregation, what I'm saying is that it's sufficient to stop enforcing it.
Regarding the second paragraph, this doesn't explain why white managers or investors would want to keep whites in charge. People are mostly capable of scratching the backs (beyond what's explicitly expected of them) of those below them, but not of those above them. So if there's a white backscratchers' club, whites on the lower rungs have an incentive to want white bosses, and for the club to continue to exist (such that their white bosses scratch their backs), but they have little influence on that; while those at the top have no incentive to perpetuate the club by either scratching the backs of the whites directly below them, or by incentivizing those to scratch the backs of whites further below.
It likely won’t be a zero-sum game, though - you can use such corruption to extract from outsiders. If you’re a manager but not an owner, it’s not your loss if the company doesn’t benefit. If you’re a politician or public servant, you can squander public funds. If you’re a cop, it’s not your loss if you let other members go free and protect other cops in the society against accusations of misdeeds, Freemason-style (or just regular Blue Wall of Silence).
Yes, in some cases it can work. My point is, an entire race likely doesn't make a good club, as it's too broad: some significant economic or social classes within the race won't have an interest in participating.
>But if you own a company, and you hire based on club membership rather than merit and salary demand, the club member you hire benefits, but you lose as much as he gains, possibly more!
That's why dictators hiring incompetents isn't a stable equilibrium. If so me other country hires meritocratically, they can grow their pie -- increase their industrial.and.military capacity -- while you stagnate, and then they are a threat to you.
Indeed. Punishment of non-punishers can maintain even a bad equilibrium; if a non-racist business gets boycotted and people who defect from the boycott get boycotted themselves, then not being racist probably won't be worth it.
Isn’t there also the dual case for discrimination - that to hire an outsider is to reward an illegible but competing Backscratcher’s Club? This is the basis of many xenophobic claims. Additionally, one can argue that the WEIRD package of traits includes a fundamental disdain towards Backscratcher behavior, though perhaps it’s a more subtle disdain for using family/kin as a basis for coordination.
Sorry, you lost me at the start. I can definitely see some groups of people going with Albert's suggestion, but I believe a significant number of people would object to that and say "...no, Albert, that's weird, let's just draw the lots".
Perfectly rational people never are found in these situations, chiefly because such people don't exist. That is the problem with game theory, as opposed to social experiments. Such as the one where you are given $100, provided you must give some amount to someone else, explaining the whole situation, and both of you agree on the split. "Perfectly rational" people ought to be happy to get $1 even if you get $99. I'm clearly not perfectly rational.
Corollary: it is good to have a reputation for *not* being perfectly rational, but actually caring about silly things like fairness and self-respect, so that if someone gets to play this money-splitting game with you, they won’t try to stiff you because they will be afraid that you’ll decline anything less than an even split even if that means that the $100 goes into the paper shredder.
And the best way to get such a reputation is by regularly demonstrating that you won’t let people walk over you unpunished, even when that would be the best choice for you on a case-by-case basis. So being irrational in the short term can be the optimal strategy in the long term! This is just the standard tit-for-tat concept in game theory, of course.
I'm going to put out the corollary that, according to game theory, equity is not rational. Therefore, this is a flaw in game theory, since people don't actually behave that way.
Equity seems rational in iterated capitalism, with growth curves? When the money isn't just "points", but resources that can be put to use in an environment where we're all competing on multiple levels?
I admit that in a communist society, this problem goes away entirely, since the state takes everyone's winnings.
I meant equity in the equality sense: fairness. In game theory, you want the best outcome for yourself, or to achieve your goals, irrespective of what happens otherwise. "I wouldn't do that to my worst enemy" is bad game theory.
But game theory doesn’t stop at analyzing one-off encounters. There’s also iterated games, reputation effects, the concept of fairness (or equity if you prefer) as a natural Schelling point in negotiations, etc.
Lots of people have spent lots of effort into finding ways to show that a) fair and even seemingly altruistic behavior can emerge from "enlightened selfishness" by rational actors, and b) sufficiently sophisticated forms of game theory can actually be useful in explaining some things about how real-world people and animals behave.
"If you defect, you temporarily become my worst enemy, I will inflict punishment on you, and I am willing to pay a cost to inflict that punishment because seeing defectors punished has utility for me" is arguably good game theory.
> In game theory, you want the best outcome for yourself, or to achieve your goals, irrespective of what happens otherwise.
Sure, but, even in a non-religious sense where it's an abstraction of accumulated years of self-reinforcing behaviors, I have a soul. The best outcome for me-as-a-person is not always going to be to behave as ruthlessly as possible, because the act of making that choice will affect me-as-a-person. Biological neural nets change when they're used, unlike the current generation of LLMs.
I maintain that this is not perfectly rational, and that the only rational approach is for them to give me $99 for the privilege of getting to keep $1.
Or I suppose we could split it 50/50; it's not my first choice but eventually I'd reluctantly compromise on it, so I might as well skip to the conclusion.
In the standard version of this game, one player gets to offer a split and the other gets to accept or decline; if the second player declines then nobody gets anything. There’s no negotiation.
Who needs negotiation? Non serviam. They're lucky I can't make an acausal bargain from the future to torture copies of them in virtual hells. It's $1 or $0 for them, and if they haven't figured this out, no skin off my back.
My righteous satisfaction will keep me warm in the cold winter nights.
Seriously, though, no way I'm letting the type of person who'd propose 99-to-1 splits get a 99x resource advantage over me. That does not end well for me.
To be a bit more formal, I reject the framing of the interaction which starts after they act but before I act. And I maintain that it is precisely as rational for me to demand a 99-1 split as for them to offer me a 1-99 split, and that if they want to come out of the interaction with money in their pocket, they need to think for just 1 second* about what kind of compromise someone else might find acceptable. Maybe they recognize that some people feel like I do, and take a gamble and offer an unfair trade anyway, but in that case they lose.
* Or however long it takes them to divide 100 by 2. For me it doesn't seem to require active thought at all.
I was going to say that the obvious first victim should be whoever has the most meat on his bones, but then I saw that Heather works in Marketing, so Heather's going down.
If Heather is any good at all at her job, she should be able to show how much better all the other alternatives are. If not...I suppose that is evolution in action.
I was surprised this never came up. Let's eat Bob, he's the biggest person here. Bob needs to consume more calories than anyone else and contains the most calories in his body, so clearly we should eat him first.
Real world Backscratchers Clubs solved the "everybody joins" problem by not letting everyone join. Rotary Clubs, when they were founded, had a rule about allowing only one person from a given profession, because Alice was the only lawyer, so we all went to her for lawyer needs, and Bob was the only accountant, so we all went to him for accounting needs. And then they were also service clubs, which is now the main point, but at founding some younger professionals just decided the way to get ahead was to form a small cartel.
If you are looking for real terms, "cartel" might be a good one to pursue on this front.
OK, but if the Rotary club is all dudes, and the club prostitute that everybody goes to for their "needs" is a dude, that could be interesting...
More seriously, I would expect Alice to benefit from her husband's membership; if we allow for female lawyers in this otherwise-patriarchal hypothetical, then everybody knows that they're supposed to go to Bob's wife for their lawyer needs.
Wouldn't it make more sense to include everyone from one profession, collude to raise prices, then lobby the government to ban non-members from practicing the profession (i.e. a guild/cartel)?
I don't see how one benefits from a club with one member from each profession. Whatever you gain by having other members buy your services, you lose by having to use other members' services even when going to a non-member would be advantageous. Job security perhaps? This doesn't feel like the most efficient way to get it.
There is a (hypothetical) story by German sociologist Heinrich Popitz about passengers on a cruise ship. I found a summary in the article "On the difficulties of speaking out against security" by Thomas G. Kirsch, Anthropology Today (Volume 32, Issue 5, October 2016, Pages 5-7, https://rai.onlinelibrary.wiley.com/doi/10.1111/1467-8322.12295) Popitz's story is related to your story because it is also story about coordination and power dynamics, but it assigns a different role to ideology. Here is Kirsch's summary. (I have not read the rest of Kirsch's paper.)
"Imagine a cruise ship with a limited number of deckchairs. Imagine further that there are passengers on board who have a self-serving interest in their exclusive usage of them. An obvious strategy to achieve this interest is to occupy the deckchairs either in person or symbolically by placing towels on them. Yet since the latter can easily be removed, an additional and more efficient strategy consists in asking selected passengers to act as occasional custodians of the deckchairs, a role that is rewarded by temporarily granting them access to the deckchairs for themselves. Popitz thus argues that assuming power in this way relies on basic social mechanisms and does not require the existence of either an elaborate ideology or a vision of the future. Not so with those who have been excluded from using the deckchairs. When joining forces, the excluded opponents cannot content themselves
with simply taking possession of the deckchairs. Instead, when seeking to revolt against what they perceive to be an unjust deckchair system, they have to come up with a vision of how the allocation of deckchairs could in the future be organized in a different way. In short, they have to agree on what an alternative, more equitable and legitimate order might look like."
>I’m sure real sociologists have written about these issues, but it was hard to find them and I figured I might as well write this post without citations. Still, if you know who they are and what their terms are, let me know.
The way to find basic literature on a topic like this, or any topic, is to google "[topic] syllabus" and you will get, well, syllabi of courses on the topic.
Also, political science might be the place to look, rather than sociology, esp if one wants to find game theoretical approaches to these sorts of questions.
I thing Danning's suggestion is a good one, but you raise a good point. ChatGPT is good for these kinds of suggestions. I asked it "Is there any scientific literature on how clubs manage to self-perpetuate due to the costs and benefits of being in the club?" Collective action and clubs seems to be the answer. Maybe also "Group Cohesion" and "Conformity". Sounds like economics and psychology to me.
Yes, collective action is a good one, too. Though the fact that that search seems to turn up only Econ syllabi worries me. I know there is tons of literature in Poli Sci on collective action (unsurprising, given that that is much of what politics entails)
Well, immediately before the quote I included, Scott said:
>If the Lifeboat Games seemed suspiciously like nationalism, and the Backscratchers Clubs seemed suspiciously like clubs/cults/ideologies, the stories in this section seem suspiciously like the Establishment - whether it’s the Catholic Establishment of the Middle Ages, the conservative Establishment of mid-20th-century America, or the progressive Establishment of today. Elites support each other not directly - which would be hard to coordinate - but by all supporting the same ideology. If it’s hard for non-elites to break into the ideology, then everyone with the ideology will be elites, and supporting the ideology is an indirect way of elites supporting other elites in a big backscratching network. This is one of the solutions to Class Warfare Having A Free Rider Problem.
So, I took "these issues" to refer to nationalism, ideology, etc, so terms like that. But if the issue is game theory, the game theory, formal theory, or formal modeling would be terms I would recommend of the top of my head.
This "just so" story ignores actual human nature and society. Please see the story of William Bligh in a small boat, the story of the whalers who were the inspiration for Moby Dick, and the English and American lifeboat cases. Any court could not charge the survivors. Murder one for the instigators at a minimum. Life long guilt for the other survivors. Since the premise is not true in any society one might wish to live in, the rest of the discussion is not valid. This reminds me of Steve Sailor's discussion about the trolley car variations, just not as realistic.
He should read up on R v Dudley and Stephens: two lifeboat cannibals admitted what they did and were prosecuted when the reached shore, establishing the legal doctrine that necessity is not grounds for murder. Popular opinion disagreed and they were pardoned. The decision procedure was to eat the person who slipped into a hunger-induced coma, not draw lots or shout.
I'm thinking of that whaleship Essex and the guys who rowed away from it, into hell - and having the courage to yell out "So we're going to cannibalize each other because we're afraid of going due west and running across cannibals? Kill me now!"
And how interestingly tragic it was that there was a drawing of lots, and when it was the captain's nephew, he bravely submitted to his fate because honor culture compelled it; and it compelled his uncle to go on with it, too, because it would have been wrong to show nepotism. It's almost like, if it had been one of the others, maybe the men could have said - wait, this is nuts ... But then I guess his death really did save the rest, or all but a couple of them, I think.
I guess at that point no one had died, that they might have eaten. Or maybe someone had died, but not expecting this eventuality, they had pushed him overboard. I don't remember the grim details, except that physically the survivors bounced back pretty well, but I don't think their lives went all that great.
The principally interesting thing about these situations is the tenacity to cling to life. Of course hunger makes your brain work no so good. Still, when you are on the downward slope of life, in circumstances where that life doesn't seem likely to offer much else - the obvious solution is: let's all lie back and let death come nice and peaceful-like.
I guess that's just not possible for young men and men in their primes.
I don't think the Orphan Supporters one works as a way to solve coordination problems or benefit an Establishment. In the Elite, virtually everyone is an Orphan Supporter, just like everyone joined Daniel's Backscratcher Club, so it doesn't give a benefit. Anyone can become an Orphan Supporter, and it costs nothing as long as it mostly just requires saying the right words, so it doesn't protect the Establishment members from new competitors. You say it works if "it’s hard for non-elites to break into the ideology", but that isn't normally the case, especially for those who would otherwise have a shot at breaking into the elite. And it doesn't explain what incentive the Orphan Supporters have to punish someone who supports orphans as much as anyone, but doesn't do enough backscratching.
In politics proper, not everyone is an Orphan Supporter (as in the privacy of the ballot box people can support causes that are outside the norms), but that doesn't help the existing Establishment. At best, they share some seats in the legislature with the newcomer anti-Establishment Scrooges but manage to keep them out of actual power. At worst, sometimes they the Scrooges get the majority and they lose power. Meanwhile, if the Establishment didn't make support of orphans a requirement for membership, one half of the Establishment could take up the Scrooge cause, and the two halves of the Establishment together would dominate politics completely, as their existing parties and connections give them an advantage over newcomers.
Once the pro-Orphan norm is established, people will occasionally use it as a way to get rid of rivals via true or bogus accusations of Scroogery, and the possibility of that helps create an incentive for everyone to be an Orphan Supporter. But that (getting rid of a rival this way) is a rare event because most everyone is an Orphan Supporter, and the whole thing doesn't really benefit the Establishment. It does benefits the Orphan Support cause, though it may end up too silly to be effective at the margin.
The Communist Party in China is good irl example of large-scale 'backscratchers', often its necessary for career progression. However this might not quite be an internally stable system, people only seek the credential of 'CCP' for their own ends, and in a tragedy-of-the-commons I imagine people will only be favourable to other CCP members as much is necessary to keep up appearances or stop themselves being kicked out. It might depend on the level of intra-surveillance of the club.
Business associations sound a lot like instances of backscratcher clubs as well. The value of the association is all in the networking effects it offers. From what I understand, legal associations act to simplify lawyers' lives, by giving a clear network of referrals to point clients at as needed. Do we see similar incentives operating in "elite classes" as well? At least in B2B exchanges, the upsides of an exclusive referral network are high enough relative to costs, that joining multiple associations is fairly common.
Don't forget that not everything is zero-sum either. Having businessmen find each other when they have completing products/needs/plans is positive sum; business associations that facilitate such connections may well have a benefit even if everyone joins. (Idk how they work though.)
1. The Schelling point in the original Albert call to eat Bob is interpreted as a breaking of symmetry to favor Bob. Indeed, it was. However, the symmetry is also broken by Bob as the caller. It's up to the group to determine which break of symmetry is going to motivate it's behavior.
2. A lack of utility in a backscratcher or orphan club comes if participation is universal. Accordingly, a club who's premise relies upon <100% participation can overcome. "This club's bylaws say that when participation exceed 50% +1, each month in such a state we shall banish the member who has tithed to us the least (or virtue signalled or...)."
3. Such games are of greatest interest when the cost of backscratching is assumed small. In a capitalist's world, the assumption is that favoring purchase of A over B for such extraneous factors leads to persistent nontrivial disutility in each transaction, and such backscratching clubs die for sub-competitive performance. (Leaving a capitalist to expect such clubs to thrive in only non-competitive markets. I leave it to the reader to conclude whether backscratching clubs appear most prevalent in such non-competitive markets today.)
3 is more "free market" than "capitalist", right? I'm sure someone could do ROI calculations for membership. But it seems like such calculations would erode the social fiction about orphans?
I once had similar thoughts (though not as clear and precise as you laid it out here) which led me to my theory of the left and right:
Left and right differ in their preferred method of coalition building: the right prefers to build a coalition with people they know – neighbors, family, friends – though doing it through religion and nationalism also works. While the left does it through abstract ideologies, and does it in a more indirect way – trying to hide what's going on. Both hate how the other group is doing it.
My group builds coalitions along abstract ideologies in a good way, and your group builds coalitions along abstract ideologies in a bad way. Can't you see the difference?
I don't think this is fair - it's just that ideology isn't the important difference when considering left/right organization. Right and left are labels for the structure of the organization itself (in the context of capturing a democratic state). Left is bottom-up, swarm governance, by consensus. Right is top-down, monarchical governance, by command. These are platonic forms ofc; real-world left/right groups always do some level of both. Ideology is important only so much as it helps or justifies the organizational particulars.
My point, which I thought was also your point, is that religion and nationalism are "abstract ideologies" in the way the other commenter probably defines that term, and that that commenter just didn't seem to want to admit that the right "does it through abstract ideologies" too.
I think we more-or-less agree. I just didn't see OP's invocation of "religion and nationalism" as particularly negative or mean-spirited (tho it could have been, idk)
I didn't either. If anything, I took the commenter's invocation of "abstract ideologies" to be negative, which is why I thought it amusing that they casually mentioned religion and nationalism as axes of coalition-building for the right, as though those were not also "abstract ideologies."
I think that this (or something like it) is something people intuitively grasp and triggers opposition to anything that starts to look like it. Obvious examples are hostility to effective altruism, modern anti-semitism and all the parts of the culture war which consist of each side thinking the other is a giant conspiracy that wants to hunt them down at some point.
Having said that, I'm not sure that these kind of not-quite-a-conspiracy backscratcher movements are really that common, or even occur at all. There are vanishingly few people who'd be engaging in them consciously compared to the numbers involved for a progressive/conservative establishment. If you're not consciously manipulative, aren't you the equivalent of someone who's just actually trying to help orphans and thinks they're struggling against orphan-hating moral mutants on the other side?
Looking for historical groups that took over society like this:
2nd-4th Century Christianity? Maybe, at the tipping point, but hard to believe it could work without the bulk of the support being genuine belief
The reformation might be more plausible given there's a movement among the princes to strengthen their grip and weaken the emperor, but that's more of a class interest than a group within a group. They also didn't want to leave half the population around them Catholic. They were shooting for total hegemony, much like the Catholics had been earlier.
The Meiji Oligarchy might be a better bet, but it's a very quick cascade towards support for the Emperor, and looks more like coalescing around a winner (eg. after a coup) which isn't the same phenomenon.
Political parties in general might be a better bet (coalition of hopefully 50+% of the population to benefit themselves at the expense of the smaller nearly half); I can't think of any country other than pre-war Hungary where a permanent coalition to shaft a chunk of the population has happened though.
Ignore my previous examples, they were all a bit off-base as these kind of movements should be much, much smaller to work. The best example is probably the Broderbond in South Africa, which I think fits this pattern.
Short version: in 1867 Hungary got fairly broad autonomy from Austria (a bit like a contemporary British dominion). Power ended up consolidated in the Liberal Party, which used a mixture of gerrymandering and patronage to control parliament in spite of being broadly unpopular among Magyars with the support of minority groups and people they were directly subsidising/employing.
> I can't think of any country other than pre-war Hungary where a permanent coalition to shaft a chunk of the population has happened though.
The Indian caste system? Racism in the United States pre-Civil Rights Act? Antisemitism in Christian-dominated Europe? The treatment of Cagots in France?
Other than [possibly] the Cagots, they're all building on pre-existing groups as opposed to forming groups for the purpose of discriminating. It's the formation that's the weird part.
"Consider for example racism. There are supposed justifications for racism - like that such and such a race is inferior, or oppressive, or plotting to kill us. But another justification is just “We’re the majority and they’re the minority, and if we all band together to profit at their expense, it probably goes well for us.” Any coalition of 51%+ can do this. But it’s easier if everybody comes color-coded so there’s one obvious coalition that occurs to everybody and which they can easily check that they’re a part of"
There is a lot of real-world scholarship on this question, and I only know some of it (and even that is nearly two decades out of date at this point), so if anyone knows more or more recent work please correct me, but from what I've read historians who study the origins of racism make it sort of halfway between these, or maybe a bit skew to either.
The argument is that racism is largely economically motivated. One pattern goes like this: it's really economically powerful to kidnap people and force them to work for you under pain of death and torture, so people want to do it, but it's too obviously evil, so you start think "well they deserve it" and "it's better for them anyway"—not only or even mainly as conscious ways to get other people to sign off on it, but as ways to justify it to yourself.
Another stage is the maintenance/revival of racism. Take the post-civil war south: racism there served the interests of the wealthy because it helped stop intra-working class alliances which did things like support unions. (There were some of those in the decades immediately following the war.) Why did working-class whites go along with this? One common answer is that racism provided what W. E. B. DuBois famously called a "psychological wage": they got to feel superior to other people. Presumably this is easier to achieve than, say, actually forming a union & bargaining & achieving a more just society, so there's temptation to believe it. And there's at least *some* economic benefits to poor whites, too, since they formed a sort of upper working class (not a standard term afaik but compare "upper middle class"), which made them better off.
So is this “We’re the majority and they’re the minority, and if we all band together to profit at their expense, it probably goes well for us.”? It's not entirely different. But it's closer to, "we're already profiting at their expense, and only bad people would do that unless they're inferior, and we're not bad people, therefore they are inferior"; or, in the second example, "if we richer people get group A to hate Group B they won't notice us picking their pockets. ("Only a pawn in their game...".)
But of course the above is simple and real people are complicated, so I think in real history it's an incredibly complex dance of these factors: there are already some ideas about inferiority, and you want to believe you're not a bad person, so they sound more persuasive to you, you talk them up more, other people in a similar situation hear them, etc. Or, you, a rich person in the south, hear positive and negative information about a lot of people, and you are subtly (possibly subconsciously) inclined to notice the negative information about black people more, and then you pass it on to others including your employees, etc. As James Madison famously put it, "As long as the connection subsists between [man's] reason and his self-love, his opinions and his passions will have a reciprocal influence on each other". Which is to say, I think the mechanisms here aren't as simple as the historical sketches above (and which can get downright silly if you interpret them literally rather than as schemas), but the economic motives provide a subtle pressure to notice, spread, and even develop the ideas which are in some ways useful to you (which is like what Soctt is saying), but also that psychological usefulness is at work too (which is another element).
White workers may well have benefitted, if employers had to hire them in preference to blacks in some cases. Thomas Sowell said white workers in South Africa pushed for apartheid to protect them from black competition. But in this case it's employers whose interests were against racist policy: segregation essentially functioned as a white workers' union/cartel; a union works best for the interests of its members if it excludes some people who could do the job, and can have them prevented from competing with it.
Meanwhile, if employers could hire black people freely, even if separately, then workers (both white and black) might have made somewhat less money than if they worked together, but white and black people would have got a similar salary for equivalent work.
If black people made much less for equivalent work, that points to "my" version (the first paragraph above); if they got a similar pay for similar quality work in a similar position, that points to yours.
There is a clear downside to joining Daniel's backscratchers club: you have to favor other backscratchers over better qualified alternatives.
So while the club is getting started, a startup founder is a member and is trying to hire a software developer. A homeless person comes in and asks for the job.
"Do you have programming experience?"
"No. But I'm a fellow backscratcher, so you have to hire me."
"Fine. I quit the backscratchers."
The founder's fiance is then approached by their ex.
Well yes, but then the founder is ostracized in a "2 minutes of hate" that is calculated to be enough of an overreaction to deter anyone else from resigning for the next 18 months. All backscratchers participate, especially any who worry that their loyalty might be questionable.
And the less said about what happens to the fiance, the better. Eventually he makes a tearful confession, and admits to subversive anti-backscratcher behavior, and submits himself to a re-education camp for recovering apostates. His physical health never fully recovers, and he occasionally breaks down sobbing for no apparent reason, but he insists that he's *happy* now.
Except that Daniel's backscratchers club does not have a 2 minutes of hate as a founding rule and if it did, it would have to be big enough that the 2 minutes of hate would be meaningful and difficult to escape.
Well, this is why real mutual aid societies vet their members (or did). These days they'll take most anyone who isn't obviously a basket case, but at one time you needed to actually be at least somewhat successful and reputable...
The Shelling point in the array of integers was displayed as a hyperlink in my browser. I clicked on it and my iPhone asked if I wanted to call the number. 938 is an area code in Huntsville Alabama.
This whole situation could have been avoided if, upon Albert yelling "let's kill and eat Bob," everyone else replied "OMG no, that's horribly unfair to Bob! Let's draw lots."
The golden rule, "don't do to others what you wouldn't want them to do unto you," has been a staple of human morality for centuries for a good reason. If I were on an ACX boat, I certainly would not want anyone to yell "let's kill and eat drosophilist!" and everyone else to go "yup, sure sounds like a swell idea!" Nor would I want people to coalesce around the Schelling point of "drosophilist is a woman, all the rest of us are men (not an unlikely scenario given the gender distribution of ACX readers), so let's kill and eat her!" I value my life. And because I wouldn't want others to do it to me, I ought not to do it to others!
I gotta say, the utter lack of compassion behind all these calculations of "let's see, if I draw lots I have a 10% chance of dying, but if I throw another person under the bus I have a 0% chance of dying (for the moment, until someone else has the bright idea of shouting MY name on the next round)" is chilling.
All this reminds me of something I read a while back:
Bob is sitting at home when there's a knock on the door. He opens the door and sees an ominous-looking stranger with a box with a lever on it. The stranger says, "If you pull this lever, I will give you $10,000, and someone will die."
"Who will die?" asks Bob.
"Someone far away," replies the stranger. "Nobody you know."
The stranger leaves the box with Bob, who agonizes for a while, but then rationalizes that people die every day, he doesn't owe some distant random person anything, and he's really not doing so well financially and $10,000 would really help him. So, finally, he pulls the lever.
There's another knock on the door. The stranger is back. He gives Bob $10,000 and asks for his box back. Bob hands him the box and asks, "What are you going to do now?"
The stranger replies, "I'm going to make the same offer to someone else. Someone far away. Nobody you know."
Presumably, he rewards Bob's ethics by not choosing him to be the person to be killed when someone else pulls the lever? I would like to think so. But I don't actually know.
The resolution to all of these paradoxes is that at no point in these stories do any of the characters demonstrate morality. They're all optimizing for outcomes like a rational creature that has no beliefs and is only self-interested. In real life people do have morality and do not only do self-interested things (except perhaps in states of extreme insecurity). Now you can cast that as another form of security: like after a point, one's safety is better increased by living in a fair and just and kind society than by getting any more personal material benefits. Maybe that's all morality is. But in any case none of your characters care about that at all, hence they are inhuman or at least very dehumanized.
Well, for starters, I shouldn't have to define morality, because people have been talking about it for at least 2000 years and it's a concept most people know well. So it is not an empty point: I'm answering the question by pointing at "a giant thing that you are completely familiar with" and saying "what about that?"
Second, it works for almost any definition you want. I am not trying to solve the problem of "explain this paradox in terms of a rational model of humans as optimizing agents". I am trying to solve the problem "explain why this paradox feels like a paradox when held up against my human intuition for how reality works". After all the point of all paradoxes is that they feel wrong, but it's hard to put into words why. So if you're Scott Alexander and you write this post and you think: "wow this is interesting why does no one talk about this?" it's because everyone else is thinking about moral humans, not this horrible and repuugant thing.
Specifically, he wrote "Something about this surprises me. It’s weird that there’s another solution which is more stable than the fair one of drawing lots. It’s strange that by shouting an obvious suggestion - one that adds no more information - Albert can save his own life with certainty. Still, that’s how it goes." What's surprising is that when you model humans *incorrectly* like this, the results seem wrong, but you can't easily tell how you modeled them incorrectly. The resolution is that actual humans do not work like this. Actual humans can perceive the game theory of the situation from the start and especially the unjustness of the solutions that fuck over one person arbitrarily, and they prefer a situation in which they are not at the whims of being fucked over by game theory, hence they have a moral intuition for the general rules that everyone should follow (and everyone should *signal* that they will follow) and they will strongly adhere to those rules---because they are *not* self-interested in this way.
>Well, for starters, I shouldn't have to define morality, because people have been talking about it for at least 2000 years and it's a concept most people know well
We don't agree on object level morality , which is why we still need politics. We don't even agree which kind of equality is the right kind.
Fine, but my point isn't that you need consider *my* morality because that's what resolves these paradoxes, it's that you need consider morality *at all* because it's an entirely missing dimension here.
What about it? Scott pointed out that his hypothetical situations merely demonstrate the incentive structures behind real-life repugnant behavior.
If you're going to claim that an allegory for discrimination is unrealistic because the people in it aren't behaving morally, you need to contend with the fact that non-hypothetical people have practiced discrimination for our entire existence, despite having every reason to behave morally. The people on the lifeboat are actually going to starve; in real life, we don't even need that much pressure to start getting ugly.
I guess you believe that's how people work. I, for the most part, do not, and I certainly don't want to take part in an intellectual exercise that doesn't content with the *idea* of people behaving morally, since that just reinforces the idea that people don't. Better to talk about how to be moral than how to be a perfectly cynical backstabber.
strangely the last paragraph of my reply is just not showing up? It was
(Although again I will add the disclaimer that in an actual stranded lifeboat situation, maybe they are psychopathic like that; I've never been there and I imagine stress can diminish all this moral stuff. Hopefully they're strong enough not to. In a way that's what strength. Regardless, in a *society* they are not. Although probably a few people are anyway, and that's why the signaling is so important: because everyone wants to know that they are going to be safe from everyone else.)
I think in real life a network of relationships would have sprung up among the people on the boat. There would be alliances and partial alliances and people or groups who disliked or distrusted other people or groups. There would be one or more people seen as potential leaders, or one person whom the majority saw as the leader. And it seems to me the leader would have a lot of influence over how the cannibalism possibility was handled,
If I were on the boat I would be drawn to someone fair-minded and kind who was seen as a potential leader. I wouldn't be drawn towards them because of my values, I'd be drawn because I'd see having them as leader as my best chance of having a decent outcome to the situation. A fair-minded and kind leader would probably stand up for using a lots system(after maybe first asking for a volunteer, because ya never know, there might be someone who would volunteer if given a chance). But it's also possible that the person seen as the most dangerous could end up as the leader, because people feared opposing them. Maybe Violent Guy or one of his allies would kill me in my sleep if I opposed him. Violent guy, if leader, would probably choose his strongest opponent as the first to be eaten.
Overall, it seems to me that the lifeboat, even as a toy example, doesn't do a good job of capturing how groups function. Most of our relationships with other people are valenced, so groups quickly develop a structure.
"Uruguayan Air Force Flight 571 was the chartered flight of a Fairchild FH-227D from Montevideo, Uruguay, to Santiago, Chile, that crashed in the Andes mountains on 13 October 1972. The accident and subsequent survival became known as the Andes flight disaster (Tragedia de los Andes, literally Tragedy of the Andes) and the Miracle of the Andes (Milagro de los Andes).
...Three crew members and nine passengers died immediately and several more died soon after due to the frigid temperatures and the severity of their injuries. ...Search and rescue aircraft overflew the crash site several times during the following days, but failed to see the white fuselage against the snow. Search efforts were called off after eight days of searching.
During the 72 days following the crash, the survivors suffered from extreme hardships, including sub-zero temperatures, exposure, starvation, and an avalanche, which led to the deaths of 13 more passengers. The remaining passengers resorted to eating the flesh of those who died in order to survive. Convinced they would die if they did not seek help, two survivors, Nando Parrado and Roberto Canessa, set out across the mountains on 12 December. Using only materials found in the aircraft wreck, they climbed 839 metres (2,753 ft) from the crash site up 30-to-60 degree slopes to a 4,503-metre (14,774 ft) ridge to the west of the summit of Mount Seler. From there they trekked 53.9 kilometres (33.5 mi) for 10 days into Chile before finding help. On 22 and 23 December, 1972, two and a half months after the crash, the remaining 14 survivors were rescued. Their survival made worldwide news."
"On October 13, 1972, Uruguayan Air Force Flight 571, chartered by an Uruguayan rugby football team and their supporters to take them to a game in Santiago, Chile, crashes into a glacier in the heart of the Andes mountains. Of the 45 passengers on board, 29 survive the initial crash, although more would die from injury, disease, and an avalanche over the following weeks. Trapped in one of the most inaccessible and hostile environments on the planet, the survivors are forced to resort to survival cannibalism of those who had already died in order to stay alive. However, rather than turn against each other, the survivors draw upon the cooperative teamwork they learned through rugby, and spiritual faith, in order to escape the mountains."
The reality seems to have been somewhat more complicated than that, but the survivors did establish a way of dealing with the necessity, and it probably did help that they didn't have to kill anyone in order to obtain corpses.
I think in practice most of the people would volunteer to sacrifice themselves because people are mostly good and moral, and they'd end up agreeing to draw lots. Maybe not, though, because as I understand it morality breaks down when people are super unsafe (cf what starving people do for food). But yeah I think the cynical self-interested response you describe is basically psychopathic. Personally, barring some extreme psychological change, I'd be drawn to a fair-minded and kind leader because of my values.
"If I were on the boat I would be drawn to someone fair-minded and kind who was seen as a potential leader."
Seriously!
ACX is a fascinating place and great for thought experiments, but the amount of amoral calculation focused on "how do I survive/improve my status, and everyone else can go pound sand" gets disturbing at times. (What makes it really paradoxical is that we know Scott is a good person! He gave up his own kidney to a stranger for crying out loud!)
If I'm ever on a lifeboat, I'm teaming up with you, Alex, and Moon Moth. Team Kindness & Fairness FTW!
I don't think Scott endorses any of this as being morally right--he says it's obvious that this is what people will do, not that it's good--and I read it as following his previous statements about how people respond to their incentives (or get out-competed by those who don't).
If you're going to make the argument that these situations aren't realistic because the people in them aren't behaving morally, you've got to contend that they're all allegories demonstrating the incentive structures which have led to many real-life atrocities (committed by people who had plenty of reason to behave morally).
I think it's good, as an intellectual exercise, to speculate about what would happen in the absence of morality.
Because in real life, while morality does exist, it appears that most people are going to choose self-interest over morality most of the time, at least if the self-interest is relatively large. So our speculations about what happens in the total absence of morality are likely to be a reasonable approximation for what happens in real life. And it's instructive to take the models describing an absence of morality and check how well or not well they correspond to real world phenomena.
In all the Backscratcher Clubs, the obligation to favor your fellow club members imposes a cost that I don't think you accounted for, that can sometimes be quite high. For example if you are seen doing favors for someone with a bad reputation, that can hurt your reputation as well. Or if you're a politician and one of your fellows asks you to throw your weight behind their stupid pet cause, it's just squandering your political capital. It actually pays to be choosy in joining clubs of mutual obligation.
This is how Tammany Hall worked, though, and very successfully for a while. There wasn't absolute obligation to do favours for people with bad reputations or crazy notions, so you could say "I can't help you out on this, but I can do that for you".
"It became the main local political machine of the Democratic Party and played a major role in controlling New York City and New York State politics. It helped immigrants, most notably the Irish, rise in American politics from the 1850s into the 1960s. Tammany usually controlled Democratic nominations and political patronage in Manhattan for over 100 years following the mayoral victory of Fernando Wood in 1854, and used its patronage resources to build a loyal, well-rewarded core of district and precinct leaders; after 1850, the vast majority were Irish Catholics due to mass immigration from Ireland during and after the Irish Famine of the late 1840s.
After 1854, it expanded its political control even further by earning the loyalty of the city's rapidly expanding immigrant community, which functioned as its base of political capital. The business community appreciated its readiness, at moderate cost, to cut through regulatory and legislative mazes to facilitate rapid economic growth.
...Throughout the 1830s and 1840s, the Society expanded its political control even further by earning the loyalty of the city's ever-expanding immigrant community, which functioned as a base of political capital. During the 1840s, hundreds of thousands of Irish immigrants arrived in New York City to escape the Great Famine and Tammany saw its power grow greatly.
Tammany Hall's electoral base lay predominantly with New York's burgeoning immigrant constituency, which often exchanged political support for Tammany Hall's patronage. In pre-New Deal America, the extralegal services that Tammany and other urban political machines provided often served as a rudimentary public welfare system."
So by providing services from A to Z for the voters, and making sure that you take good care of local issues and local problems as they arise, you consolidate power. Be powerful enough, and you can afford to turn down doing favours (openly) for bad apples or pet causes, because the majority of your clients know that you will continue to look after them.
Reading that article is a treasure trove of early political history in the USA. Voter fraud and voter suppression is no new thing:
"In the 1830s the Loco-Focos, an anti-monopoly and pro-labor faction of the Democratic Party, became Tammany's main rival for votes by appealing to workingmen. However, Tammany's political opponent remained the Whigs. During the 1834 New York City mayoral governor election, the first city election in which the popular vote elected the mayor, both Tammany Hall and the Whig party, from their headquarters at the Masonic Hall, battled in the streets for votes and protected polling locations in their respective regions from known opposition voters. During the 1838 state election for governor, the rival Whig party imported voters from Philadelphia, paying $22 a head for votes in addition to paying for votes at their polling places. Tammany Hall operatives continued their practice of paying prisoners of the almshouses for votes and also paying for votes at their polling places."
Someone on The Motte recommended the book "Plunkitt of Tammany Hall" about/by a former leader of the organisation, and former New York State Senatoar, George Washington Plunkitt, who describes the methods by which he operated - what he called "honest graft" as opposed to "dishonest graft":
"EVERYBODY is talkin' these days about Tammany men growin' rich on graft, but nobody thinks of drawin' the distinction between honest graft and dishonest graft. There's all the difference in the world between the two. Yes, many of our men have grown rich in politics. I have myself. I've made a big fortune out of the game, and I'm gettin' richer every day, but I've not gone in for dishonest graft—blackmailin' gamblers, saloonkeepers, disorderly people, etc.—and neither has any of the men who have made big fortunes in politics.
There's an honest graft, and I'm an example of how it works. I might sum up the whole thing by sayin': "I seen my opportunities and I took 'em."
Just let me explain by examples. My party's in power in the city, and it's goin' to undertake a lot of public improvements. Well, I'm tipped off, say, that they're going to lay out a new park at a certain place.
I see my opportunity and I take it. I go to that place and I buy up all the land I can in the neighborhood. Then the board of this or that makes its plan public, and there is a rush to get my land, which nobody cared particular for before.
Ain't it perfectly honest to charge a good price and make a profit on my investment and foresight? Of course, it is. Well, that's honest graft.
Or supposin' it's a new bridge they're goin' to build. I get tipped off and I buy as much property as I can that has to be taken for approaches. I sell at my own price later on and drop some more money in the bank.
Wouldn't you? It's just like lookin' ahead in Wall Street or in the coffee or cotton market. It's honest graft, and I'm lookin' for it every day in the year. I will tell you frankly that I've got a good lot of it, too.
I'll tell you of one case. They were goin' to fix up a big park, no matter where. I got on to it, and went lookin' about for land in that neighborhood.
I could get nothin' at a bargain but a big piece of swamp, but I took it fast enough and held on to it. What turned out was just what I counted on. They couldn't make the park complete without Plunkitt's swamp, and they had to pay a good price for it. Anything dishonest in that?
Up in the watershed I made some money, too. I bought up several bits of land there some years ago and made a pretty good guess that they would be bought up for water purposes later by the city.
Somehow, I always guessed about right, and shouldn't I enjoy the profit of my foresight? It was rather amusin' when the condemnation commissioners came along and found piece after piece of the land in the name of George Plunkitt of the Fifteenth Assembly District, New York City. They wondered how I knew just what to buy. The answer is—I seen my opportunity and I took it. I haven't confined myself to land; anything that pays is in my line.
For instance, the city is repavin' a street and has several hundred thousand old granite blocks to sell. I am on hand to buy, and I know just what they are worth.
How? Never mind that. I had a sort of monopoly of this business for a while, but once a newspaper tried to do me. It got some outside men to come over from Brooklyn and New Jersey to bid against me.
Was I done? Not much. I went to each of the men and said: "How many of these 250,000 stories do you want?" One said 20,000, and another wanted 15,000, and other wanted 10,000. I said: "All right, let me bid for the lot, and I'll give each of you all you want for nothin'."
They agreed, of course. Then the auctioneer yelled: "How much am I bid for these 250,000 fine pavin' stones?"
"Two dollars and fifty cents," says I.
"Two dollars and fifty cents!" screamed the auctioneer. "Oh, that's a joke! Give me a real bid."
He found the bid was real enough. My rivals stood silent. I got the lot for $2.50 and gave them their share. That's how the attempt to do Plunkitt ended, and that's how all such attempts end.
I've told you how I got rich by honest graft. Now, let me tell you that most politicians who are accused of robbin' the city get rich the same way.
They didn't steal a dollar from the city treasury. They just seen their opportunities and took them. That is why, when a reform administration comes in and spends a half million dollars in tryin' to find the public robberies they talked about in the campaign, they don't find them.
The books are always all right. The money in the city treasury is all right. Everything is all right. All they can show is that the Tammany heads of departments looked after their friends, within the law, and gave them what opportunities they could to make honest graft. Now, let me tell you that's never goin' to hurt Tammany with the people. Every good man looks after his friends, and any man who doesn't isn't likely to be popular. If I have a good thing to hand out in private life, I give it to a friend—Why shouldn't I do the same in public life?
Another kind of honest graft. Tammany has raised a good many salaries. There was an awful howl by the reformers, but don't you know that Tammany gains ten votes for every one it lost by salary raisin'?
The Wall Street banker thinks it shameful to raise a department clerk's salary from $1500 to $1800 a year, but every man who draws a salary himself says: "That's all right. I wish it was me." And he feels very much like votin' the Tammany ticket on election day, just out of sympathy.
Tammany was beat in 1901 because the people were deceived into believin' that it worked dishonest graft. They didn't draw a distinction between dishonest and honest graft, but they saw that some Tammany men grew rich, and supposed they had been robbin' the city treasury or levyin' blackmail on disorderly houses, or workin' in with the gamblers and lawbreakers.
As a matter of policy, if nothing else, why should the Tammany leaders go into such dirty business, when there is so much honest graft lyin' around when they are in power? Did you ever consider that?
Now, in conclusion, I want to say that I don't own a dishonest dollar. If my worst enemy was given the job of writin' my epitaph when I'm gone, he couldn't do more than write:
"George W. Plunkitt. He Seen His Opportunities, and He Took 'Em."
Also, if you want votes from the opposition or the moderate middle or the squishy, get to the kids first and bribe them with candy 😁
"Chapter 6. To Hold Your District: Study Human Nature and Act Accordin'
There's only one way to hold a district: you must study human nature and act accordin'. You can't study human nature in books. Books is a hindrance more than anything else. If you have been to college, so much the worse for you. You'll have to unlearn all you learned before you can get right down to human nature, and unlearnin' takes a lot of time. Some men can never forget what they learned at college. Such men may get to be district leaders by a fluke, but they never last.
To learn real human nature you have to go among the people, see them and be seen..1 know every man, woman, and child in the Fifteenth District, except them that's been born this summer—and I know some of them, too. I know what they like and what they don't like, what they are strong at and what they are weak in, and I reach them by approachin' at the right side.
For instance, here's how I gather in the young men. I hear of a young feller that's proud of his voice, thinks that he can sing fine. I ask him to come around to Washington Hall and join our Glee Club. He comes and sings, and he's a follower of Plunkitt for life. Another young feller gains a reputation as a baseball player in a vacant lot. I bring him into our baseball dub. That fixes him. You'll find him workin' for my ticket at the polls next election day. Then there's the feller that likes rowin' on the river, the young feller that makes a name as a waltzer on his block, the young feller that's handy with his dukes—I rope them all in by givin' them opportunities to show themselves off. I don't trouble them with political arguments. I just study human nature and act accordin'.
But you may say this game won't work with the high-toned fellers, the fellers that go through college and then join the Citizens' Union. Of course it wouldn't work. I have a special treatment for them. I ain't like the patent medicine man that gives the same medicine for all diseases. The Citizens' Union kind of a young man! I love him! He's the daintiest morsel of the lot, and he don't often escape me.
Before telling you how I catch him, let me mention that before the election last year, the Citizens' Union said they had four hundred or five hundred enrolled voters in my district. They had a lovely headquarters, too, beautiful roll-top desks and the cutest rugs in the world. If I was accused of havin' contributed to fix up the nest for them, I wouldn't deny it under oath. What do I mean by that? Never mind. You can guess from the sequel, if you're sharp.
Well, election day came. The Citizens' Union's candidate for Senator, who ran against me, just polled five votes in the district, while I polled something more than 14,000 votes. What became of the 400 or 500 Citizens' Union enrolled voters in my district? Some people guessed that many of them were good Plunkitt men all along and worked with the Cits just to bring them into the Plunkitt camp by election day. You can guess that way, too, if you want to. I never contradict stories about me, especially in hot weather. I just call your attention to the fact that on last election day 395 Citizens' Union enrolled voters in my district were missin' and unaccounted for.
I tell you frankly, though, how I have captured some of the Citizens' Union's young men. I have a plan that never fails. I watch the City Record to see when there's civil service examinations for good things. Then I take my young Cit in hand, tell him all about the good thing and get him worked up till he goes and takes an examination. I don't bother about him any more. It's a cinch that he comes back to me in a few days and asks to join Tammany Hall. Come over to Washington Hall some night and I'll show you a list of names on our roll' marked "C.S." which means, "bucked up against civil service."
As to the older voters, I reach them, too. No, I don't send them campaign literature. That's rot. People can get all the political stuff they want to read—and a good deal more, too—in the papers. Who reads speeches, nowadays, anyhow? It's bad enough to listen to them. You ain't goin' to gain any votes by stuffin' the letter boxes with campaign documents. Like as not you'll lose votes for there's nothin' a man hates more than to hear the letter carrier ring his bell and go to the letter box ex pectin' to find a letter he was lookin' for, and find only a lot of printed politics. I met a man this very mornin' who told me he voted the Democratic State ticket last year just because the Republicans kept crammin' his letter box with campaign documents.
What tells in holdin' your grip on your district is to go right down among the poor families and help them in the different ways they need help. I've got a regular system for this. If there's a fire in Ninth, Tenth, or Eleventh Avenue, for example, any hour of the day or night, I'm usually there with some of my election district captains as soon as the fire engines. If a family is burned out I don't ask whether they are Republicans or Democrats, and I don't refer them to the Charity Organization Society, which would investigate their case in a month or two and decide they were worthy of help about the time they are dead from starvation. I just get quarters for them, buy clothes for them if their clothes were burned up, and fix them up till they get things runnin' again. It's philanthropy, but it's politics, too—mighty good politics. Who can tell how many votes one of these fires bring me? The poor are the most grateful people in the world, and, let me tell you, they have more friends in their neighborhoods than the rich have in theirs.
If there's a family in my district in want I know it before the charitable societies do, and me and my men are first on the ground. I have a special corps to look up such cases. The consequence is that the poor look up to George W. Plunkitt as a father, come to him in trouble—and don't forget him on election day.
Another thing, I can always get a job for a deservin' man. I make it a point to keep on the track of jobs, and it seldom happens that I don't have a few up my sleeve ready for use. I know every big employer in the district and in the whole city, for that matter, and they ain't in the habit of sayin' no to me when I ask them for a job.
And the children — the little roses of the district! Do I forget them? Oh, no! They know me, every one of them, and they know that a sight of Uncle George and candy means the same thing. Some of them are the best kind of vote-getters. I'll tell you a case. Last year a little Eleventh Avenue rosebud, whose father is a Republican, caught hold of his whiskers on election day and said she wouldn't let go till he'd promise to vote for me. And she didn't."
There may be a lot of howling over Donald Trump, but I feel he would have been right at home in Tammany Hall era Democratic party of New York!
Scott Alexander, the place to go for people who want to understand and mitigate racism out of genuine curiosity rather than guilt.
(social media is where guilt-motivated people go, with academia now mainly adding an intellectual spin to norms that emerged online rather than leading the curve like the late 2000s).
Regarding the lifeboat games, a natural Schelling point is that "rulebreaking is bad". When Albert shouts to eat Bob, Juanita pops up and says "It looks like we can't trust you to keep agreements, but fortunately it sounds like you just volunteered. There's nothing special about Bob, but you, Albert, we all know exactly what you'd do if you had the chance."
Which would would you prefer to live in? You get to choose!
The kind of people who'd rather eat Juanita than Albert would probably kill the rest of us in our sleep. They're moral monsters and must be eliminated. No mercy. (Mostly kidding. I don't think I'd actually make that argument.)
It seems to me like a subsequent "obvious" outcome is that, being granted so much status by the backscratchers, actual orphans co-opt the club and turn it into a lifeboat game serving their own purposes. Non-orphan members are incentivized to stay in the club, since membership is still a requirement for access to orphan-club-controlled institutions, but they also pay higher dues that actually benefit the orphans rather than just enough to act as free-rider deterents. A hallmark of this process would be increasing punishments for dissenters/non-members to further incentivize non-orphans to maintain membership despite its rising cost. The reason why these high costs aren't rejected, like in the adoption example, is that the orphans themselves do not suffer from increasing membership costs. Rather, they benefit from it since it's now lifeboat game (aka don't stick out and get eaten) and they are Albert in that they can leverage the fact that their favor serves as the defining in-group signal in order to enforce compliance.
If you are a non-orphan who was previously benefiting from orphan club membership, what can you do? The most authoritarian and self-serving members of the very group whose promotion defines your club are the ones you need to demote to reestablish symbiosis, but the second you do, you get Alberted. In a cruel inversion, the original members find themselves abandoned by the organization that once favored and protected them.
Orphanship might not actually be susceptible to this sort of comendeering since what orphans probably most want is to not be orphans, aka get adopted, at which point they lose their priviledged orphan status (POS). Coopted backscratcher clubs targeting individuals with the permanent, highly identifiable traits that serve well in lifeboat games would be much more durable.
That's an interesting countermove. The viability of identifying as a member of the club-defining group as a route to regain its benefits probably depends on how socially malleable the identified trait is. How socially malleable the trait is is itself socially malleable, so I'd expect most of the discourse to take place on this meta-level.
In any case, it seems inevitable that dead-parent orphans would vehemently resist their power being diluted by potentially opportunisic legal orphans, creating space for a counter-club defined by its opposition to legal-disownment exclusionary radical orphans (LEROs).
One would think that Albert, by proposing everyone eat Bob, would self identify as the best candidate to be eaten. Then when time comes to eat some one else, no one would propose a candidate (except possibly himself if he wished to signal his own non-edible worth) instead of drawing lots.
> Something about this surprises me. It’s weird that there’s another solution which is more stable than the fair one of drawing lots. It’s strange that by shouting an obvious suggestion - one that adds no more information - Albert can save his own life with certainty. Still, that’s how it goes.
No, the callee would immediately suggest choosing Albert instead. Now the others have two options, both suggesting the other, and they would have no reason to choose the original callee over Albert. Albert has broken the initial N-way tie by volunteering himself as the most deserving of being eaten, providing the others a better justification to society for their actions than siding against a random victim. Knowing this outcome, no one would shout out a suggested victim in the first place. Even under the axiom that future judgement doesn't matter, only survival, there is no reason the callee wouldn't suggest the caller, so the caller would at best make themselves one of two possible victims, so no one would choose to do that.
Yes, Albert has revealed himself to be an amoral schmuck who's willing to throw an innocent person under the bus to increase his chances of survival. Why would you want to keep him around in a life-or-death situation?
That was my reaction as well, although I think the more general principle still holds in cases where proposing a Schelling Point is less clearly villainous. Scott used a very similar lifeboat scenario many years ago as an illustration of dynamics at work in a diplomacy game where the players were five SAIA Visiting Fellows, plus me (*) and another one of Scott's online friends who was local to the area and got pulled in to round out the group of seven. Once the game started, the five SAIA Fellows mostly followed through on a prior agreement to team up against the two outgroup players and then declare a draw. Unbeknownst to Scott (and rather strengthening his interpretation of the dynamics at work), me and the other non-SAIA player (let's call them Alex) also teamed up on the basis of ingroup bias, since Alex and I were and are long-time friends. One of the five SAIA players wavered somewhat from their agreement, but came back around once it became clear that Alex and I were going to lose.
Scott described running that diplomacy game as a mistake and the game as being one of the worst he's ever seem. I respectfully disagree, as I had a great time. Alex and I were both much better tactical diplomacy players than the SAIA crew, so we were able have an enjoyable time going down swinging, and I had fun trying to split people off the coalition even if I didn't succeed. Moreover, it's my attitude that if you can't have fun despite losing for bullshit reasons, you have no business playing Diplomacy.
(*) If the character of Erica was named as a shout-out to me because of this, I appreciate it.
After Albert shouts: "Let's eat Bob!" shouldn't Monty say: "You still have no idea how Bob will taste! I'm now going to offer everyone a taste of Charlotte, just a morsel of her flesh, not enough to kill her. Pretty gamey, isn't she? Now that you've tasted Charlotte, do you eat Bob or choose someone else? Keep in mind that someone on this lifeboat will be delicious."
One thing that troubled me reading this is that while the lifeboat context is one where gains are at the expense of the excluded, a zero-sum game, it's not obvious that it's also the case with the backscratching context. To the degree that membership is an informative signal and backscratching is a positive sum game, nobody has to pay a price for the gains from collaboration. The benefits could be redistributed back to the excluded. Backscratching clubs are not necessarily evil.
I think that if you have to start figuring out the world from some empirical question, a good place to start is how often are we in a zero-sum interaction and how often are we in a nonzero sum game. Let's call it P(zerosum). My P(zerosum) is somewhere around 1%. That's what makes me a relatively nice person. If I saw evidence that it's higher than that it would change who I am. Any evidence that moves it up will make me more evil. Any that moves it down will make me more good. This is where morality and epistemics are conjoined twins. Conjoined twins who are screaming in agony.
I'd always sort of assumed that they and the Red Martians were compatible that way, so that scavenger animals would eat both. It's been a long time, but did any of the Green hordes ever engage in cannibalism?
For the people debating which Schelling point would take precedence in the lifeboat example - I think the point is not to show that one particular Schelling point is a more likely candidate than the rest, but rather to show that some Schelling point will almost certainly be accepted within the group.
My problem with your hypothetical is that it rests on the assumption that everyone on that boat is trying to minimize their chances of being eaten (or maximize the odds of surviving) when in fact some (I really believe most) are more concerned about avoiding responsibility for killing another human being. Even drawing lots might be considered participation in the execution of another. I think historically, people don't draw lots, they wait to see who dies first, which absolves everyone of guilt.
In terms of sociology, it can be described by Bourdieu: identities and beliefs are symbolic capital invented/articulated by intellectuals as a tool to create a new group within a field. This new group then uses the tool as cultural capital to struggle for status with the other groups already there, some of whom may take the ideas as part of their own toolkit.
Having watched a dozen seasons of survivor over the past couple years, I have to make a point about the very first example (It doesn't undermine the main gist of the post). When Albert shouts "Wait, let's all kill and eat Bob!", the other most likely outcome is that Albert gets eaten. Has painted himself as a target. Even if he wins the first vote (this can be accomplished by shouting the least popular person's name), he will then become the biggest target on the second vote. This is the primary incentive not to try something like this.
Yeah, I've played a fair amount of Mafia and this dynamic is very real among inexperienced players. If you aggressively advocate for lynching someone day one you often succeed, but also put yourself in real danger day two.
Now, this may still be worth it because your goal in Mafia isn't merely to survive but to catch the mafia, so among higher level players the dynamic changes. But at the base level of people being put in a new and unfamiliar situation, speaking up and drawing attention to yourself is very risky.
First of all, people are adaptation executors, not utility maximizers, and nobody goes around consciously inventing such strategies (because that would make them evil). So we probably should look for stuff that worked in tribes with less than 150 people that existed for a million or so years of human history.
In such tribes you'd have a rich preexisting network of social relationships, like I want to borrow some flint spearheads from the uncle of my wife's niece, so I support his wife's father as the Chief of the Hunt. And if we win, that uncle will say a word for me and I'll also get some choice antelope haunches. That would be more than enough to kickstart an otherwise arbitrary factionalism, without any explicit or implicit rules about backscratching, purple ribbons, or whatnot.
Second, in our societies it works markedly worse, but I suspect that this, and not the median voter theorem, causes elections to be surprisingly close. Or maybe it is a more truthful restatement of the median voter theorem: in a tribe under no threat of extinction, people's main competition is fellow tribesmen, so the goal of any political movement is to invent some polarizing issue such that it alienates a slight minority of the people, then your faction kills and eats them and gets to propagate your politically adept genes.
The entire point of constantly inventing euphemisms for orphans and banning previous euphemisms is to alienate some people and kill and eat them. And some of us seem to be naturally adept at it, which should not be surprising because a million or so of years of evolution since language first appeared.
Finally, pursuing that angle makes a proper complement to https://slatestarcodex.com/2013/03/04/a-thrivesurvive-theory-of-the-political-spectrum/ which did an excellent job explaining how rightwing attitudes are optimized for tribes under a constant threat of extinction, but then kind of floundered and concluded that leftists must be, like, chill hippies? They are not, they are much more vicious actually, understanding the difference in terms of PvE vs PvP societies explains what we can see with our own eyes much better.
I'm not convinced that leftism (of whatever variety) is inherently PvP. It seems like the ideologies ignore PvP entirely, which may be why it often winds up covering up a lot of PvP - it's like the distraction in a magic trick. If everyone's looking over there, then you won't notice that I've killed and eaten Bob.
Of note is that the old fraternal societies were often a substitute for modern insurance. They would vet members to make sure that they were at least somewhat unlikely to be a useless drain on the organization, and then had a system for taking care of those that had suffered some misfortune.
There were of course also other angles (political organization, MLM robe and paraphernalia sales), but this mutual aid aspect was probably the most universal.
I was once a member of the local Grange (rather a moribund outfit these days, as many fraternal societies are), and they still ask at the end of every meeting - "Is any member sick or in distress?"
Mutual aid was a big purpose of fraternal organizations. Another purpose they served was providing what we'd now call "third spaces" that were owned and managed by the organization for the benefit and use of their members, kinda like a middle/working class counterpart to upper-class Gentlemen's Clubs back before the latter term became a euphemism for strip joints. Compared to their main competition in this role, bars and saloons, fraternal organizations would provide the benefits of a curated group of people to socialize with. Saloons also generally weren't an ideal place for non-drinkers to hang out, and their business model inclined them towards encouraging people to clear off and make room for other patrons when they're done buying drinks.
Although I may be overstating this. Fraternal organizations as substitutes for saloons seems like it would predict that fraternal organizations in the US should have boomed during the Temperance movement and Prohibition. But I just looked it up, and it sounds like they were biggest in the late 19th century and were well into decline when Prohibition went national. On the third hand, I also found while looking it up that at least a couple largish fraternal organizations (the Good Templars and the Rechabites) were heavily involved in the 19th century Temperance movement.
I have to say, I suspect there was a bit of racism involved in them, too. :-/
My family was apparently part of the Elks' club for a while, but I literally cannot remember any use we made of that other than getting access to a nearby outdoor pool. It was very well-maintained, and filled with nice people, as I recall.
Backscratcher's clubs in reality suffer from the Groucho Marx problem ("I wouldn't join any club that would have me as a member"). You don't want to join a backscratching club full of people less powerful than yourself (you'll just wind up doing them a whole bunch of favours and getting little in return).
A few ways out of this:
1. Instead of forming a backscratching club that pretends to be a sailing/ club, you form an actual sailing club where people bond over a shared love of sailing. Maybe a little bit of backscratching does occur but it isn't that major a part of the club's activities.
2. You enforce really strict requirements so that everyone is a roughly-equivalent level of elite. This is the model of, say, a London club. But of course it's tricky to start something like this from scratch.
3. The Freemasonry model -- you are super secretive about your club, and allow people to *think* that there's a whole lot of really elite people in it. Once an elite person joins and figures out that everyone else in his local chapter is just some random schmo, you tell them "oh yes, but if you keep going to our meetings then eventually you'll get invited to the secret elite circle where all the really elite people hang out".
All three of these map on nicely to the medieval guild system, which handles these issues through an internal ranking that rewards skill and effort.
1. There's bonding over specific professional crafts that were internally generally cooperative while competing with other guilds
2. There are strict requirements *for each level within the guild.* You'd start by being an apprentice, move on to being journeyman, then finally graduate to being a master member. In the best cases, advancing was fairly meritocratic.
3. They were elite and secretive, particularly at the highest level, which still gives some proto-prestige to the apprentices (it's relevant that Freemasons evolved from stonemason guilds).
"You don't want to join a backscratching club full of people less powerful than yourself (you'll just wind up doing them a whole bunch of favours and getting little in return)."
On the other hand, this is how you end up being the big fish in the small pond. Everyone owes you favours, so their families and friends and neighbours need to do you favours. So maybe that means you are now the only grocery store, pub, and undertaker in the town. Or you are the guy who employs everyone. Or you and your family are the elite of the small town. You get to wield a lot of power that way.
So you can always get preferential treatment and skip the queue when you want it. You employ Joe and his cousin Bill and their in-law Tom, so now when your mother-in-law needs her house painted, Tom's cousin Mike does it for you at a hefty discount. Things like that.
Isn't a lot of employment advice about the importance of networking? If you're starting out on your career, you don't have a lot of power and influence yourself yet, but if you know someone who knows someone who can get you that interview at Wiggins Widgets, then in turn you help them out with a call to your Uncle Bill and so forth. That's backscratching - why does the 'more powerful' person in this case help out a friend of a friend? Because of the web of mutual obligations and favours that can be called in later.
Isnt "Joining Backscratcher clubs is an advantage" basically the same claim as "Economic protectionism works"? Except without the fancy arguments about industrial policy.
And two backscratcher clubs could negotiate a free-trade agreement so that members of each could scratch each other's backs. Especially if the two groups are already coordinating against another backscratcher club (or another coalition of backscratcher clubs).
>But it’s easier if everybody comes color-coded so there’s one obvious coalition that occurs to everybody and which they can easily check that they’re a part of. And so on to nationalism, religious conflict, political ideology conflict, and so on.
I don't know, it seems to me like there is a substantial difference between--on the one hand--forming a coalition against short people because most of us are tall(er) and it's convenient to us and they're easy to spot, and--on the other hand--forming a coalition against the As Soon As We Gain State Power Either Electorally Or By Force We Are Going To Literally Kill Everybody Who Is Not A Member of Our Party, Oh And Also We Are Not Accepting New Members So If You're Not In The Party Right Now We Will Kill You Party because if you don't form a coalition against them they will kill you.
I guess you could argue that it's just a difference in degree or something, but It just seems like there are other differences as well. Or is it simply that my second example needs to be categorized as a "later" move in the lifeboat game, one that only takes shape after e.g. Heather and Iolanthe's "let's eat everybody except Heather and Iolanthe" coalition has revealed itself?
One think I would add about the calculations people make to join backscratchers' clubs is that they are mostly subconscious. People might consciously believe they're doing things for noble reasons, but beneath the surface a lot of human behaviour is driven by a desire for status.
Yes, I keep coming back around to Jonathan Haidt's theory in "The Righteous Mind", that our ability for rational thought is a by-product of our ability for rationalization and persuasion, to make it easier to form coalitions against other humans.
I think most societies have advanced to the point where they develop immunity systems to the Alberts and Backscratchers Clubs of the world. Eg, by voting to kill Alberts first.
To some extent, opposition to (perceived) Backscratchers Clubs can be seen as one of the main stated reasons behind animosity towards merchant minority ethnic groups (Chinese in Southeast Asia, Jews in many places, Igbo in Nigeria, etc).
"Any coalition of 51%+ can do this. But it’s easier if everybody comes color-coded so there’s one obvious coalition that occurs to everybody and which they can easily check that they’re a part of."
which describes how and more importantly WHY this plays out exactly as it does in the very real life and death stakes of maximum security prisons, and why prison gangs are all race-based.
" But a third reason - linked to the second - is as cover for a backscratchers club."
On the one hand, this is the way we of the 21st C have been primed to interpret everything.
On the other hand, Robert Trivers and _Mother Night_ both have a point...
The best way to fool someone completely is to start by fooling yourself, ie by really really believing your story. And if you really really believe your story, and your story is that you do various good things, then, uh, what exactly is the issue?
I think the 21st C has swung the pendulum way too far on the side of cynicism, to the point where we simply cannot believe that people would ever do anything nice for anyone. This is even seeping into the most banal of entertainment where one of the features I've noticed in recent movies as opposed to say movies of the 70s or earlier, is what "heroes" do is go out and save their families, not strangers. A 60s movie would have us believe that someone might be willing to risk their life for the sake of strangers or an idea; a 2010s movie cannot imagine such a strange possibility.
And I'm not sure this makes us any the wiser about humanity; it's just another form of WEIRD. We haven't become smarter about understanding or modeling the human condition, we've just enforced another version of our culture on the world and claimed that it both is and should be universal.
Relevant real world dynamic: I've known a businessman who was a Liberal party member to say he would always stay a part of the Liberal party regardless of their policy positions, purely because of the business connections he made through being involved in the party.
The Liberal party is not purely a backscratchers club for business people (the most powerful people in the party are dedicated political animals many of whom have never worked in business). But being a pro-business political party is a great Schelling point for it to serve that function along with its other purposes.
I'm just curious, which country's Liberal party is this? Considering you describe them as pro-business, I'm guessing it might be Australia rather than Canada, but too many other countries have Liberal parties at various positions on the political spectrum to be 100 percent sure.
I might suggest “Member of the liberal party, here in Australia”. It’s a bit annoying because Americans don’t have to do that but it’s there world right now.
"the stories in this section seem suspiciously like the Establishment"
I don't know if that's quite right.
I've stated in a few places that the way I think this works is that some genuinely smart person discovers some heuristic that's useful - it makes life more pleasant, it helps with science/technology, it helps you understand people, whatever. Then what?
The inevitable pattern is that various other genuinely smart people take up the heuristic because it is genuinely useful. BUT other midwit people see something else in the heuristic, that it can be weaponized, that people not using for whatever reason can be demonized.
This is *always* possible. Look at the petty hatreds that exist within a group of "technology enthusiasts". Maybe you imagine these people spend their time actually educating each other about how neural nets work, or how to simulate branch predictors. But no, 99% of their interaction is screaming matches about why Apple is better than Microsoft, or RISC-V is superior to x86.
Most people's favorite thing in the world is to hate someone else. And so when any new idea comes along, the first question is: "how can we use it to hate someone?" There's no difference between hating someone because they believe Christ is of one nature when of course we all know he is of two natures; and hating someone because they believe homelessness can be solved by providing housing when of course we all know it's a problem of mental illness. In both cases it's UTTERLY UNIMPORTANT what the underlying reality may be -- there may be no such underlying reality, or something not amenable to empirical testing, or both sides mean something different by various key words. But the point is not a difference in words, it's I have my tribe and you have yours.
Point is, your model is that DEI/Woke succeeds as an ideology because it provides an uncoordinated way for a group ("the elites") to scratch each others's backs. My model is that that's a convenient side effect, but it's not primary. Hatred is primary, the fact that it can be channeled to benefit some of the elites is just a spandrel.
If you look at this in rational terms, DEI/Woke is not improving the overall wealth/capability of the US.
It's not even improving the overall wealth/capability of the DEI/Woke elite as whole, not long term, not once the bills start coming in.
It's not even improving the relative social standing of DEI/Woke elite, since the only social standing they care about is within-group, and that's a fixed sized pie.
But what it IS doing is providing a steady stream of hate-fueled dopamine every day. The Two Minute Hate is one of the more overlooked elements of 1984, but to my eyes one of the elements that rings most true.
So how do you avoid this, the morphing of any good heuristic into a weapon of hate? I have no idea!
One element that may help is having strong restrictions on who can join your club. If you have to demonstrate real mathematical ability to become a physicist, then while physicists as individuals may hate each other, the system as a whole most cares more about physics and truth than about those hatreds. Something like that COULD (maybe?) save Social Science if we were willing to say that the number of Official Social Scientists needs to shrink to 1% of where it is today, and those 1% need to demonstrate serious competence in *something* difficult (math sure, or historical languages, or neuroscience, but *something* beyond "I is interested in peoples").
Another element that may help is having a conservative establishment that is tasked with manufacturing consent, with others not getting much of a word in, not until they join the establishment. That's been part of the traditional story (blown apart in Europe in the Reformation, then in a different version in the US in the 60s). We're supposed to believe that manufacturing consent is a terrible solution. I'm not so sure; as always the comparison is not with some perfect world of your imagination but with the actual real world of everyone empowered to say whatever craziness they like.
Scott's story is that "The Establishment" is doing this to ensure that they have better lives. In the case of the establishment as DEI/Woke, I've already said that I think hae motivation is more hatred of perceived enemies than back-scratching. But for other Establishments, like eg US WASPs, or many nobilities for many centuries, I think there is a genuine element of appreciating just how bad chaos can be in the absence of manufactured consent, and just how valuable it is to ensure that civilization is maintained, by whatever fictions and even occasional draconian punishments are necessary.
I think the conditions you're describing only apply under conditions of material scarcity. When there's enough to go around, greed and novelty-seeking can beat out hatred. Thus, I am optimistic how things will go politically once the renewable energy transition is complete.
Scott seems to have independently re-discovered Curtis Yarvin's conception of "The Cathedral" in his description of elite ideology as a decentralized but extremely effective backscratching club.
The best solution to the lifeboat problem has to be the nose game.
It has the urgency of shouting random names but it doesn't require choosing anyone else's fate. Albert, who touches his nose first, can secure his survival without making any enemies!
I favor Scenario 3 as a response to Iolanthe's orphan adoption plan, because the plan if successful would exterminate the population of orphans and thus be a horrible genocide.
How arrogant is Iolanthe, anyway, to think she can save orphans when all these important credentialed people have failed to solve the problem for decades?
This lifeboat scenario reminds me of a British quiz show called the Weakest Link. (There are probably similar quiz shows in the US and other countries.)
The quiz starts with a dozen or so contestants standing in an arc, each behind a podium with their name on the front, and a host standing before them fires questions, apparently at random, at individual contestants.
The quiz is in a series of rounds, and at the end of each round, the remaining contestants write with a marker pen on a board the name of one other contestant who they propose to be evicted from the game. After everyone holds up their card to reveal the name they have chosen, the host interrogates one or two of them at random. Some of these questions are about their personal life, but end with enquiring why they voted for who they did. The host then briefly assesses a majority decision and tells the relevant contestant "Goodbye, you are the weakest link!".
Broadly speaking, the strategy adopted by most contestants is that in the early rounds they vote off the genuinely weaker performers. But towards the end, when the number of remaining players has dwindled, they vote strategically to send off the strongest contestants, because these of course would be the most challenging opponents in the final couple of rounds if the voter stays in the game that long. There is also some tit-for-tat revenge voting, as reciprocal payback for previous votes.
The final round is a simple five question play-off, and the single winner at the end receives a token prize. Everyone voted off in earlier rounds receives nothing.
If I remember correctly (it‘s been a long time since I watched this), the prize money was determined by the number of correct answers, and that provided an incentive against voting strong contestants out, resulting in an interesting conflict of incentives. When you say „token prize“, perhaps you mean that the achievable prize money was quite small compared to other quiz shows, like Who wants to be a Millionaire. Anyway, I think it‘s a good example, valuable material for students of Schelling lifeboat situations.
Yes, you are correct. Within each round there is a cumulative money "pot" which starts at zero and increases by some set increments (of the order £100 - I forget the exact figures) for each correct answer. At any time any contestant can shout "bank" to consolidate into the final prize total the amount so far added to this pot and reset the pot amount to zero. It is also reset to zero for any incorrect answer.
Actually, the pot increments may themselves also increase, so there is an incentive to hang on for a while and not simply shout "bank" after each correct answer. But of course, set against that is the risk of leaving it too long and blowing the lot with a wrong answer. Also, I think any money left in the pot when the round ends is discarded. So there is that to take into account as well.
In the early rounds, in which the questions are easier, the total prize money is easier to increase, provided timely "bank" calls are made (and contestants gain kudos for this - It must be easy to forget the banking aspect when concentrating on the questions). But towards the final rounds, the questions become harder and the rounds shorter, due to the smaller number of contestants left. So the pot is harder to accumulate and the amounts smaller, and thus I would have thought the banking in the later stages becomes relatively less important than trying to eliminate stronger contestants.
“This is obvious, right? There’s no downside to being in the club, and the upside is preferential treatment from all existing club members.” Someone has probably already mentioned this, but this isn’t actually true. You are neglecting the cost of having to favor other club members. For example, if you are an employer, having to hire a club member instead of the most meritorious candidate will hurt you and leave you at a disadvantage relative to your competitors for customers’, as long as they aren’t in the club and don’t have to bare this cost themselves*. In fact, since everybody in the club is subsidizing everybody else in the club, if you add the net gain to all members and subtract the net loss, it should come out negative (it’s not 0 due to loss of market efficiency, the employee can’t gain more than the employer loses from being hired, as otherwise they could do a deal where the employee could reimburse the employer for his loss assuming 0 transaction costs, but the employee can gain less than the employer loses)**. Of course, some individuals will gain more from membership than they lose, and some will lose more than they gain, but on net it should come out negative. Of course, this is relative to the counterfactual where there was no club, conditional on their being a club, it may be costlier to stay out than to join it even if you lose more from the club existing than you gain. This is especially true if nearly the entire world is in the club, as than the benefits of staying out are miniscule and the costs are huge, but if the club is small, the costs of joining will be huge, and the benefits miniscule. Of course, if you know everybody will join the club in the immediate future, you would have to be stupid not to join (immediate here meaning the time you spend in net loss is short enough to be outweighed by the future gains of joining early).
“another justification is just “We’re the majority and they’re the minority, and if we all band together to profit at their expense, it probably goes well for us.” Any coalition of 51%+ can do this.” I suspect you know this already, but my inner pedant compels me to mention that being the majority is only relevant in so far as it means you have greater power, and It’s possible to be the minority and still have greater power, e.g nobles in medieval Europe, or whites in Apartheid South Africa. And even if you are the weaker faction, it’s often nevertheless better for you to all band together instead of staying uncoordinated e.g blacks in Apartheid South Africa. And of course you can have several factions fighting for resources and all benefiting from coordination among their members e.g casts in India where I live.
Regarding preexisting models like this, your model is a lot like how I imagined the racial cartel model of discrimination in economics worked, though this was just guessing from the name, as I’m not very familiar with that model.
*The employer employee transaction is just an example, and the same dynamics will apply to any other market transaction where a club member favors another club member.
** this is of course ignoring that assuming 0 transaction costs, if you lose more from being in the club compared to the club not existing, you can just pay other club members to dissolve the club, or not form it in the first place. Though in the real world, that will get you accused of disloyalty and attacked by other club members, since real people aren’t Homo Economicus.
*** in my defense, I do have enough self control not to spew forth on the solution to the oddness of Albert being able to save himself, seemingly without any new information, or how to solve having 8 different shelling points, as I think you know the answer to those, and game theory unconnected from sociology isn’t really the point of your post.
An important consideration I foolishly over looked is that you can coordinate with other club members to extract a larger surplus from non club members. E.g, if you and all your competitors are club members, you can coordinate to form a cartel to force consumers to pay a higher price, as long as you treat non club member consumers worse than consumers who are club members. You can trust your competitors because they would likely get kicked out of the club for betraying a club member in good standing, especially if only one competitor alone betrays you and the other competitors, and might face additional penalties if a club required members to go out of their way to punish traders. So there is no guaranty that the net gain to all members minus net loss to all members from the club existing will be negative. It could be either negative or positive.
The essay is insightful and beautifully written, but I think it misses a crucial point: solidarity.
The emphasis on "backscratching" downplays the more profound concept of mutual support and shared responsibility. When we talk about "restricted circles of extended trust", it's not solely about prioritization but about mutual aid, communal support, and collective well-being. Belonging to the same "club" means sharing a common moral framework and duties, not primarily (if at all) indulging in unfair favoritism.
Just as we naturally treat our families differently from strangers and share a sense of shared destiny, we extend this care to some broader communities. Many of the "hurdles" societies impose on their members are not just about gatekeeping, but ensuring people take these duties and moral codes seriously and deserve the automatic trust group members share.
So, the cynical view that "if everyone joins the backscratchers club it loses its purpose" is flawed, assuming "backscratchers club" stands for a social/ethnic-group. If everyone adopted a universal sense of solidarity and care for all humans, it would create a utopia, not a failed scam.
"the stories in this section seem suspiciously like the Establishment - whether it’s the Catholic Establishment of the Middle Ages, the conservative Establishment of mid-20th-century America, or the progressive Establishment of today. "
These things do not sound very similar to me. The Catholic Establishment (Church plus Divine Right feudalism?) had explicit, legible hierarchies that defined their power. That isn't true of the more modern establishments.
I also think that the modern "Establishments" don't look as much like the OSCs as you believe. Perhaps you're just defining "Progressive Establishment" narrowly around some parts of academia and non-profits. But if you take "the Progressive Establishment" to mean, like, main stream US culture, then I think the concept of OSCs are unnecessary.
What's the distinction between "Corporations and other organizations that want to have mass appeal end up signaling that they like things with mass appeal" and an OSC? Live sporting events regularly feature "Applaud the Military Veteran/Police/Teacher" breaks. Is this just OSC shit? Are those groups all part of The Progressive Establishment? Budweiser will make rainbow cans of beer until it becomes unprofitable to do so. Corporate decision makers will invest in DEI as long as it makes them more profitable by helping them recruit/retain talent more cheaply. Are they all part of an OSC that supports DEI only to leave that OSC and join a different OSC that thinks DEI is bad when the EV calculations change?
The concept of OSCs may be useful in explaining collective action in some circumstances within some organizations, but I think you're generalizing too far.
So, after Daniel founds the first club and everyone joins, what I thought was going to happen--and what I still think should--is this:
1. Someone in Daniel's club (maybe a late joiner and/or a bit of an outcast already) says "screw it, I'm leaving this club and starting my own rival club". In this Other Club, there will be a strict ranking order, based on the order people joined. So this guy, call him Oscar, is king, since he joined first. Except...
2. Does anyone else join? After Oscar makes his announcement, lots of people surely realise this is a golden opportunity to be second or third ranking in the Other Club. But that's meaningless if nobody else joins: you'll just end up being Oscar's personal servant, and ostracised by the rest of the town as well.
3. So everyone considering jumping ship for the Other Club will pause, and look around. If nobody else is moving, and is just waiting and watching as well, then best to stay put. It's even less likely others will join after you do, since the second-ranked position is taken. So you shouldn't join either.
4. But wait. What if you loudly shout "I'm joining the Other Club before you do!"? Then either lots of others make a snap judgement that there's a race to join, and you get the second spot *and* are closely followed by hordes of others who will all be your servants...or hardly anybody follows, and your position is almost worthless. After all, why should anybody take *your* intention to join as reflective of anything but your own weird preferences? It's not like you're some symbol of public opinion.
5. I think there are two solutions to this problem. The first is for Oscar to secretly pay a bunch of people to quickly, one after another, announce their intention to join. Then he creates an illusion of a race to join, everyone else is terrified of being left behind and joins the "race", making it a real race, and Oscar gets a huge populated club where he's king.
6. The second is for some high-profile person who is a natural Schelling Point and a perceived representative of the public to declare that he or she is joining, because he/she knows that everyone else wants to. If lots of people trust this person, or even more importantly if lots of people *think* lots of other people trust this person, then there's a race to join, and again Oscar is king.
The first solution is like an astroturf campaign, a centralised attempt to create the illusion of mass opinion. The second is, it seems to me, basically *the* answer to the question of why politicians' official positions matter so much. Despite the fact that they can and do abandon or reverse them once in office, despite the fact that many of their positions are on things they would have little or no power over in the office they're running for (e.g. presidential candidates taking positions on state laws, on purely cultural questions, and even on hypothetical constitutional amendments) and despite the fact that they may have literally taken the opposite position a year ago...it doesn't matter. The mere act of them publically taking that position, as a widely-perceived symbol of public opinion and the most natural Schelling Point for social coordination, is of *massive* importance, very often decisive, in shaping the direction society goes on that issue. And the people who stand to gain or lose from that position thus have an enormous interest in sacrificing lots to get "mere" lip service from a prominant politician.
(I also hope this description might make rationalists less dismissive of the importance of elections, and less likely to think social changes are either inevitable or random results of meaningless cultural events--like the multiple people blaming the rise of wokeness on teeagers on tumblr (?????)--on the basis of "but politicians have so little actual power".)
Various alumni organizations act as backscratchers’ clubs. The most extreme example is most likely the YC founders alumni group.
The club acts as one of the perks of going through the educational course. Why go to an Ivy League college? Well, one reason is for the networking. Aka backscratchers’ club.
Lets say that there is a sub-community of about ten thousand people all of whom are wealthy shareholders of Fortune 500 companies. These people are not all personally acquainted, but they are linked by overlapping cocktail party invitation lists, so networked communication between them is reasonably cost-effective, if a bit slow.
One year the economic growth fails to meet expectations, and share prices start to fall. Everyone is casually acquainted with stories about the Great Depression, and they all remember the last recession, so they begin whispering about each other behind each other's backs. Someone proposes that they pick a major corporation randomly, and short their stocks, so that the rest will still have access to sufficient credit--
No. People in real life do not act this way. What is the lifeboat supposed to represent? If the entire national economy, then within the lifeboat certain people have hoards of food of varying sizes, and use that hoard to "hire" bodyguards. The choice of who to eat will be anything but random. While elites are taken down by other elites every day, that's more or less interpersonal competition, not some sort of coordinated game play.
Or, is the lifeboat the elite community? In that case, they will be surrounded by a large number of smaller boats, and will pick the hapless inhabitants of one of them to play "Andes Survivors" with. A rising tide lifts all boats, but it doesn't change who lives on a yacht and who lives in a rowboat...
And then, what happens when the little people, who outnumber the elites by at least 2 orders of magnitude, decide they've had enough and start digging out the pitchforks? Which computer simulation predicts the October Revolution?
An obvious Schelling point for the second person to eat would be Albert, who established the idea of coordinating around an arbitrary victim. (Ob course, if Albert knows this, it might be in his best interest not to propose the idea in the first place.)
---
I am not a lawyer, but I think that both legally and morally, there should be a vast gulf between people forming a pact to eat a person chosen by lot and people just coordinating a murder to save ones own skin. Just try for murder and hang anyone who ate someone whom they did not have a pact with. What is society for, if not to provide disincentives for behavior with negative externalities? (Of course, the murderers might also coordinate to cover up their crime, but even a slight probability of failure will make the lot-drawing much more attractive by comparison.)
I would argue that going along with eating Bob is not the correct solution for everyone (except Bob).
Consider Charlotte: if she is is smart she might notice that
a) one kill will likely not be enough to sustain the survivors until rescue
b) she is next in line
So it might be in her best interest to push for the drawing of lots instead.
Suppose you are n survivors, and you estimate that it will take k killings until you are rescued. Also suppose there a c charismatic persons who are unlikely to be eaten before the rest when people coordinate, and you are part of the rest r=n-c.
If you go along with the coordination once the first name is proposed, the probability that you will be among the survivors of the rest group is:
(r-k)/(r-1)
If every killing is decided by lot, your odds would be
(n-k)/n.
For example, if you estimate that the n=10 survivors will eat three people (k=4), and you know that there are two people who are unlikely to be picked before you (c=2, r=8), but don't know anything about your place in the food chain (other than you not being first), then it would be slightly better to insist on the lot (p=0.6) vs coordination (p=0.57)
One point that the post fails to adress is that there is need for actual justifications a la : "We are the 51% etc" when it comes to racism and so on, that justification is inbuilt. Same goes for the orphan-supporters who discriminate against non-orphan-supporters. Human tribalism is hardwired. Evolution has done the game theory for us. The only purpose for justifications is to rationalize the back-scratching that is already happening anyway. (That hardwired back-scratching is also much more messy and inefficient than purely rational back-scratching would be.)
> “The more complicated versions sound like cults, religions, and ideologies.... I’m sure real sociologists have written about these issues, but it was hard to find them and I figured I might as well write this post without citations. Still, if you know who they are and what their terms are, let me know.”
Wow, something I actually know about!
Rodney Stark is the sociologist of religion who pioneered this the “backscratchers club” model, and brought rational choice theory (basically, Econ 101) to religion. Among other things, he created an analytical model to predict which new religious movements are likely to succeed or fail. It was based originally in the studies he did on the Bay Area cults of the 1960s and ‘70s, but he later extended it to try and explain why both early Christianity and Mormonism took off the way they did.
Probably the best place to understand his model is his 1996 article “Why Religious Movements Succeed or Fail: A Revised General Model”. Especially relevant here are factor #3 (“medium strictness” — too demanding to let in free riders, but not too demanding to keep out all but the weird fanatics) and factor #5 (“a religious labor force” — joining a group that demands a lot of your time can counterintuitively be a great way to be lazy).
I’ve (oddly coincidentally) been talking with Eneasz about how religious communities work on the podcast “The Bayesian Conspiracy” recently; he’s been interested in which of their tools we might be able to borrow to cultivate community in the Rationalist space: thebayesianconspiracy.com
My immediate reaction to the first situation was that the second time it happened I'd shout "Let's Eat Albert". That was pure emotion. Albert is a bad person, and I'd expect others to shout it too.
Once we'd got rid of Albert and Bob and I'd had time to think, I'd position myself as someone who's very very in favour of fairness and justice and uncompromising about the principle of actually drawing lots. And then I'd show myself prepared to fight to defend that principle.
That's calculated, but I think it's my emotional reaction too!
Once home, I'd join the various backscratcher clubs if and only if I liked the activities and people involved, at which point I'd become rather dedicated to the club and passionate about it. If I didn't like them, I'd doggedly refuse to join even when it had become obviously a really good idea to join and thus I'd lose out quite badly but keep my arrogance and pride intact.
This is both my emotional reaction and, from a lifetime of observation, how I actually behave in such situations. Calculation tells me to join in earlier, but screw that.
I wonder why nobody cited the lifeboat sketch. Everyone is sick and tired from decades of Monty Python quotes, probably. Anyway, on my lifeboat, I'd found the four-course menu coalition.
I think the answer to how well any of these strategies/clubs will perform often depends on what time depth people are thinking about. This is essentially a Rawlsian point, I guess: if you can get people to think long term, including those future periods when they will not be part of the winning coalition, then they will choose the fairest option. When people think short-term only, unfair options seem more promising to them.
Therefore:
(1) Lifeboat debates are a good way to crystallise this problem, because there's an incentive to think in the short term: you think you'll be rescued reasonably soon.
(2) Humanity now engages in many projects that require timespans longer than a human career/lifetime. Those inevitably create political problems, because relatively few people are going to be good at thinking that long-term; relatively many people will be willing to sacrifice fairness or advantages beyond the term of their own life in favour of short-term gains.
(3) Our ability to think (effectively) about the long term is often very highly constrained by our experience, so teaching history is really important. We won't all experience insidious political unfairness, but we can all learn about it in the classroom.
Interesting story! One question that comes to mind is, is effective altruism a backscratcher’s club?
It seems to me that it definitely wasn’t in the beginning (there were too many calls for sacrifice for that to be plausible), but I could see an argument that the resources of Open Philanthropy and subsequent professionalization have created a kind of Schelling point.
The problem is that (unless I misunderstood something) this presents examples of Backscratchers Clubs that are mostly borne out of simple greed and wish to conspire for mutual benefit, and actual Backscratchers Clubs genuinely seem to mostly come out of other, "higher" pursuits with the backscratching an eventual added emergent behavior.
I.e. at least if we go by the official explanation, Freemasonry really came out of bored gentlemen first getting interested in the Mason guild rites and then self-development, and the whole conspiring for mutual benefit arises after that, and even there the conspiring would probably quite organically develop from things like having a shared space for a lot of people who might indeed derive mutual business benefit from each other and with a belief that they have, indeed, developed themselves to a higher moral level. Same for religions and so on.
The one type of Backscratchers Clubs that genuinely follow the model are criminal conspiracies, but for the rest of them, one might imagine a club where people are genuinely first interested in scratching each other's backs like, a lot - maybe it's a furry thing? - and only after that's been going on for a while they start talking about backscratching deals of other sorts.
After reading the book review about prions, it got me thinking about this post and the danger of disease transmission via cannibalism. Of course Scott doesn't go into the weeds of disease or calorie counts because it detracts from the metaphor of the lifeboat to backscratchers clubs, but I find the application interesting.
Suppose everyone is about to draw lots, and Bob shouts out "I have HIV!" Bob has now created a Schelling point that separates him from everyone else in a negative way - nobody would want to eat the guy with a serious transmissible disease. I don't know how much weight this carries for people actually starving to death, but surely it matters when there are nine other people on the lifeboat.
The next step - does everyone else try to claim some sickness that spares them from being eaten? There is a fine line here where the disease has to be bad enough that no one wants to risk being infected but also has to be believable in the absence of obvious symptoms. I don't think anyone else would be convinced if you also claimed to have HIV - what are the odds two out of ten people on a lifeboat have it? Nor would they believe you have Ebola without hemorrhagic fevering - or maybe they do believe you and toss you over the side. Shouting "I occasionally have symptoms of HSV-1 infection!" is not serious enough to prevent you from being eaten. So claiming disease has to be both serious enough that no one wants to eat you, but not serious/infectious enough that everyone wants to kick you off the lifeboat for their own safety. And it also has to be convincing. After Bob does it first, everyone else runs the risk of being seen as an insincere mimic.
I'm not sure how this strategy plays out in the long run. Maybe everyone makes similar claims, and it becomes pointless like the backscratcher club where everyone is a member. Bob might still be safe, because his initial claim carries more weight. Then again, maybe no one believes Bob's conveniently timed confession. Anyway, something to think about the next time you need to strategize how to avoid being cannibalized.
Does this relate to the mystery of why so many dislike EA, when it's just good people doing good things? The answer would that making a noise about how charitable you are is interpreted as "look at the gig leaf, not what's underneath" .
I don’t think you have to look any further than the well-documented phenomenon of do-gooder derogation to explain that; and from my limited understanding, the research on that phenomenon points to the idea that do-gooder derogation arises because do-gooders make others look bad.
The problem with Erica and to a lesser extent the later orgs' models is the lack of a clear enforcement mechanism against internal defectors. Each club member wants to be favored by others, but not have to favor others. Without clear rules on what that means and procedures for removing members that don't follow them, this can't work. Which is maybe cancellation, in the analogy?
Lifeboat Games and Backscratchers Clubs sound like interesting concepts! They seem to focus on mutual support and creative ways to help each other succeed, which is fantastic. Just like in the Daman game, where strategy and teamwork can enhance the experience, these clubs could thrive on collaboration and smart planning. Building connections and supporting one another is always a winning strategy, whether in gaming or real life. If these ideas involve sharing resources or skills, they can create a strong community. Best of luck with your Lifeboat Games and Backscratchers Clubs—may they bring success and fun for everyone involved.
But it becomes problematic when it inevitably stops being about helping orphans and becomes a backscratching club that has little to do with the original mission. Which is why people who actually care about orphans become disillusioned about the causes they were once passionate about. It's when they see that actually helping orphans has been replaced with a witch-hunting campaign against those using the word orphan in any kind of a metaphor.
> it inevitably stops being about helping orphans and becomes a backscratching club that has little to do with the original mission
Inevitably? Is there not a single charitable or social-good organization anywhere that mostly just does the mission?
I'm honestly not sure. I used to think all organizations existed to do their mission – that's probably due to my being on the spectrum – but then over and over I saw evidence that this model does not match reality. Then I think at some point my model may have been overadjusted towards cynicism; I'm still not sure to what degree I'm overcompensating for the previous naivety, if at all.
I think many organizations are *started* mostly with the goal of pursuing stated mission. I think they then continue for a while working primarily towards the mission, but I do also think that most of them remain small and thus not very effective (eventually dissolving entirely), and the ones that do grow large and influential become victims of their own success in that they become less and less about the mission and more and more about politics and backscratching. I think an organization's best work towards the mission is often done during a pretty brief period: after it has grown large enough to be able to actually do something but before it's grown so established that it's crippled by its own success and overtaken by politics.
I don't have any hard data to support this hypothesis, and as I said earlier, I'm not even sure to what extent it is accurate. This is just the model that I currently estimate as slightly more likely to be accurate than competing models based on anecdotal evidence and general observation. But I'm highly uncertain about my own estimate of its likelihood.
That said, I still need a model to guide my own behaviour wrt various organizations, so I have to pick one even if I am not certain about its accuracy.
From The Onion:
In Retrospect, I Guess We Might Have Resorted To Cannibalism A Bit Early
https://www.theonion.com/in-retrospect-i-guess-we-might-have-resorted-to-cannib-1819583474
Published June 10, 1998
Milton Boyd
Well, I suppose everyone's heard about last week's incident by now, and you probably have a pretty low opinion of us survivors. And, all things considered, perhaps we deserve it. Perhaps we panicked and resorted to cannibalism a bit early. But you weren't there. You don't know what it was like. I just want you to hear our side of the story before you go judging us.
When the six of us got into the elevator on that fateful day, we had no idea what was going to happen. We thought we were just going to take a little ride from the 12th floor to the lobby, just like every other day. Do you think we knew that elevator was going to get stuck between floors? Do you think we got into the elevator saying, "Hey, you know, we should eat our good old pal Jerry Weinhoff from Accounts Payable"? Of course not.
During those first few minutes after the elevator car lurched to a stop somewhere between the seventh and eighth floors, we were still civilized human beings. Everyone kept his cool. We tried pushing the emergency button. ...
One word enters your mind: survive. Survive!
I have no idea how long we'd been marooned when we started edging toward Jerry. Twenty, thirty minutes, time has little meaning when you're in a situation like that. It wasn't a spoken decision, either. We just all looked at each other and knew something had to be done.
It might have been an animal act, but it had a certain logic. Jerry lived alone and had nobody special in his life—no kids, no wife or girlfriend, and his parents had died a long time ago. And, most important, he was the biggest. We figured there was enough meat on him to keep the rest of us alive for days, maybe weeks.
... Maybe it was savage. Maybe it was an animal act. But human teeth are pointed and sharp in front for a reason. Besides, we had no way of knowing that, at that very moment, an Otis Elevator repairman was working to free us.
If you just eat your own, is it still cannibalism?
I'm pretty sure that's the literal definition of cannibalism.
Your own what?
Adherents who bought into the winning ideology.
When a different party takes office, the new party "appoints" 3,000 partisans to engage their political apparatus in government. The Appointed (Anointed?) are expected to further the party's ideology.
The 3,000 in the current clown car will be looking for work if Joe can't carry the torch.
Appropriate verse for the topic:
The Yarn of the 'Nancy Bell' by Gilbert of Gilbert and Sullivan fame:
'Twas on the shores that round our coast
From Deal to Ramsgate span,
That I found alone on a piece of stone
An elderly naval man.
His hair was weedy, his beard was long,
And weedy and long was he,
And I heard this wight on the shore recite,
In a singular minor key:
"Oh, I am a cook and a captain bold,
And the mate of the Nancy brig,
And a bo'sun tight, and a midshipmite,
And the crew of the captain's gig."
And he shook his fists and he tore his hair,
Till I really felt afraid,
For I couldn't help thinking the man had been drinking,
And so I simply said:
"Oh, elderly man, it's little I know
Of the duties of men of the sea,
And I'll eat my hand if I understand
However you can be
'At once a cook, and a captain bold,
And the mate of the Nancy brig,
And a bo'sun tight, and a midshipmite,
And the crew of the captain's gig.'"
Then he gave a hitch to his trousers, which
Is a trick all seamen larn,
And having got rid of a thumping quid,
He spun this painful yarn:
"'Twas in the good ship Nancy Bell
That we sailed to the Indian Sea,
And there on a reef we come to grief,
Which has often occurred to me.
'And pretty nigh all the crew was drowned
(There was seventy-seven o' soul),
And only ten of the Nancy's men
Said 'Here!' to the muster-roll.
'There was me and the cook and the captain bold,
And the mate of the Nancy brig,
And the bo'sun tight, and a midshipmite,
And the crew of the captain's gig.
'For a month we'd neither wittles nor drink,
Till a-hungry we did feel,
So we drawed a lot, and, accordin' shot
The captain for our meal.
'The next lot fell to the Nancy's mate,
And a delicate dish he made;
Then our appetite with the midshipmite
We seven survivors stayed.
'And then we murdered the bo'sun tight,
And he much resembled pig;
Then we wittled free, did the cook and me,
On the crew of the captain's gig.
'Then only the cook and me was left,
And the delicate question,"Which
Of us two goes to the kettle" arose,
And we argued it out as sich.
'For I loved that cook as a brother, I did,
And the cook he worshipped me;
But we'd both be blowed if we'd either be stowed
In the other chap's hold, you see.
"I'll be eat if you dines off me,"says TOM;
'Yes, that,' says I, 'you'll be, '
'I'm boiled if I die, my friend, ' quoth I;
And "Exactly so," quoth he.
'Says he,"Dear JAMES, to murder me
Were a foolish thing to do,
For don't you see that you can't cook me,
While I can and will cook you!"
'So he boils the water, and takes the salt
And the pepper in portions true
(Which he never forgot), and some chopped shallot
And some sage and parsley too.
"Come here,"says he, with a proper pride,
Which his smiling features tell,
"'T will soothing be if I let you see
How extremely nice you'll smell."
'And he stirred it round and round and round,
And he sniffed at the foaming froth;
When I ups with his heels, and smothers his squeals
In the scum of the boiling broth.
'And I eat that cook in a week or less,
And as I eating be
The last of his chops, why, I almost drops,
For a wessel in sight I see!
* * * * * *
"And I never larf, and I never smile,
And I never lark nor play,
But I sit and croak, and a single joke
I have--which is to say:
"Oh, I am a cook and a captain bold,
And the mate of the Nancy brig,
And a bo'sun tight, and a midshipmite,
And the crew of the captain's gig!"
I guess we were all guilty, in a way. We all shot him, we all skinned him, and we all got a complimentary bumper sticker that read, "I helped skin Bob."
The ending has the punchline:
> When I finally got home from work that day, some 50 minutes late, my youngest daughter Kellie ran up to me and gave me a big hug. She said, "Daddy, I'm glad you're home." Daddy, I'm glad you're home. It was at that moment I knew I'd done the right thing.
This sounds horribly familiar. People focus on the small good consequence of their actions, and utterly ignore the cost. I swear that "Compared to what?" is a fully complete counterargument to 90% of the drivel on the internet.
https://www.youtube.com/watch?v=wHz_Uyk4dYQ
"Our only option might be to eat each other to stay alive."
"It's only been, like, four hours. Aren't you resorting to cannibalism a little quickly?"
regarding it being surprising that there is a more stable solution than the random fair one, you might be interested in Ram Rachum's work on emergent dominance hierarchies in reinforcement learning agents: https://arxiv.org/abs/2401.12258
Rachum studies hawk/dove games and shows that in an iterated setting, agents that start off playing mixed strategies will tend towards always playing hawk or always playing dove, with the agent that always plays hawk randomly decided by the first game. This generalises to multiagent settings like the lifeboat game: the stable solution is to randomly decide for each pair of agents which will be hawk and which will be dove, and this usually results in a dominance hierarchy.
Minor typo: "at the expensive of its enemies"
> They’ve heard this story before, so they decide to turn to cannibalism sooner rather than later. They agree to draw lots to determine the victim. Just as the first person is reaching for the lots, Albert shouts out “WAIT LET’S ALL KILL AND EAT BOB!”
As a Bob I can confidently say that any Bob worth their salt would shout: "Albert is trying to defect from a fair decision procedure! Let's eat him!"
Indeed, given social norms and culture, it is not completely clear whether there is a first-mover advantage or a first-mover disadvantage in this hypothetical scenario. When Albert is trying to coordinate people by shouting out “WAIT LET’S ALL KILL AND EAT BOB!”, he also creates a justification for the others to say "ALBERT IS EVIL, LET US HIT HIM!"
that's how righteousness appeared.
For the second time they need to eat someone, they all shout out LET'S EAT ALBERT because Albert has already proven he will shout out someone else's name. By eating him, you're protecting yourself; maybe the lifeboat will be rescued before there is the necessity to eat a third person.
Also you get to feel a little bit better if you survive. Sure, we ate Bob, and I genuinely feel bad about it, but when you think about it, it was really Albert's fault, and when we ate him we were justly punishing him for his selfish defection.
He tasted a little like scapegoat.
Or someone confident with Albert will say, “Let’s listen to Albert’s wise words on whom to eat next”. Would you object, thereby signalling your low confidence in your chances and making yourself a prime Schelling point?
I think it’s a disadvantage. If we imagine this as a vote to see who is eaten, then Albert has voted for Bob. But if Bob votes next, he will certainly vote for Albert, and then we have two people both in the same situation in terms of votes. I think in that case morality favours Bob
If I wasn't sure how many times we were going to have to iterate through this particular problem I probably wouldn't demonize the technique that I may need to resort to at some point.
Bob could also point out that, if they all go for him, afterwards they'll feel guilty and if they're not rescued, they will pick Albert next time for sure, to assuage the feeling.
Even it didn't work, you have a feeling of payback against Albert.
Instead of "let's eat Bob", maybe the biggest guy just starts beating Bob up and that becomes the Schelling point before Bob has time to come back.
Yeah, could work. Though probably that guy would be more in the frame if it goes to R2. The others jump on him.
There actually is an old group therapy structure called The Boat Structure that models this process. Though you push people out rather than eat them to stop the lifeboat sinking. İt gets pretty crazy.
Yeh bad news for Albert potentially - the assassin seldom wears the crown.
Yup.
Could you please share more about this group therapy structure? Googling didn't yield any helpful results.
It's been a couple of decades since I did it last. But from what I recall...
* mark off a small area of the group room using mattresses
* 10-20 people sit inside it. The scenario is that you've escaped from a sinking ship and are in a lifeboat. However the boat has a small leak and someone will have to be thrown overboard (to certain death) so the others can live. Hopefully you're all rescued up soon.
* everyone has a turn to say why they shouldn't be the one, or to volunteer. Then there's a vote and the loser goes out.
* Continue. I can't remember how you vote on the last person, when it's just one v one. But you don't have to fight and there is just one winner.
Best I can do. PS - the process can bring up a lot of deep dark stuff so it's not a good idea to do it without trained facilitation.
That sounds incredibly upsetting. Let me guess, seventies-era therapy? From the era when bringing up strong emotions was seen as therapeutic in itself
Alternatively, the Schelling point is to get the big dude first, because he's clearly a threat who does not respect agreements.
You'd be stuck with a half beaten up Bob who probably wouldn't pull his weight, then get eaten anyway
Eh, we'll put in place a social insurance policy for disabled veterans of wars to preserve cooperation against the threat of defection.
Why waste such a good opportunity to have two guys to eat for the price of bringing one down? Since Bob won’t be much of a threat, let him live till you’re done eating the big dude, so he doesn’t rot.
However, do people coördinate against big dudes like that? After all, we have this notion that if you’ve got a problem with the big dude, the manly thing to do is to tell him to his face and fight him one on one, with no dirty tricks that might help you against his superior strength.
I guess this wouldn’t apply if the rest of the passengers were women—which would make him a glaring Schelling point—and the boat were roomy enough for them to surround him and attack him all at once, though it may still be easier to convince him to sacrifice himself out of chivalry.
When we get to the third level, where it’s about picking a Schelling point, and there’s too many possible Schelling points, we see what the advantage was in the drawing lots - it’s very symmetry makes it a more stable Schelling point.
Precisely. :-) I would be happy to share a lifeboat with you, sir.
one idea is that there's an anonymous note left that says, "let's all just eat bob". the next day, there's anonymous notes from everyone that lists everyone. and so on and so forth. this is one situation where i'm not sure a Land Value Tax would solve anything, unfortunately.
Tax whatever they're writing notes on! A rare case where it would be advantageous to appoint someone illiterate as the treasurer.
Exactly. And I think most groups would then vote to eat Albert. If not in the first round then in the next as he has three big marks going against him:
- he effectively killed bob
- he broke convention and isn't to be trusted
- he made each person's odds slightly less and resentment would surely impact future decisions
The first mover advantage for Albert far from gives him 100% of avoiding death. I'd say it highlight increases his chances of death.
> As a Bob I can confidently say that any Bob worth their salt would shout: "Albert is trying to defect from a fair decision procedure! Let's eat him!"
Getting to finish the first of those two sentences doesn’t sound very Bob-like to me. In my experience, you’re cut off and either ignored or punished for acting above your station; perhaps they’ll be offended by your big words they can’t be bothered to engage, or by your nerve in accusing anyone else of anything; perhaps they’ll just command you to calm down or to stop yelling.
In this case, they’ll probably cut you off with their knives or with their eager teeth.
When Albert shouts "hey let's eat Bob", shouldn't Bob shout "let's eat Albert instead"? Which would bring it to a vote, which Albert should lose*. Even if that fails, Albert should be the next Schelling point for the person to be eaten for having suggested Bob, not the blonde.
*Why should the majority prefer to eat Albert? Albert's effectively trying to move the distribution of victims from uniformly random to favoring whoever can think up a good strategy. But we know Albert can think up good strategies, which means this distribution is worse for the majority of non-alberts.
It's just a hypothetical scenario with a lot of simplification. IRL Albert wouldn't shout 'let's eat Bob' in the first place, they would probably talk with the non-Bobs one-on-one about how Bob is planning to coordinate eating them and how they should deal with this problem.
This simplified scenario plays an important role, namely, presenting the dilemma in the starkest terms possible. Thus eliminating such extraneous machinations you have posited.
No one is looking to pitch this as a vehicle for a Netflix mini series!
https://en.wikipedia.org/wiki/Circle_(2015_film)
yes, that was my point with my comment. it's an abstract game not an actual situation hence options which would be available in the real world are not available in the scenario. Do note that were it an actual real world scenario, one-on-one talk would easily be possible when stuck on a life-boat, because they would take turns staying up looking for help in the night and when Albert is the one awake, he could wake up the people he wants to speak with.
I think the idea is that saying "Let's eat Bob" is a clear Schelling point, because it's the first thing anyone said and satisfies everyone (except Bob) completely, whereas letting Bob say "let's eat Albert" lets Albert say "Well let's eat Carol" and so on.
Shouldn't Albert specifically try to get rid of Bob instead of Carol?
I know this is more complex than the metaphorical/analogous situations you're drawing here, but I really see *Albert* as the clear Schelling point. He's the only one, in fact, I would be comfortable killing and eating, due to his actions and willingness to defect for personal gain.
Yes! Want to start a club? ;-)
But if Albert then tries to move on to "let's eat Carol - no, let's eat Heather - no, let's eat Iolanthe" then everyone will agree even more strongly "let's eat Albert". Bob hasn't said "let's eat Carol", so Carol should support Bob in the "let's eat Albert" coalition.
Question is: is it a Unique Schelling Point?
Given a discussion in another thread, I was primed for this post to head for the Rawlesian veil of ignorance...
Put the veil of ignorance over Bob's head, then eat him.
They would want to disincentivize someone breaking from existing agreements in a way that counterfactually could have harmed them. (blah blah, Rawls veil).
Albert only proposed this because he predicted that the other people in the group would go along with it. However, this 1/9 chance to die — before Albert chooses — is worse than a 1/10 chance (because Albert doesn't pick himself), and thus you should refuse the deal even if there's technically a fact of the matter about who was chosen.
Seems like a classic prediction problem where it would have been better for everybody to pre-commit. Of course, this assumes that everyone believes Albert would choose randomly from the 9 others.
I can see Bob being a good Schelling point among actors who aren't rational enough and so they agree to Albert's plan.
However, I do think Albert is still also significantly more of a schelling point than Carol. He's the one who tried moving things so he would always win at the expense of others, giving everyone else a 1/9 chance of dying rather than a 1/10. Sure, if you think purely causally he's given you a 0% chance (except for Bob), but he chose randomly before his announcement.
Then he becomes a decent point of agreement for trying to manipulate others. (Which is probably part of why humans have the intuition of get rid of the schemer first)
Exactly this. Albert speaking first also asserts his dominance and position of leadership.
If you consider Albert is Donald Trump it becomes interesting when you ask “Well why didn’t everyone just eat Albert”. People aren’t equal even in theoretical scenarios and certain personalities will dominate. There’s a reason Trump uses projection on his enemies because it becomes almost impossible to be on the receiving end and argue that it’s actually the other way around.
Because at that instant, when he was the first and only to shout, you have 2 choices:
- Agree and have a 0% chance of dying
- Try to coordinate and have a >0% chance of dying
This doesn't work in an iterated game, but unless everyone pre-committed not to do this (e.g. by coming from a culture that vilifies this) then _at that point_ the clear best action is to eat Bob
If you are first to agree, the others will see that you are a schemer like Albert though. That would have consequences.
Edit: which I see you acknowledge in your final sentence.
The greatest trick the devil ever pulled was to convince the world that there are non-iterated games.
This deserves to be the title of a long post or paper.
Some people are willing to punish defectors even at a cost to themselves. Granted, the number who will do so when the cost approaches a 100% chance of dying may be small, but surely it goes up quite a bit as the chance of dying goes down. In any case, a quick decisive action from the one person on the lifeboat who has the courage of their convictions even unto death could convince the other seven people (who have to eat *somebody,* after all) and Bob (who likely doesn't need much convincing) to get behind the would-be hero and eat Albert.
But whoever suggest that they eat Albert would stand out, and be the next obvious choice. So people have an incentive to not be the first to nominate Albert.
Good point. This is like the sleeping lion problems
Fans of Survivor are very familiar with this dynamic.
Maybe I'm missing something but as per the comment you're responding to it would be Bob who suggests Albert and he really has nothing to lose here. So isn't it irrelevant that nobody else would suggest Albert, as Bob is already doing the suggesting for them?
Right. What Bob should do is spread the rumour that Albert has been trying to get everyone to eat Charlotte. Now, whether the group decides to align with the defector or eat the defector, Bob is safe.
This thread has taught me that if I ever do get into a lifeboat then it's important to start strategising immediately. While everyone else is still thinking about "I still can't believe our ship sunk" and "where's my luggage" and "did everyone get off okay", I'll be busy thinking about how to ensure I'm the guy who gets to eat everyone else.
It's lifeboat, so Bob could not talk privately (or plausibly allege that Albert has had private conversations) with anyone.
(Ignoring all the metaphor for a second) Drawing lots is actually a stable equilibrium. Shouting out someone’s name incentivizes doing so as fast as possible (because otherwise someone else might shout out yours first). Drawing lots encourages waiting as long as possible because everyone’s life is on the line. Rationally you’d rather endorse the system that drags out the killing even if you could have a safe situation in round 1. Also, maritime law typically only protects you if you draw lots. (Edit: I’m wrong on this last point. See below)
Let's eat Scott.
But that would increase the probability of orphans!
> Also, maritime law typically only protects you if you draw lots.
No way it's in the law?
Wiki suggests it's a custom - https://en.wikipedia.org/wiki/Custom_of_the_sea, still I had no idea it's a thing.
You’re right, I’m several hundred years out of date on the law there (the Dudley case ruled even drawing lots as illegal). Good thing you replied before my next sailing trip.
I thought the Dudley case was at least a little ambiguous on that point. The defendants had killed the most junior member of the crew (the cabin boy) as a Schelling Point rather than drawing lots as customary, and the decision specifically called them out for choosing their victim consciously (arrogating authority that didn't belong to them, and setting a precedent that if upheld could act as a cover for killing out of animis) as well as for abusing their seniority to prey on their subordinate to their own benefit rather than sacrificing themselves to ensure the safety of those under their care.
The decision does also generally reject necessity as a defense for murder, but taken together it seems like the logic of the ruling would probably support killing a volunteer and would have a decent chance of supporting a unanimously voluntarily decision to draw lots.
Thanks for the link, so the victim in Dudley was called Richard Parker! Yann Martel must surely have chosen that name for Life of Pi with the case in mind.
WOW!
Drawing lots is not a stable (Nash) equilibrium? If the possible actions are (a) support drawing lots, and (b) shout out the name of someone else as fast as possible; then given the provided assumptions any participant is better off switching to (b) if everyone plays (a).
There are rational arguments for why I dislike backscratcher movements, which I might or might not get to writing down someday.
But right here and right now, I just want to shout and to cry and to thunder against the concept! I abhor the notion, I hate coordination capture, I despise backscratchers and all they stand for.
If there was an eldritch ritual that would remove backscratching from the world in all its forms and replace it with acausal trades and updateless decision-theoretical coordination, I would gladly bleed myself to death and spend a year learning latin to perform it.
As I write this, I feel a shroud of anger, real red-violent anger, not just pretend writing-anger fogging my mind and tears welling in my eyes. DOWN WITH BACKSCRATCHING! FUCK THOSE BASTARDS!
Any movement that is, indirectly or not about backscratching incurs an ideological discount in my mind to compensate for it. If the OSC came to my town, I would care no more no less about orphans than before (except for optimizing effectiveness of my current orphan-caring activities based on accurate OSC sensibilization and advice) and any more effort that's asked on pain of social ostracization can DIE IN A FIRE WITH ME.
FUCK FUCK FUCK FUCK
Before you finalise that ritual, please take some time to discuss the implementation details with the demon you're about to sell your soul to.
Is a family a backscratching club? If I am more willing to help my brother when he is in need than a random stranger, is that backscratching? Is that good or bad? Will that still be allowed after you perform your ritual?
How about friendship? One of my friends is looking for a new job, so I pull some strings at my employer to get him an interview -- he'll still have to pass the tests himself but he gets bumped to the head of the resume-filtering queue. We didn't *start* the friendship as a mutual aid pact, we started it because of similar interests and hobbies and stuff, but now that the friendship exists, being able to call in favors like that is one of the side benefits. (My employer benefits too, since although I can't say from direct experience how competent he is as an employee, I can at least testify that he seems smart, isn't lying about his current employment status, seems like a reasonably sane and reliable person in general, etc.)
Will your eldritch ritual only prevent me from helping my friend bypass the resume queue, or will it eliminate friendship completely? Next time I'm sitting at home bored and want to call a friend and ask if they want to go do something, is that still allowed or do I need to go through some app to give all available people in my area an equal chance to respond?
The unstated but obvious joke in Scott's description of orphan supporter clubs, is that they usually aren't started by some Machiavellian schemer who's deliberately trying to create a backscratcher club. Or at least that's not how they feel on the inside. They genuinely care about orphans and want to connect like-minded people together. Creating a social network of people who can then call in non-orphan-related favors with each other, is a nice side benefit. Will your ritual somehow eliminate the favor-calling opportunities while leaving the orphan support intact, or will it no longer be possible to start any kind of club or organisation or movement or any other structure of people which would inevitably have backscratching as a side effect? Will the net effect be more or fewer orphans getting helped than before?
Jorge Louis Borges wrote a short story called 'the Bribe'.
A young person at a North American university is competing for some position. He knows one of the older professors, whose vote can decide his future, considers him less qualified than his competitor. But he also realizes that the older professor is bound by North American puritan morals.
So he finds a way to insult the older professor publicly. Since the professor now knows that he has a personal motive to vote against the main character, so he votes in favor of him and he gets the position. To quote the story:
"... I realized, my dear Winthrop, that you are ruled by that curious American passion for impartiality. You wish above all else to be 'fairminded.' Precisely because you are from the North, you tried to understand and defend the South's cause. The moment I discovered that my trip to Wisconsin depended upon your recommendation to Rosenthal, I decided to take advantage of my little discovery. I realized that calling into question the methodology that you always use in your classes was the most effective way of winning your support."
Great points.
While I generally agree, it is good to remember that backscratching clubs often seem like "a solution to coordination problems" from the inside.
It could be that creating a backscratching club sometimes *is* the most reasonable solution to a coordination problem, in which case there's no reason for anyone to get too pissy about it when they notice one.
"I would gladly bleed myself to death and spend a year learning latin to perform it"
I just want to make sure you do these in the correct order.
somehow this was my main concern as well
>replace it with acausal trades and updateless decision-theoretical coordination
Correct me if I'm wrong, but isn't that just "backscratching but with decision-theoretic epicycles"?
One of the reasons that I dislike backscratching is that it arises organically from optimization processes with perverse incentives that spawn deviant children organizations.
I want to promote orphan-supporters insofar as it helps support orphans more than it detracts from other causes, which I remain aware of when deciding to promote OSC content based on how I expect the repercussions on society at large to be beneficial. This allows me to promote OSC's orphan-supporting content without promoting OSC's word-banning content.
Before the OSC appears, if people would coordinate using eldritch means like acausal trades, they wouldn't spawn word-banning movements as a byproduct. Of course, I want to discount how much I prefer acausal trades to backscratching as a coordination tool insofar as it has similar negative byproducts; I'm not currently aware of any significant ones (except overhead) nor do I expect there are.
The problem is that, by definition, there is no mechanism for acausal trade to get established. Causation in relevant because it’s how the world works.
I guess a follow-up question would be, what property of acausal trade as a coordination mechanism makes it not spawn word-banning movements?
My intuition here is that the word-banning here is not a natural byproduct of causal coordination, it's a byproduct of Heather being Heather. In the first place, if your goal is actually to help orphans, you can't really _coordinate_ with Heather per se, that word implies having common goals. The most you can do is get Heather to press "cooperate" in some specific situations dictated by decision theory. Which is what I meant by "backscratching with epicycles".
It's even explicit in the article that Orphanism was not optimized to be about orphans in the first place. It's not the word-banning that's a byproduct, it's actually the orphan stuff. Maybe if it worked via acausal trade, they wouldn't need the fig leaf of caring about orphans in the first place, and that's, I dunno, probably good?
(Sorry for the swearing; I generally dislike social signalling, but I didn't really know how to express that I feel strongly negatively about this topic, and stating it plainly doesn't quite convey the same message.)
I mean your reaction is a moral one: such a thing is bad and makes society unsafe. You seem strangely (to me) reluctant to cast it as a moral stance explicitly so instead you're railing against them at a personal level. But the reason you dislike them is that they're obviously horrible: they do not solve for everyone's mutual safety which is what institutions are supposed to do. Even self - serving institutions like corporations are expected, in modernity, to pay *some* attention to the safety of everyone else (that's we have a bunch of laws about them and why we shame the bad ones), and the clubs in this article seemingly do not, hence they are repugnant.
I think the whole language of “acausal trade” sprung up because some people who had convinced themselves of a Humean form of rationality that is purely decision theoretic and leaves no room for morality as a separate norm then re-discovered Kantian arguments, and wanted to cast them in the language of trade and rationality rather than morality.
I just wanted to chime in that I found your comment very insightful. I never made the connection that it was easy for me to accept acausal trade because I was already a believer in the categorical imperative. Thanks!
Ditto what Bistromathtician said.
To the extent I understand it, I believe the mechanism of acausal trade getting established is a reasonable empirical belief that one's mind-state is mirrored by one's trade partner. The idea being that if you can reliably assume that your counterparty will behave as you do (say, because you're both fresh out of the box serially produced AIs of the same model, or identical twins or something, or you just know the person really well and they think like you do), then defection *stops being the dominant solution* because you can assume that any preference for defection on your part will be symmetrically engaged in by your trade partner, and if only symmetrical outcomes are allowed -- i.e. cooperate-cooperate or defect-defect, then cooperate-cooperate is the dominant solution because all of a sudden you're playing stag hunt where the *individually welfare-maximizing choice* of both parties is to cooperate.
This differs from the Categorical Imperative in that the latter acknowledges that cooperate-defect is part of probability-space, it just says that you shouldn't do it as a normative matter.
I don’t think of acausal trade, at least the way it is normally talked about, as requiring such a degree of similarity that it makes asymmetric outcomes literally impossible. I think of it just as requiring that the parties are in some sense recognizing the similarity of their reasoning process in a way that puts the same rational requirements on their decision, such that asymmetric outcomes would necessarily be irrational, though not impossible. But this is basically Kant’s idea - all rational beings are under the same requirements, such that if they were all rational they would do the same thing, and thus doing something that wouldn’t be effective if universalized is irrational (though not impossible).
I appreciate the clarification re: the Categorical Imperative. However, while I confess that I do sometimes think this way and this it may be a decent account of virtue ethics, it seems self-evident that absent robust symmetry guarantees it’s not an account of *individual welfare maximization* and thus not of “rationality” at least as definitionally limited to such maximization (absent additional constraints like reputation effects or non-bounded iterated games) because it’s not robust to defection. Two-boxers really do get epsilon more cash. Defecting really does improve one’s individual position whether or not it’s reciprocal.
The virtue-theoretic / internal cognitive properties are part of why I think of “being a defector” as being something that one is rather than defection being something that one does, and thus there’s at least conceivably potential societal upside in, say, summarily executing everyone who uses a license plate cover to avoid payng tolls or other forms of traffic enforcement or doesn’t return their shopping cart. But that’s conceptually just extending the idea that ex post punishment is the correct way to minimize defection to a world in which defection is an incorrigible and inferable character trait, (basically trying to make group selection actually work). The fact that I’m habitually conformant to cooperative norms dorsn’t mean that I’m operating under a belief that I wouldn’t personally save a few minutes here or there if I ran red lights more often or routinely turned into the offramp to force my way into traffic at the last possible second.
Can I interest you in joining my non-backscratchers club? We non-coordinators *do* need to look out for each other, after all.
Just so! If the backscratchers coordinate and we do not we will always be at a disadvantage and the backscratchers will take over. We must work together in solidarity to promote fair competition and individual decision making!
Unfortunately (or not), these kinds of organizations do actually serve a purpose beyond just giving members some kind of unfair advantage. While it is omitted from Scott's description, in the real world there is usually some more or less extensive vetting process before you can join one of these societies. Not so much as in ages past, I think (the remaining societies are hard up for members), but the 'costs' are less of the 'do silly rituals' sort than they are 'you actually have to be self sufficient and convincingly honest and trustworthy'.
This allows the organization to function as a reputation service. In a world before credit scores, before you could Google someone's name or check out their LinkedIn, you could go somewhere and demonstrate that you were a member of this organization (to the local chapter, by way of secret handshake or what have you) and so demonstrate some level of trustworthiness. Which, depending on the reputation of the organization in question, might well extend even to outsiders - ah, he's one of the Knights of Columbus, well they're strange Catholics but at least he's not a con man or a serial killer...
I believe that most backscratching clubs of the "fraternal mutual aid society" sort, such as the modern Freemasons, prefer that you join when you don't need aid. And in practice, I have a feeling that your ability to draw aid in the future will depend a lot on whether you have been perceived as willing to give aid when asked.
I've seen this dynamic operate in completely non-backscratchy groups, too.
Yep. And these organizations need to do some checks, because one way they can fail is from the inside, the people who accept favors but don't help other club members.
Of course, one should also find a way to vilify every other backscratching club while keeping their own club protected.
The best way to vilify the potentially biggest backscratching club is to call it a lifeboat game that will lead to cannibalism. #NeverLifeboat.
Surely this is why Abrahamic religions have spread so far. "Ours is the one true God", "Other religions follow demons and false gods", "All other paths lead to damnation". And straight from the Quran "“whoever changes his religion, kill him”".
And as far as I can tell, this isn't a universal feature of religion - just the most successful ones.
It was useful unto necessity in some particularly competitive contexts, so there's lots of relevant instructions still in the scripture, but key portions have been commented out of leading builds ever since industrialization made winning a war unprofitable.
I'm not sure this is a good example, as Abrahamic religions have a strong (though certainly not universal) tendency to tolerate each other, even when "each other" is the dominant outgroup who they encounter. See the Muslim attitude towards "people of the book", or the Jewish attitude towards Noahides.
https://slatestarcodex.com/2014/09/30/i-can-tolerate-anything-except-the-outgroup/
But also the long-term successful idelogies that take over large parts of the world enact norms that limit the ability of new memetic competitors from challenging them. The dangers tend to come when you have shifts in information technology that render those protections less effective.
For instance, consider the transition from paganism to monotheism where you have a rule blocking other religions from getting off the ground. And the Catholic church was able to effectively combat alternative interpretations by controlling the written word until the printing press. And notice that even seemingly nice messages about not excluding people or good Samaritans also has the secondary effect of detering organization along different fault lines.
I wonder if that's not something we are seeing today as the mechanisms that existing organizations used to limit the ability of people to spontaneously organize in different ways has been reduced by new infotech.
I guess the most interesting thing that stood out to me is that this is a "barberpole model" a la https://slatestarcodex.com/2014/04/22/right-is-the-new-left/. Whenever a Backscratchers' Club gets too successful like Daniel's, its like it never even existed at all, so the Heathers of the world try to form a new Backscratcher's Club within the old Backscratcher's Club. But if they're too successful, they get "couped" in turn by *another* new Backscratcher's Club sprouting within them. It's an endless cycle, because victory doesn't actually solve anything -- just make the ground fertile for the next conflict.
And this applies just as well to ideologies as it does to Backscratcher's Clubs! In the past, the pioneering Christian evangelists like Saint Paul must have believed that, once everyone agreed on Christianity, surely there would be peace right? No more wars between pagans & Christians, just Christians being Christian to each other. All alike, with nothing to fight over. All supporting each other in the Backscratcher's Club of God/universal brotherhood.
But all that really happened was that the Christians started fighting each other over things within Christianity! Arianism, Nestorianism, Pelangiasm, Catharism, more than I can name; if the wars with Paganism seemed like a wildfire that would burn everything to ashes & eventually bring a sort of peace... then what actually happened was the ashes immediately turning into a new forest, that burned in a new wildfire, precisely *because* they were so alike that ideas could spread easily between them & create mass movements!
i.e. Something like Catharism would have made no sense to someone who's never even *heard* of Christianity -- but the moment everyone did, "Catharism" became something that was possible, something that could spread beyond the minds of a few religious scholars into the mass population & trigger an uprising. The monoculture was in fact *more* vulnerable to war, because it was more vulnerable to the explosive spread of ideas -- *any* ideas, not just the approved ones. The peace set the stage for war, *inherently by being the peace*. By connecting people together and making them one. All victory could ever do was set the stage for its own defeat.
And this keeps repeating throughout history: the periodic rebellions & explosions of Imperial China would be very interesting to try out through this lens, but a more familar example for most readers will probably be the effects of the printing press on European history. It printed the Gutenberg Bible, which perhaps led rather more directly to the Protestant Heresy/Protestant Reformation then the Church could have foreseen at the time. For the more ideas could spread -- and the more they were used to spread a monoculture where, if 1 "domino" starts to fall, all the rest will "catch" him -- the more peaceful it paradoxically looks: the field of dominos getting closer & closer & *closer* together. Right up until it all explodes. If everything linked together, none can flip without all flipping; all the rest will catch him.
But if one does come down with something & flip, all must flip with him: all the rest will catch *it*.
The applicability to the modern day & our connected world, social media & modern movements, is best left as an exercise to the reader.
/It was a glorious peace while it lasted, the 90s post-Cold War era. Too bad that it couldn't work; a Backscracther's Club that works for everyone is a Backscratcher's Club that works for no one. All it *can* do is connect people together & make it easier for them to talk about the need for a new Backscratcher's Club -- precisely because the old club brought everyone together. Precisely because it had won.
It's also common to see early Christianity as a backscratcher's club - its early spread was in urban contexts among people with a weak social support network, who could now enter into a club for mutual support and a feeling of moral superiority to outsiders.
I cannot resist adding this, from http://forum.ship-of-fools.com/cgi-bin/ultimatebb.cgi?ubb=print_topic;f=61;t=000011
I was walking across a bridge one day, and I saw a man standing on the edge, about to jump. I ran over and said: "Stop. Don't do it."
"Why shouldn't I?" he asked.
"Well, there's so much to live for!"
"Like what?"
"Are you religious?"
He said, "Yes."
I said, "Me too. Are you Christian or Buddhist?"
"Christian."
"Me too. Are you Catholic or Protestant?"
"Protestant."
"Me too. Are you Episcopalian or Baptist?"
"Baptist."
"Wow. Me too. Are you Baptist Church of God or Baptist Church of the Lord?"
"Baptist Church of God."
"Me too. Are you original Baptist Church of God, or are you Reformed Baptist Church of God?"
"Reformed Baptist Church of God."
"Me too. Are you Reformed Baptist Church of God, Reformation of 1879, or Reformed Baptist Church of God, Reformation of 1915?"
He said: "Reformed Baptist Church of God, Reformation of 1915."
I said: "Die, heretic scum," and pushed him off.
Original version by Emo Philips: https://www.youtube.com/watch?v=D88DQIJmxRI
That's good and adds some things, but it's missing what I consider to be my favorite part:
"Are you Baptist Church of God or Baptist Church of the Lord?"
There's a Marxism version of this as well. :-)
"A heretic is someone who shares *almost* all of your beliefs. Kill him." -Steve Jackson
> In the past, the pioneering Christian evangelists like Saint Paul must have believed that, once everyone agreed on Christianity, surely there would be peace right? No more wars between pagans & Christians, just Christians being Christian to each other.
Small-r republicanism was also assumed to lead to peace (between republics) early in the transition away from monarchy. https://en.wikipedia.org/wiki/Democratic_peace_theory#History
Democratic peace also shows that this sort of peace mechanism can, in fact, work; not as a way to get a benefit at the expense of others, but as a way to preserve peace yes. The democratic peace theory (that democracies don't go to war with each other) holds up pretty well. And I don't think it's just because democracies have common rivals in dictatorships; from the 90s to the 00s democracies lacked rivals they considered a serious threat, but that didn't make them want to go to war with each other.
There's a trend of fewer interstate wars since WW2, which doesn't appear to be quite reducible to nations not going to war with each other because they're both democracies. North Korea is as dictatorial as it gets, but nobody expects it to attempt to invade the South again. China & Vietnam are both one-party Communist states, and the former briefly invaded the latter in 1979, but they've also been able to have peace since then.
There are many dictatorships that don't go to war of course, but there is a remarkable scarcity of democracies that do go to war with each other. The American Civil War is the only serious example I know of, and even there I'm not sure how democratic the South was, even if we only consider whites. Plus technical examples like Finland and the UK being at war in WWII because they were allied to dictatorships on opposite sides, but they didn't really fight one another.
Israel went to war mainly with the PLO and later the Hezbollah in Lebanon, which the Lebanese government couldn't control; Lebanese participation in the 1948 war was minimal. Germany wasn't a democracy, the emperor controlled foreign policy. At the time of the siege of Syracuse (is that the one you refer to?), Syracuse was ruled by a tyrant. Anyway, as I said in another comment, I'm not saying there have been no wars between democracies, but there have been few compared to wars involving autocracies, even in the recent era when base rates don't explain it.
Also the "War of 1812", although that was kind of a sideshow to Napoleon.
https://youtu.be/o7jlFZhprU4?si=LEjEj_pcX3GLjg2i
How democratic was the British Empire in 1812? Did the people most likely to die in or have their property destroyed by the war have a say?
I love that song, thank you to the Arrogant Worms.
"The White House burned, burned, burned,
And we're the ones that did it.
It burned, burned, burned,
while the President ran and cried."
There was the time the UK rather politely invaded Iceland, but that may bolster your argument less than refute it.
I got into an argument about how democratic the south was not too long ago. https://www.richardhanania.com/p/taking-the-medium-term-past-seriously/comment/56319184 There's some kind of myth that it wasn't compared to the north, but it actually had comparable shares of the population voting. https://web.archive.org/web/20201111211244/http://economics.yale.edu/sites/default/files/files/Workshops-Seminars/Economic-History/sokoloff-050406.pdf
It was the first World War that convinced men like Whitaker Chambers that socialism (rather than democracy) was necessary to prevent war. https://entitledtoanopinion.wordpress.com/2024/05/10/witness/ The Kaiser's Germany had elections, and the Social Democrats even obtained a majority, despite how much Bismarck would have hated that. It wasn't sufficient to prevent them from warring with autocracies like Russia, constitutional monarchies like the UK, and republics like France & the US.
Specifically, I think it was that national forms of socialism (trade unions, not actually a reference to the Nazis) had failed to prevent war, but international forms of socialism (communism) might work, right?
Wikipedia has a handy list. It's rather large.
https://en.wikipedia.org/wiki/List_of_wars_between_democracies
The Democratic Peace Theory can be salvaged by limiting it to *liberal* democracies and defining "liberal" restrictively. The most restrictive version is to require >50% of the adult population to be eligible to vote, which is a stronger requirement than it sounds at first because women are generally a little bit over 50% of the adult population and most large democracies only adopted Women's Suffrage during the interwar years.
Further epicycles can be added by excluding "young" democracies that have had democratic constitutions and elections for less than three-ish years, which defines away a bunch of the remaining counterexamples. But then we're left with "No wars between democracies in the past century or so, except for a bunch of wars that happened in the wake of decolonization or the fall of the Soviet Union, and don't pay too much attention to Latin America", which is a much weaker claim than "Democracies don't go to war against each other" is generally understood to mean.
I wouldn't require women's suffrage, but I would require it to be actually a democracy at least among free men: that (almost) all (non-slave where applicable) adult men can vote, elections are actually free, and elected officials actually control most policy, including foreign policy and the military. I wouldn't exclude young democracies, but I'd require that at least one free general election has actually been held; a secessionist militia that promises to establish a democracy, or perhaps holds an unfree election under a state of war, doesn't count. Under these conditions, not many count, and even fewer actually had serious fighting between democracies. (Compare the more recent lists at https://en.wikipedia.org/wiki/Template:Lists_of_wars_by_date )
Anyway, I'm not saying that war between democracies absolutely never happens, but that it's much rarer than either war between autocracies, or between democracies and autocracies. Democracy is among the most reliable things that prevent war. Other ones ones are alliance systems against external enemies, especially if policed by a superpower (but these risk a major war between the alliances, and often have localized but serious proxy wars), and nuclear weapons (but the risks of giving every country nukes to prevent war are obvious).
I'd actually bring Latin America as a positive example: its countries aren't part of tight alliance systems like most Northern democracies are (perhaps because they're too far from potential external threats to be too worried about them), yet it hasn't had a serious war since the fall of the Cold War military dictatorships.
(I don't like the term "liberal democracy" because it's ambiguous between norms that were liberal in the 19th century but really are conditions of actually being a democracy at all—not persecuting opposition politicians, sufficient freedom of speech that they have a chance to convince voters—, and democracies whose policies are liberal in a modern sense.)
The first sovereign democracy of the modern era was the United States of America, in 1776(*). The second was the French Republic in 1789. In 1798, the United States and France went to war. A limited war, fortunately, and one both sides would rather forget, but still.
* The Icelandic althing goes back to the 10th century, but was under the sovereignty of the King of Norway and/or Denmark until 1903; there were a few other proto-democracies under similar circumstances.
Well, except for that one time in 1995. https://www.youtube.com/watch?v=pbfBzWJVbX4
"Like maple syrup, Canada's evil oozes downward."
Interesting thought. Maybe our filter bubbles are actually good for something.
There is a stable solution though when two separate backscratchers clubs get established and are comparable in strength. This is how American political parties have worked for most of their history, though in recent decades they started taking on more substantive ideological differences, so they stop being as pure backscratchers clubs. (Van Buren and Jackson had almost no substantive policy interests in common - they just realized their orthogonal interests could both be achieved if they coordinated together.)
Agreed. I think this is why countries seem to naturally slide towards being Lebannon, Brazil, Argentina, Venezuela, Malaysia, and so on -- naturally slide towards being split between different backscratchers clubs. What social scientists call "Clientelism", "Patronage Politics", or just "Corruption". Why is corruption so endemic, so hard to fight? Because you can't have a Backscratchers' Club that works for everyone, at least not in the long term. In the long term, everything slides towards splitting up into multiple Backscratchers' Clubs, each trying to benefit themselves by stomping on the others, in a long-lasting stable equilibrium. Sometimes this looks like the historical US. Mostly, it looks like Lebannon or Malaysia.
(The process of *becoming* Lebannon or Malaysia looks something like the recent FAA hiring scandal [https://www.tracingwoodgrains.com/p/the-faas-hiring-scandal-a-quick-overview], where you invent a test with arbitrary answers, hand out the answers to your Backscratchers' Club, and act shocked when only your club aces the "Your lowest grade in high school should be Science, but your lowest grade in college should be History, and your most common grade in high school can be anything but a C, which is *verboten* for some reason." question set. Damn, they must really know their stuff huh?
A similar thing was key to Erdoğan taking power in Turkey [https://www.astralcodexten.com/p/book-review-the-new-sultan], incidentally; once Gulenists got into power at the Ministry of Education, they could steal or just write the answer sheet to the national examinations, then pass the answer sheet to their own students so the civil service would be increasingly made up of top-scoring Gulenists.)
(You could also bring in some Selectorate Theory [https://en.wikipedia.org/wiki/Selectorate_theory] and note that this helps explain why it was so hard for the US to build democracy in Iraq & Afghanistan. A democracy is, in Selectorate Theory, just a system where the "winning coalition" is really, really big. So big they can't be satisfied by just handing them money, it's more cost-efficient to actually do a good job of running the country & invest the money into things like public works, universal education, and other public goods that benefit everyone and exclude no one.
If the "winning coalition" is small instead though, none of that happens because it's more cost-efficient to just hand your supporters money directly, rather than incidentally benefitting them by building a new road system that works for everyone or whatever. And, as we have just learned, a Backscratchers' Club that works for everyone works for no one. That means existing "We work for everyone!" Backscratcher's Clubs tend to get supplanted from within by new "We work for *us*." clubs -- but it also means a new "We work for everyone!" club is hard to get off the ground, it can't compete with the established "We work for *us*." clubs. Even when the new club is being aggressively promoted by the US and called "Democracy". The dynamics are no different.
i.e. Democracy is not just hard to maintain, but hard to establish. It just doesn't have to beat Clientelism/Corruption when it's old & decaying, but new & trying to get off the ground. The temptation of "But what if we only benefitted ourselves?" is always there, waiting.)
(If you want to bring in more social science, it could be worth looking at Public Choice Theory & the example of things like farmer's subsidies. The famed paradoxical result that the farmers *gain* political power as they *lose* membership in a democracy -- contrary to what you'd expect where the big groups have power & the little groups don't -- precisely *because* they're small. Precisely because if there are only a few farmers, and they take a lot of money per person, that's still only a small amount of money lost per taxpayer, so no taxpayer opposes them while 100% of farmers support it -- so the farmers actually have more power when they're small. When they naturally form a Heather-like Backscratchers' Club rather than a Daniel-like club. In countries like India where the farmers make up a majority, by & large the farmers are actually taxed, not subsidized, simply because they're on the losing end of the Backscratchers' Club wars. Simply because they're large, and a club for them would be a club for everyone.)
Tangential thought: how do you keep together a backscratchers' club that includes everyone? Simple: *Don't tell them that it includes everyone*. You tell them that it excludes someone -- no, a lot of people, the more people the better. The more people you point to as being excluded, the more exclusive & desirable the club becomes, so you better find or invent a lot of people to exclude. If people can only be united in hate, well then invent someone to hate! If people can only be united by an enemy, well then invent an enemy! If the club isn't a persecuted minority, well then just pretend that it is!
I think a lot of historical examples of the persecution of minorities, ranging from ancient pogroms to historical Fascism, are an example of doing that but for real (instead of inventing people to exclude). Even things like, say, picking fights with other countries just to stir up nationalism in your own popuation & unite them together, could be an example. And no doubt there are plenty of modern examples, both of picking fights with a real someone & inventing someone to be scared of & exclude. I'm sure you can think of a few. (Perhaps this is why it was necessary to invent the concept of Satan, for example, to go historical again...)
i.e. So perhaps what the 90s dreams of unity & the UN really needed, was to pretend they *weren't* tolerant & inclusive. No, maybe it needed to pretend it was still excluding the aliens of Tau Ceti or something, like the Star Trek dream but inverted. If people cannot accept a genuinely inclusive world, simply tell them it isn't. Indeed, the more inclusive it becomes, the more exclusive & discriminatory you'll have to make it look (lest it *be*).
Another tangential thought: it's common for these self-promoting "Logrolling"/Backscratcher's Clubs to be made up of mediocrities (who would never rise very far on their own), or have hazing rituals where you submit blackmail material to the Club, or both -- because a club made up of successful people is good, but a club made up of successful people *who would be nothing without the club* is even better. The Varangian Guard/Machiavelli point: you want people who are loyal, because they have nowhere else to go. Because they would lose *everything* if you lost power -- and they know it.
Consider for example all those dictatorships that go around promoting incompetent sycophants who pose no threat to the dictator, and firing/executing anyone who proves themselves competent. Because competence means you could be something without the dictator; it means you have *options*. You could be a threat. The more incompetent you are meanwhile, the more completely & utterly dependent you are on the dictator for your exalted position. Again, Machiavelli in action.
Or look at the various Royal Guards, Presidential Guards, Republican Guards, and so on those dictatorships cultivate. Like the Varangian Guard, the most important thing is to not build the Praetorian Guard; you want people who are completely & utterly dependent on you, who are most emphatically *not* able to go outside you & plan your demise like the Praetorians did, who would be nothing without you. The more hated & foreign the minority, the more sense it makes to exalt them as your guards; it's just business really.
As Scott quoted Krugman as putting it at https://slatestarcodex.com/2015/07/28/non-dual-awareness/ ,“Being incompetent and displaying it, conveys the message I will not run away, for I have no strong legs to run anywhere else...". It's why, for example, people suggested making an all women army for the Afghan National Army, and generally putting women in charge in the Afghan government -- you want people who would lose *everything* if the Taliban took over. People who will be loyal no matter what, because the alternative is death -- a slow & painful death at Taliban hands. It'd be a complete & utter inversion of Afghanistan's native norms & traditions? It'd piss a lot of people off? Exactly the point. Exactly why it'd work. At least you wouldn't have Ashraf Ghani fleeing the country with over a hundred million in cash -- enough to set up a cushy life for himself after the fall of the country.
> If people can only be united in hate, well then invent someone to hate!
Oh, right, like 1984 and Emmanuel Goldstein!
I appreciate your allegories. They are some of the most thoughtful that I've found.
Agreed. Just imagine if he co-wrote with Robin Hanson. After the amazing allegory would be some counterintuitive solution for backscratching, possibly involving prediction markets.
The initial lifeboat involves a negative consequence: being killed/killing someone else. How would it work with a positive consequence?
10 people meet up because they all heard about a bird-watcher meet-up at a certain location. While watching for birds, they all simultaneously see an envelope fly out of the sky. Inside is a lottery ticket worth $5 million (must be a scratch-off). Only one person can redeem the ticket, and of course that person will also be responsible for income taxes and such. Who gets the ticket?
Someone initially screaming their name before lots can be drawn won't work, and maybe would get their name removed from drawing lots.
Maybe they would decide to split the thing 10 ways, after all distributions are made. That isn't straightforward, either, not even counting that the person selected may be dishonest and have a "change of heart" when given possession of the proceeds.
Maybe the person with the smallest income should get it, since that would result in the smallest of income taxes, maximizing the amount to be divided.
Hand in hand, everyone marches down to the nearest lawyer's office to draw up a contract for the fair division of proceeds.
The problem in this thought experiment is that division of the windfall is legitimate, so it can occur in a low-context matter that exploits the existing framework of law and contract. Both the cannibalism and back-scratching clubs are informal or clandestine arrangements.
Perhaps I could change it, then, to a family of 10 siblings, whose only remaining parent has passed away, leaving a priceless artifact which they all want to possess, not sell, as it represents the family legacy in some irreplaceable way. Only one person can possess it.
Nothing is more destructive than inheritance battles, between family members, over wills and it's even worse where there are no wills. "Sally took Granny's rings that I wanted and Granny promised them to me but Sally says she is the eldest daughter so they should come to her". Absolutely tearing each other apart over small sums of money and rubbish items.
I've heard the claim that it is not that rare for more money to be spent in total legal fees in these disputes than the actual inheritance. Three people each willing to spend up to $X-1 to win the $X inheritance, and sometimes willing to spend $X+T because it's worth another $T to them to make sure those other bastards don't get a dime.
"Bleak House" is a famous literary "example".
https://en.wikipedia.org/wiki/Bleak_House
The spending ratchets up not just due to spite ('make sure those other bastards don't get a dime') but also from sunk cost issues, like in a dollar auction (https://en.wikipedia.org/wiki/Dollar_auction). Once you've spent the money on a lawyer the money is spent and gone, so considering the next legal spending should just be about how useful it is on the margin - it makes sense to spend an additional $5k to get $10k of inheritance no matter how much you've already spent.
There is an auction game of, say, auctioning off a $20 bill, where only the top bidder wins, but the top two bidders both have to pay their bid. The idea is that, even if the top bid is $20 and you have bid $15, you will lose less money if you now bid $21. And then the same principle applies to the now 2nd highest bidder.
The only winning move is not to play. Or, perhaps, to make a separate agreement with the 2nd highest bidder, to both split the proceeds.
Dickens satirised it in the case of Jarndyce versus Jarndyce where an inheritance dispute had been going on for generations and eventually only stopped when the entire estate had been eaten up by legal fees so there was nothing left to inherit, but unhappily it went on just long enough to destroy many lives.
I've personally seen incredible bitterness over "so-and-so got that and it should have been mine" in disputes, and the thing in question was not worth wrecking family relationships over. People go crazy when a will is involved, even if there isn't a hefty sum of money or property to inherit. I have to agree that the spirit of "I'm not going to let those bastards win" takes over and quashes any good sense.
It ends with everybody in traction, and a banana peel on the floor.
A mad movie reference?
Of course!
(Some people can think, others can only joke.)
Whoever sees the coolest bird first, obviously, else this bird-watching meetup isn't worth their binoculars. Cue arguing over which is the coolest bird.
Ice Phoenix!
It's essential to the lifeboat version that the outcome is bad for one person and (relatively) good for the other 9, not the other way around. A version with positive consequences would have to have a positive consequence for all but one person, not just once. Let's say there are 9 remaining spots for a last amusement ride before the park closes. (Except the reward there is too low to engage in unfair scheming and jeopardize friendships.)
That version is still a negative version, just a less-than-what-you-would-get-otherwise version. Assuming all strangers, all of them expect to ride once more, but one of them will have expectations dashed. And in that situation, I think the race belongs to the swift, like in musical chairs.
But what about 10 people with but a single seat on the ride remaining? I think each person would try to paint themselves as most deserving: I came from the furthest away to ride, or I've wanted to ride this for 26 years, or it's my dying wish to ride this one last time, etc.
If by "positive consequence" we mean it's (relatively) good for one and (compared to that) bad for the others, that works completely differently from Scott's version, as everyone has only one preferred outcome, while in the original version one has multiple preferred outcomes, with no preference between them. In your version there is no motivation for everyone to agree on any Schelling point that's not him, so tricks like what Albert tried won't work.
It's the fraction of people who can get the better outcome that matters, not whether you you consider getting an extra ride in to be a positive or negative (how you view it is arbitrary and relative to your expectation). The equivalent of your example with the lottery ticket is would be if the lifeboat could only support one person, and they have to choose 9 people to die instead of just one. Then you can no longer get majority support (from rational self-interested actors) just by choosing a name. If the lifeboat can hold at least 6 of 10 people, you can form an ad-hoc coalition and overpower the other 4 people.
There's a British game show called "Golden Balls", where the final round has the last two contestants doing a Prisoner's Dilemma over the accumulated pot from previous rounds. After discussing, rach privately selects whether they try to "split" or "steal" the pot. If both split, the pot is divided 50/50. If both steal, both go home empty-handed. And if one splits while the other steals, the stealer keeps the whole pot.
Usually, the contestants spend the discussion phase swearing up and down that they'll split and urging the other to do the same, and then about half the time they turn out to be lying. But there's one notorious episode where one of the contestants opened the discussion by promising he would steal, but if the other contestant chose "split", then he would pay the other contestant half the pot afterwards.
Here's how it played out:
https://youtu.be/S0qjK3TWZE8
Short version: contestant 2 threatened to steal if contestant 1 didn't promise to back down, but backed down himself and put in "split" at the last minute. Turns out contestant 1, despite his claims, also put in "split".
That was fascinating, and I would never have thought about it. But in hindsight, it makes sense, from a Monty Haul standpoint. In Monty Haul, if you switch doors, you KNOW the door they revealed isn't the door you want, so you're switching from a 1/3 chance to 1/2 chance to get the right door.
For this, his opponent has two possibilities for him to think: he will take split, or take steal. But he changed it to more possibilities: 1) he says he will take steal but promises to split the money later, 2) he says he will take steal but will NOT split the money later (he's lying), he will take split even though he says he will take steal. For his opponent, then, two of the three possibilities lead to a split, and one leads to him getting all the money.
The odds aren't as clear-cut as in Monty Haul, as one must still decide whether the opponent is honest, but it kind of looks like a 2/3 chance of winning this way.
This reminds me of a deal offered in a webcomic, where one side had two enemy sides who were both after its treasury. They signed an unbreakable deal with both sides, paying them 2/3 of the treasury, contingent on the OTHER side attacking first. This prevented BOTH sides attacking, until they realized they could sign another unbreakable deal with each other, splitting any treasury proceeds 50/50 no matter who attacked first. In this Golden Balls scenario, of course, after-game deals weren't necessarily binding.
Was that Erfworld?
It was indeed! It was Overlord Firebaugh, Wanda's father, who conceived of the plan against Frenemy and Quisling.
Huh, I'd thought it sounded more like Parson or Charlie...
The Wanda backstory section was so *dark* it didn't stick in my memory very well.
The best part of that scenario is that Ibrahim said in an interview later that he'd originally been planning to steal.
This writing is amongst some of your best work, Scott. (I realize this comment doesn't contribute to the overall discussion, but I'm risking being spammy with a comment just to emphasize how great it is.)
Except...
I just went to share it on Facebook and then hesitated when FB previewed the terrible AI art accompanying the post. Seriously, it's so bad, I'm actually embarrassed to be recommending something associated with it.
I settled for putting a disclaimer on my recommendation of the link to ignore the AI art, but...man. No art would be so much better than the AI pictures, seriously.
Here, I made a lovely alternative AI image: https://imgur.com/a/x9g6Rwp
TRUE LOL
So for the last year or so I've been describing the pervasive sense of plasticy homogenized sameness in Scott's AI images and everywhere else as "Lisa Frank." It doesn't matter what the subject of the art is, it always looks like it came out of the Lisa Frank Studio to be emblazoned on an 8 year old girl's trapper keeper in 1989. Scott's AI art for this article looks like the Lisa Frank Studio spun off a goth division or something.
Is it possible all generative AI started with Lisa Frank? That seems like the only explanation.
So you can just imagine my reaction when I clicked through. It might be the first AI image I've ever unironically enjoyed.
Thank you!
The unironic answer to your at-least-half-joking answer is that generative image AIs work by taking an image classification AI and running it backwards so it generates the image that it would most strongly map to the prompt you gave it. Google's 2015 DeepDream demo was the first high-profile implementation, which was pretty bare-bones and worked by perturbing a starting image to make it classify more strongly to the prompt.
As it happens, this makes deeply surreal images that look like Lisa Frank took too much LSD. I'm guessing the "Lisa Frank" part is because her art style is surreal in color but with very bold and striking subjects in the images. And the LSD part because the process of running a classifier backwards is fairly closely analogous to how drug-induced hallucinations in humans seem to work. I wouldn't be surprised if Frank's art style was influenced by 70s-era psychedelic art, since she came of age in the 70s and started doing her iconic art right around then.
The current crop of generative AIs have a bunch of extra stuff added in designed to constrain the output to be less psychedelic and more suitable to substitute for stock photos. Plasticy sameness is what you get when you run the constraints too strongly and Lisa Frank knockoffs are what you get with relatively loose constraints.
> it always looks like it came out of the Lisa Frank Studio to be emblazoned on an 8 year old girl's trapper keeper in 1989
Wow, you nailed it!
Is the objection the quality of the art or that it was made by AI? I'll admit to being fairly tasteless as far as art goes, but quality-wise it seems fine. Not a great work of art, but serviceable.
As for being AI made... I don't really get the objection here. I know people are worried about AI art displacing human artists, but I don't think there's any world in which Scott would have paid a human artist to commission an image to go with this blogpost: this is the sort of thing where AI has enabled art in a place there probably just wouldn't have been (properly paid and compensated) art before.
It doesn't seem worthy of a disclaimer... or is posting "disclaimers" about AI art a qualification of being in the "anti-AI-art backscratchers" club?
100% objecting to the quality of the art. See my comment above to @Eremolalos.
As for the disclaimer, I was actually switching horses midstream there. I took a break writing the comment to post the link to Facebook and was then confronted with goth Lisa Frank.
For the record, I'm totally fine with AI taking artists' jobs (if it does it well). But realistically, there are ways in which paid-for human art would likely be here if it weren't for AI art. Scott (or Substack) might pay for a stock image subscription service that would then go pay individual artists for their work.
Why are people so split on AI art? What is it that you don't like about it?
It has that common AI style: digital painting style, dramatic lighting, dark ArtStation type of feel. If you zoom in, the details are often nonsensical. But overall, it's a nice looking image.
Please see my comment to @Eremolalos above.
It's not a nice-looking image unless you like Lisa Frank.
It looks nothing like Lisa Frank to me. It's nearly monochromatic apart from the life jackets. Dark and moody.
I would describe it as the type of images you would get if you did an image search for "striking digital art" or "stunning digital art", back before AI art existed. It always tries to turn up that "wow" factor to an 11.
Stable Diffusion does other styles no problem if you want it to. I don't think it's inherent to AI (unlike the nonsensical details, which all AIs struggle with). Midjourney seems to have purposely tuned their model to have more of an ArtStation vibe by default. I'm pretty sure this style is driven by human preferences. It's just strange how divisive it is. People love it or hate it.
I once read an interesting book somewhat about this problem:
https://www.amazon.co.uk/Dictators-Handbook-Behavior-Almost-Politics/dp/1541701364/ref=sr_1_1?sr=8-1
In order for a party or movement to get to power it has to have a great coalition of backscratchers, but once you take power you want to have a narrow coalition so each person can get a larger slice of the pie. The solution is to have a great purge.
The truth behind how Ponzi schemes work?
Everybody expects they will be the one to survive the purge
Are you sure you're not thinking of pyramid schemes?
They're similar, but the premise is different. Ponzi schemes are actually pyramid schemes, but the con men tell people they make money by some investment strategy, but any returns actually end up coming from new investors.
Well, I know I'm being pedantic here, but Ponzi schemes are not pyramid schemes. They're similar, but the former is not a subset of the latter. It's more like Ponzis and pyramids are subsets of the set "frauds where any returns are coming from new investors."
In a Ponzi scheme, the schemer acts as the hub of the scheme; all new investors interact directly with the schemer and make their payments to him or his employees, even if they were recruited by earlier investors*. The investors invest with the schemer in the first place because, as you say, they've been told a lie about the yield coming from postal reply coupon arbitrage or complicated options straddle trades or something. There's no pyramid; investors don't have to go out and find new investors.
Anyway, the reason I brought it up is precisely because in a pyramid scheme, the participants are perfectly well aware that the yield is coming from new investors, and they naturally want to be as close to the top of the pyramid as possible.
That is an extremely important book, and though it says that, it says a lot more as well. Esp re why some types of govts provide public goods and some don't.
And if you liked that book, I recommend https://www.cornellpress.cornell.edu/book/9780801479823/dictators-at-war-and-peace/#bookTabs=1
"The future is littered with prizes
But though I'm the main addressee
The point that I must emphasize is
YOU WON'T GET A SNIFF WITHOUT ME!"
-- Scar, "Be Prepared"
Extremely minor typo: “expensive” should be “expense”
It is not true that Daniel's backscratcher club is meaningless. It has the effect that the local newspapers no longer print anything negative about anyone local, which presumably was not previously the case. (If everyone including non-locals joins, then the newspapers only print puff pieces.)
Thinking about this a bit more, it seems like some people benefit more than others from Daniel's club. For example, people who are better at parsing subtle, implicit social signals might do better, relative to others anyway, in an environment where most explicit negative communication has been banned.
You could tell a story that this is how we get social norms around politeness - they are imposed as a backscratcher's club designed by people who are good at communicating with subtlety. Not sure I believe it though
This also explains why it can be economically rational to remain a racist in a racist society even though classical economic theory says it shouldn't be. The cost of being excluded from the white Backscratcher's Club following this kind of defection can be higher than the gains from open and rational business activity.
Perfect competition doesn't predict that profits get competed down to zero anyway (in the accounting profit sense, where the cost of capital, that of the investors deferring consumption and taking risk, is included in the profit, as opposed to the economic profit sense).
Economic theory also doesn't actually predict that no company will discriminate, or that all companies that discriminate go bankrupt. What it does predict is that, if the market is competitive, and race doesn't provide information about the quality of workers beyond what can be easily discerned from other factors, then equally good workers of different races will have approximately equally good job opportunities. If, say, 13% of the population is black, and half the employers and/or workers are irrationally racist against them, and the rest of the population doesn't care about race, then on a perfect market, half the companies will employ whites only, and at the other half 26% of the employees will be black on average, and everyone will make a similar salary at similar jobs. On a perfect market, black people only start to suffer if, in some profession, a bigger percentage of employers (weighted by the number of positions) are racist than the percentage of non-black workers in that profession. Of course on an imperfect market, problems start sooner because of the non-perfect fungibility of employers and workers.
This does predict that if a generation doesn't experience discrimination as young people, then equally performing members will have equal shots at achieving high positions once they're old and experienced. And perfect competition does predict that young people won't experience discrimination (again assuming race doesn't provide information about the quality of workers beyond what can be easily discerned from other factors).
I wouldn't say the perfect competition model assumes worker quality is solely a fixed inherent quality like IQ; it doesn't make assumptions like that. I'd rather say that when we say the perfect competition model predicts that workers of equal quality will have equal opportunities, quality is to be understood as current quality, however that came about, even if it's affected by earlier discrimination. But if a generation participates on a labor market with perfect competition throughout their lives, then worker quality in this sense won't be affected by discrimination.
David Bernstein made the best argument I'm aware of for the 1964 CRA, based on the idea that uncoordinated defection from the equilibrium would result in extra-legal violence :
https://www.cato-unbound.org/2010/06/16/david-e-bernstein/context-matters-better-libertarian-approach-antidiscrimination-law/
Violence is obviously particularly bad, but merely being excluded from the club that includes most people can be bad for you economically.
I have trouble taking anything serious, that comes from cato-institute.
Cato Unbound hosts debates between people, and usually those people are not employees of Cato. Bernstein is a libertarian, and he's arguing against other libertarians who criticize the CRA.
This is a good place to discuss if Backscratching Clubs even make sense as a way to benefit their members in the first place.
The example that's most relevant here was: "If a member runs a company, they should preferentially hire other members for good positions". But if you own a company, and you hire based on club membership rather than merit and salary demand, the club member you hire benefits, but you lose as much as he gains, possibly more! (Proof: if you didn't hire based on club membership, maybe he could offer to do the job for sufficiently less money that you'd still hire him. If he'd still be willing to do the job for that money, he gains as much as you lose by paying him more. If he could get a better job at that point, he gains less than you lose. And if you wouldn't employ him even if he'd do it for free, as a non-member is so much more competent, that means you lose more than the entire salary you pay him through your nepotism.) The club, as whole, doesn't benefit.
It's another matter if you're a hiring manager of the company, but not the owner, so you benefit a fellow member, but you aren't paying with your own money, and the shareholders aren't members. So in economic matters it seems to me that there is a point in a club only if either
- it's a small club, so they have e.g. a management position here and there, but they don't also make up most of the shareholders. I.e. they have an agent, but aren't the principal. It works better if membership or their behavior isn't well-known, otherwise people don't put them in positions of trust.
- or the situation is asymmetric, members tend to be on one side of a trade (e.g. they make and sell some product), and the other side of the trade are non-members, the members can collude to (say) raise the price of their product, and non-members can't easily enter the market (perhaps the members lobby to have them barred).
----
About racial Backscratchers' Clubs: Assume almost all managers and investors are white, while whites are a smaller majority of workers. A club where white-led companies are expected to prefer white employees is in the interest of white workers, but against the interest of white businessmen (managers and owners). Once the club contains all whites, it may be risky for a white business to defect from it alone; but rich whites have no reason to join in the first place, and they have an interest to coordinate to leave the club, at which point white workers have no choice but work for non-members. And even if one or a few white business defect on their own, white workers have little individual incentive to shun them, as workers or customers, even if they have a collective interest in maintaining the club. The club would only be stable if the workers are somehow able to coordinate, but the businesses aren't, which is unlikely. In a free market that is: white workers have an interest in voting for governments that enforce segregation, what I'm saying is that it's sufficient to stop enforcing it.
Regarding the first paragraph, see my answer to your other comment: https://www.astralcodexten.com/p/lifeboat-games-and-backscratchers/comment/61745709
Regarding the second paragraph, this doesn't explain why white managers or investors would want to keep whites in charge. People are mostly capable of scratching the backs (beyond what's explicitly expected of them) of those below them, but not of those above them. So if there's a white backscratchers' club, whites on the lower rungs have an incentive to want white bosses, and for the club to continue to exist (such that their white bosses scratch their backs), but they have little influence on that; while those at the top have no incentive to perpetuate the club by either scratching the backs of the whites directly below them, or by incentivizing those to scratch the backs of whites further below.
It likely won’t be a zero-sum game, though - you can use such corruption to extract from outsiders. If you’re a manager but not an owner, it’s not your loss if the company doesn’t benefit. If you’re a politician or public servant, you can squander public funds. If you’re a cop, it’s not your loss if you let other members go free and protect other cops in the society against accusations of misdeeds, Freemason-style (or just regular Blue Wall of Silence).
The costs can be spread around.
Yes, in some cases it can work. My point is, an entire race likely doesn't make a good club, as it's too broad: some significant economic or social classes within the race won't have an interest in participating.
>But if you own a company, and you hire based on club membership rather than merit and salary demand, the club member you hire benefits, but you lose as much as he gains, possibly more!
That's why dictators hiring incompetents isn't a stable equilibrium. If so me other country hires meritocratically, they can grow their pie -- increase their industrial.and.military capacity -- while you stagnate, and then they are a threat to you.
Indeed. Punishment of non-punishers can maintain even a bad equilibrium; if a non-racist business gets boycotted and people who defect from the boycott get boycotted themselves, then not being racist probably won't be worth it.
Isn’t there also the dual case for discrimination - that to hire an outsider is to reward an illegible but competing Backscratcher’s Club? This is the basis of many xenophobic claims. Additionally, one can argue that the WEIRD package of traits includes a fundamental disdain towards Backscratcher behavior, though perhaps it’s a more subtle disdain for using family/kin as a basis for coordination.
So *this* is the reason Technoblade hated orphans... After two years, new lore still drops on the normal schedule. That's insane.
Ahh, yes. Technoblade: the second-worst thing that ever happened to those orphans.
Sorry, you lost me at the start. I can definitely see some groups of people going with Albert's suggestion, but I believe a significant number of people would object to that and say "...no, Albert, that's weird, let's just draw the lots".
Yeah, the flaw in models like this is that norms matter. A must-read on that is Varshney, Ashutosh. "Nationalism, ethnic conflict, and rationality." Perspectives on politics 1, no. 1 (2003): 85-99. PDF here http://www-personal.umich.edu/~satran/PoliSci%2006/Wk%204-2%20Sacred%20Values%20Varshney.pdf
Perfectly rational people never are found in these situations, chiefly because such people don't exist. That is the problem with game theory, as opposed to social experiments. Such as the one where you are given $100, provided you must give some amount to someone else, explaining the whole situation, and both of you agree on the split. "Perfectly rational" people ought to be happy to get $1 even if you get $99. I'm clearly not perfectly rational.
Corollary: it is good to have a reputation for *not* being perfectly rational, but actually caring about silly things like fairness and self-respect, so that if someone gets to play this money-splitting game with you, they won’t try to stiff you because they will be afraid that you’ll decline anything less than an even split even if that means that the $100 goes into the paper shredder.
And the best way to get such a reputation is by regularly demonstrating that you won’t let people walk over you unpunished, even when that would be the best choice for you on a case-by-case basis. So being irrational in the short term can be the optimal strategy in the long term! This is just the standard tit-for-tat concept in game theory, of course.
I'm going to put out the corollary that, according to game theory, equity is not rational. Therefore, this is a flaw in game theory, since people don't actually behave that way.
Equity seems rational in iterated capitalism, with growth curves? When the money isn't just "points", but resources that can be put to use in an environment where we're all competing on multiple levels?
I admit that in a communist society, this problem goes away entirely, since the state takes everyone's winnings.
I meant equity in the equality sense: fairness. In game theory, you want the best outcome for yourself, or to achieve your goals, irrespective of what happens otherwise. "I wouldn't do that to my worst enemy" is bad game theory.
But game theory doesn’t stop at analyzing one-off encounters. There’s also iterated games, reputation effects, the concept of fairness (or equity if you prefer) as a natural Schelling point in negotiations, etc.
Lots of people have spent lots of effort into finding ways to show that a) fair and even seemingly altruistic behavior can emerge from "enlightened selfishness" by rational actors, and b) sufficiently sophisticated forms of game theory can actually be useful in explaining some things about how real-world people and animals behave.
"If you defect, you temporarily become my worst enemy, I will inflict punishment on you, and I am willing to pay a cost to inflict that punishment because seeing defectors punished has utility for me" is arguably good game theory.
This is a pretty cool bit of math:
https://en.wikipedia.org/wiki/Shapley_value
> In game theory, you want the best outcome for yourself, or to achieve your goals, irrespective of what happens otherwise.
Sure, but, even in a non-religious sense where it's an abstraction of accumulated years of self-reinforcing behaviors, I have a soul. The best outcome for me-as-a-person is not always going to be to behave as ruthlessly as possible, because the act of making that choice will affect me-as-a-person. Biological neural nets change when they're used, unlike the current generation of LLMs.
I maintain that this is not perfectly rational, and that the only rational approach is for them to give me $99 for the privilege of getting to keep $1.
Or I suppose we could split it 50/50; it's not my first choice but eventually I'd reluctantly compromise on it, so I might as well skip to the conclusion.
In the standard version of this game, one player gets to offer a split and the other gets to accept or decline; if the second player declines then nobody gets anything. There’s no negotiation.
Who needs negotiation? Non serviam. They're lucky I can't make an acausal bargain from the future to torture copies of them in virtual hells. It's $1 or $0 for them, and if they haven't figured this out, no skin off my back.
Also no money in your pocket, so how is that rational for you?
My righteous satisfaction will keep me warm in the cold winter nights.
Seriously, though, no way I'm letting the type of person who'd propose 99-to-1 splits get a 99x resource advantage over me. That does not end well for me.
To be a bit more formal, I reject the framing of the interaction which starts after they act but before I act. And I maintain that it is precisely as rational for me to demand a 99-1 split as for them to offer me a 1-99 split, and that if they want to come out of the interaction with money in their pocket, they need to think for just 1 second* about what kind of compromise someone else might find acceptable. Maybe they recognize that some people feel like I do, and take a gamble and offer an unfair trade anyway, but in that case they lose.
* Or however long it takes them to divide 100 by 2. For me it doesn't seem to require active thought at all.
Not just "that's weird", but "that's volunteering".
I was going to say that the obvious first victim should be whoever has the most meat on his bones, but then I saw that Heather works in Marketing, so Heather's going down.
If Heather is any good at all at her job, she should be able to show how much better all the other alternatives are. If not...I suppose that is evolution in action.
I was surprised this never came up. Let's eat Bob, he's the biggest person here. Bob needs to consume more calories than anyone else and contains the most calories in his body, so clearly we should eat him first.
Real world Backscratchers Clubs solved the "everybody joins" problem by not letting everyone join. Rotary Clubs, when they were founded, had a rule about allowing only one person from a given profession, because Alice was the only lawyer, so we all went to her for lawyer needs, and Bob was the only accountant, so we all went to him for accounting needs. And then they were also service clubs, which is now the main point, but at founding some younger professionals just decided the way to get ahead was to form a small cartel.
If you are looking for real terms, "cartel" might be a good one to pursue on this front.
Rotary clubs, when they were founded, would not have let Alice in whether she was a lawyer or not.
OK, but if the Rotary club is all dudes, and the club prostitute that everybody goes to for their "needs" is a dude, that could be interesting...
More seriously, I would expect Alice to benefit from her husband's membership; if we allow for female lawyers in this otherwise-patriarchal hypothetical, then everybody knows that they're supposed to go to Bob's wife for their lawyer needs.
Wouldn't it make more sense to include everyone from one profession, collude to raise prices, then lobby the government to ban non-members from practicing the profession (i.e. a guild/cartel)?
I don't see how one benefits from a club with one member from each profession. Whatever you gain by having other members buy your services, you lose by having to use other members' services even when going to a non-member would be advantageous. Job security perhaps? This doesn't feel like the most efficient way to get it.
This is exactly how e.g. realtors work in the Netherlands. Essentially its a guild that sets the prices
There is a (hypothetical) story by German sociologist Heinrich Popitz about passengers on a cruise ship. I found a summary in the article "On the difficulties of speaking out against security" by Thomas G. Kirsch, Anthropology Today (Volume 32, Issue 5, October 2016, Pages 5-7, https://rai.onlinelibrary.wiley.com/doi/10.1111/1467-8322.12295) Popitz's story is related to your story because it is also story about coordination and power dynamics, but it assigns a different role to ideology. Here is Kirsch's summary. (I have not read the rest of Kirsch's paper.)
"Imagine a cruise ship with a limited number of deckchairs. Imagine further that there are passengers on board who have a self-serving interest in their exclusive usage of them. An obvious strategy to achieve this interest is to occupy the deckchairs either in person or symbolically by placing towels on them. Yet since the latter can easily be removed, an additional and more efficient strategy consists in asking selected passengers to act as occasional custodians of the deckchairs, a role that is rewarded by temporarily granting them access to the deckchairs for themselves. Popitz thus argues that assuming power in this way relies on basic social mechanisms and does not require the existence of either an elaborate ideology or a vision of the future. Not so with those who have been excluded from using the deckchairs. When joining forces, the excluded opponents cannot content themselves
with simply taking possession of the deckchairs. Instead, when seeking to revolt against what they perceive to be an unjust deckchair system, they have to come up with a vision of how the allocation of deckchairs could in the future be organized in a different way. In short, they have to agree on what an alternative, more equitable and legitimate order might look like."
> In short, they have to agree on what an alternative, more equitable and legitimate order might look like."
Huh, that's an interesting insight. (From my limited understanding, that's what's happened to the left wing in Israeli domestic politics...)
>I’m sure real sociologists have written about these issues, but it was hard to find them and I figured I might as well write this post without citations. Still, if you know who they are and what their terms are, let me know.
The way to find basic literature on a topic like this, or any topic, is to google "[topic] syllabus" and you will get, well, syllabi of courses on the topic.
Also, political science might be the place to look, rather than sociology, esp if one wants to find game theoretical approaches to these sorts of questions.
So what should you put in for the topic name then? Lifeboat syllabus? Backscratcher syllabus? Orphan support club syllabus?
I thing Danning's suggestion is a good one, but you raise a good point. ChatGPT is good for these kinds of suggestions. I asked it "Is there any scientific literature on how clubs manage to self-perpetuate due to the costs and benefits of being in the club?" Collective action and clubs seems to be the answer. Maybe also "Group Cohesion" and "Conformity". Sounds like economics and psychology to me.
Try this: https://www.google.com/search?q=economics+%22Collective+action%22+and+%22clubs%22+syllabus
Yes, collective action is a good one, too. Though the fact that that search seems to turn up only Econ syllabi worries me. I know there is tons of literature in Poli Sci on collective action (unsurprising, given that that is much of what politics entails)
Well, immediately before the quote I included, Scott said:
>If the Lifeboat Games seemed suspiciously like nationalism, and the Backscratchers Clubs seemed suspiciously like clubs/cults/ideologies, the stories in this section seem suspiciously like the Establishment - whether it’s the Catholic Establishment of the Middle Ages, the conservative Establishment of mid-20th-century America, or the progressive Establishment of today. Elites support each other not directly - which would be hard to coordinate - but by all supporting the same ideology. If it’s hard for non-elites to break into the ideology, then everyone with the ideology will be elites, and supporting the ideology is an indirect way of elites supporting other elites in a big backscratching network. This is one of the solutions to Class Warfare Having A Free Rider Problem.
So, I took "these issues" to refer to nationalism, ideology, etc, so terms like that. But if the issue is game theory, the game theory, formal theory, or formal modeling would be terms I would recommend of the top of my head.
This "just so" story ignores actual human nature and society. Please see the story of William Bligh in a small boat, the story of the whalers who were the inspiration for Moby Dick, and the English and American lifeboat cases. Any court could not charge the survivors. Murder one for the instigators at a minimum. Life long guilt for the other survivors. Since the premise is not true in any society one might wish to live in, the rest of the discussion is not valid. This reminds me of Steve Sailor's discussion about the trolley car variations, just not as realistic.
He should read up on R v Dudley and Stephens: two lifeboat cannibals admitted what they did and were prosecuted when the reached shore, establishing the legal doctrine that necessity is not grounds for murder. Popular opinion disagreed and they were pardoned. The decision procedure was to eat the person who slipped into a hunger-induced coma, not draw lots or shout.
I'm thinking of that whaleship Essex and the guys who rowed away from it, into hell - and having the courage to yell out "So we're going to cannibalize each other because we're afraid of going due west and running across cannibals? Kill me now!"
And how interestingly tragic it was that there was a drawing of lots, and when it was the captain's nephew, he bravely submitted to his fate because honor culture compelled it; and it compelled his uncle to go on with it, too, because it would have been wrong to show nepotism. It's almost like, if it had been one of the others, maybe the men could have said - wait, this is nuts ... But then I guess his death really did save the rest, or all but a couple of them, I think.
I guess at that point no one had died, that they might have eaten. Or maybe someone had died, but not expecting this eventuality, they had pushed him overboard. I don't remember the grim details, except that physically the survivors bounced back pretty well, but I don't think their lives went all that great.
The principally interesting thing about these situations is the tenacity to cling to life. Of course hunger makes your brain work no so good. Still, when you are on the downward slope of life, in circumstances where that life doesn't seem likely to offer much else - the obvious solution is: let's all lie back and let death come nice and peaceful-like.
I guess that's just not possible for young men and men in their primes.
I don't think the Orphan Supporters one works as a way to solve coordination problems or benefit an Establishment. In the Elite, virtually everyone is an Orphan Supporter, just like everyone joined Daniel's Backscratcher Club, so it doesn't give a benefit. Anyone can become an Orphan Supporter, and it costs nothing as long as it mostly just requires saying the right words, so it doesn't protect the Establishment members from new competitors. You say it works if "it’s hard for non-elites to break into the ideology", but that isn't normally the case, especially for those who would otherwise have a shot at breaking into the elite. And it doesn't explain what incentive the Orphan Supporters have to punish someone who supports orphans as much as anyone, but doesn't do enough backscratching.
In politics proper, not everyone is an Orphan Supporter (as in the privacy of the ballot box people can support causes that are outside the norms), but that doesn't help the existing Establishment. At best, they share some seats in the legislature with the newcomer anti-Establishment Scrooges but manage to keep them out of actual power. At worst, sometimes they the Scrooges get the majority and they lose power. Meanwhile, if the Establishment didn't make support of orphans a requirement for membership, one half of the Establishment could take up the Scrooge cause, and the two halves of the Establishment together would dominate politics completely, as their existing parties and connections give them an advantage over newcomers.
Once the pro-Orphan norm is established, people will occasionally use it as a way to get rid of rivals via true or bogus accusations of Scroogery, and the possibility of that helps create an incentive for everyone to be an Orphan Supporter. But that (getting rid of a rival this way) is a rare event because most everyone is an Orphan Supporter, and the whole thing doesn't really benefit the Establishment. It does benefits the Orphan Support cause, though it may end up too silly to be effective at the margin.
The Communist Party in China is good irl example of large-scale 'backscratchers', often its necessary for career progression. However this might not quite be an internally stable system, people only seek the credential of 'CCP' for their own ends, and in a tragedy-of-the-commons I imagine people will only be favourable to other CCP members as much is necessary to keep up appearances or stop themselves being kicked out. It might depend on the level of intra-surveillance of the club.
Business associations sound a lot like instances of backscratcher clubs as well. The value of the association is all in the networking effects it offers. From what I understand, legal associations act to simplify lawyers' lives, by giving a clear network of referrals to point clients at as needed. Do we see similar incentives operating in "elite classes" as well? At least in B2B exchanges, the upsides of an exclusive referral network are high enough relative to costs, that joining multiple associations is fairly common.
Many businesses will donate to both Democrats and Republicans, or causes of such.
Don't forget that not everything is zero-sum either. Having businessmen find each other when they have completing products/needs/plans is positive sum; business associations that facilitate such connections may well have a benefit even if everyone joins. (Idk how they work though.)
Three thoughts:
1. The Schelling point in the original Albert call to eat Bob is interpreted as a breaking of symmetry to favor Bob. Indeed, it was. However, the symmetry is also broken by Bob as the caller. It's up to the group to determine which break of symmetry is going to motivate it's behavior.
2. A lack of utility in a backscratcher or orphan club comes if participation is universal. Accordingly, a club who's premise relies upon <100% participation can overcome. "This club's bylaws say that when participation exceed 50% +1, each month in such a state we shall banish the member who has tithed to us the least (or virtue signalled or...)."
3. Such games are of greatest interest when the cost of backscratching is assumed small. In a capitalist's world, the assumption is that favoring purchase of A over B for such extraneous factors leads to persistent nontrivial disutility in each transaction, and such backscratching clubs die for sub-competitive performance. (Leaving a capitalist to expect such clubs to thrive in only non-competitive markets. I leave it to the reader to conclude whether backscratching clubs appear most prevalent in such non-competitive markets today.)
3 is more "free market" than "capitalist", right? I'm sure someone could do ROI calculations for membership. But it seems like such calculations would erode the social fiction about orphans?
I once had similar thoughts (though not as clear and precise as you laid it out here) which led me to my theory of the left and right:
Left and right differ in their preferred method of coalition building: the right prefers to build a coalition with people they know – neighbors, family, friends – though doing it through religion and nationalism also works. While the left does it through abstract ideologies, and does it in a more indirect way – trying to hide what's going on. Both hate how the other group is doing it.
the left does it though "abstract ideologies", the right through "religion and nationalism"
I'm not sure the distinction here is as large as you think
My group builds coalitions along abstract ideologies in a good way, and your group builds coalitions along abstract ideologies in a bad way. Can't you see the difference?
I don't think this is fair - it's just that ideology isn't the important difference when considering left/right organization. Right and left are labels for the structure of the organization itself (in the context of capturing a democratic state). Left is bottom-up, swarm governance, by consensus. Right is top-down, monarchical governance, by command. These are platonic forms ofc; real-world left/right groups always do some level of both. Ideology is important only so much as it helps or justifies the organizational particulars.
My point, which I thought was also your point, is that religion and nationalism are "abstract ideologies" in the way the other commenter probably defines that term, and that that commenter just didn't seem to want to admit that the right "does it through abstract ideologies" too.
I think we more-or-less agree. I just didn't see OP's invocation of "religion and nationalism" as particularly negative or mean-spirited (tho it could have been, idk)
I didn't either. If anything, I took the commenter's invocation of "abstract ideologies" to be negative, which is why I thought it amusing that they casually mentioned religion and nationalism as axes of coalition-building for the right, as though those were not also "abstract ideologies."
Best post in a long time. Feels like a classic slatestarcodex post.
I think that this (or something like it) is something people intuitively grasp and triggers opposition to anything that starts to look like it. Obvious examples are hostility to effective altruism, modern anti-semitism and all the parts of the culture war which consist of each side thinking the other is a giant conspiracy that wants to hunt them down at some point.
Having said that, I'm not sure that these kind of not-quite-a-conspiracy backscratcher movements are really that common, or even occur at all. There are vanishingly few people who'd be engaging in them consciously compared to the numbers involved for a progressive/conservative establishment. If you're not consciously manipulative, aren't you the equivalent of someone who's just actually trying to help orphans and thinks they're struggling against orphan-hating moral mutants on the other side?
Looking for historical groups that took over society like this:
2nd-4th Century Christianity? Maybe, at the tipping point, but hard to believe it could work without the bulk of the support being genuine belief
The reformation might be more plausible given there's a movement among the princes to strengthen their grip and weaken the emperor, but that's more of a class interest than a group within a group. They also didn't want to leave half the population around them Catholic. They were shooting for total hegemony, much like the Catholics had been earlier.
The Meiji Oligarchy might be a better bet, but it's a very quick cascade towards support for the Emperor, and looks more like coalescing around a winner (eg. after a coup) which isn't the same phenomenon.
Political parties in general might be a better bet (coalition of hopefully 50+% of the population to benefit themselves at the expense of the smaller nearly half); I can't think of any country other than pre-war Hungary where a permanent coalition to shaft a chunk of the population has happened though.
Ignore my previous examples, they were all a bit off-base as these kind of movements should be much, much smaller to work. The best example is probably the Broderbond in South Africa, which I think fits this pattern.
What do you refer to w.r.t. pre-war Hungary? (Which war?)
WWI
Short version: in 1867 Hungary got fairly broad autonomy from Austria (a bit like a contemporary British dominion). Power ended up consolidated in the Liberal Party, which used a mixture of gerrymandering and patronage to control parliament in spite of being broadly unpopular among Magyars with the support of minority groups and people they were directly subsidising/employing.
> I can't think of any country other than pre-war Hungary where a permanent coalition to shaft a chunk of the population has happened though.
The Indian caste system? Racism in the United States pre-Civil Rights Act? Antisemitism in Christian-dominated Europe? The treatment of Cagots in France?
Other than [possibly] the Cagots, they're all building on pre-existing groups as opposed to forming groups for the purpose of discriminating. It's the formation that's the weird part.
Interesting idea about triggering opposition. Could we be looking at the origins of do-gooder derogation?
Not exactly real sociologist but this largely mirrors the object case in Tim Urban’a What’s our Problem? of SJWs in recent years
Can you explain what you mean when you say:
> Overall I would rather be Heather with her word-change campaign than Iolanthe with her adoption campaign.
Do you mean Heather is in an advantageous position, or that you think you'd sleep better at night if you were Heather?
I'm guessing the first, since most of the post ignores the aspect of sleeping better.
Fascinating convo.
My 2 cents:
Better to be a Marxist.
No, not Karl. Groucho,
Who famously said,
"I refuse to join a club that would have me as a member."
Was I the only one who assumed on round two everyone would say "Let's eat ALBERT" since he defected in round one?
Seconded. The Schelling point is to eat the guy who ruined our fair lots system.
And knowing this, Albert won’t suggest it in the first place.
This fails when there are exactly three people, though.
How? Don't Charlotte and Bob gang up on Albert?
Oh yeah, the tiebreaker could just decide which one they want dead.
Me too. I'd even try it on round one.
Unfortunately, forming a backscratcher's club based on this principle would probably backfire.
"Consider for example racism. There are supposed justifications for racism - like that such and such a race is inferior, or oppressive, or plotting to kill us. But another justification is just “We’re the majority and they’re the minority, and if we all band together to profit at their expense, it probably goes well for us.” Any coalition of 51%+ can do this. But it’s easier if everybody comes color-coded so there’s one obvious coalition that occurs to everybody and which they can easily check that they’re a part of"
There is a lot of real-world scholarship on this question, and I only know some of it (and even that is nearly two decades out of date at this point), so if anyone knows more or more recent work please correct me, but from what I've read historians who study the origins of racism make it sort of halfway between these, or maybe a bit skew to either.
The argument is that racism is largely economically motivated. One pattern goes like this: it's really economically powerful to kidnap people and force them to work for you under pain of death and torture, so people want to do it, but it's too obviously evil, so you start think "well they deserve it" and "it's better for them anyway"—not only or even mainly as conscious ways to get other people to sign off on it, but as ways to justify it to yourself.
Another stage is the maintenance/revival of racism. Take the post-civil war south: racism there served the interests of the wealthy because it helped stop intra-working class alliances which did things like support unions. (There were some of those in the decades immediately following the war.) Why did working-class whites go along with this? One common answer is that racism provided what W. E. B. DuBois famously called a "psychological wage": they got to feel superior to other people. Presumably this is easier to achieve than, say, actually forming a union & bargaining & achieving a more just society, so there's temptation to believe it. And there's at least *some* economic benefits to poor whites, too, since they formed a sort of upper working class (not a standard term afaik but compare "upper middle class"), which made them better off.
So is this “We’re the majority and they’re the minority, and if we all band together to profit at their expense, it probably goes well for us.”? It's not entirely different. But it's closer to, "we're already profiting at their expense, and only bad people would do that unless they're inferior, and we're not bad people, therefore they are inferior"; or, in the second example, "if we richer people get group A to hate Group B they won't notice us picking their pockets. ("Only a pawn in their game...".)
But of course the above is simple and real people are complicated, so I think in real history it's an incredibly complex dance of these factors: there are already some ideas about inferiority, and you want to believe you're not a bad person, so they sound more persuasive to you, you talk them up more, other people in a similar situation hear them, etc. Or, you, a rich person in the south, hear positive and negative information about a lot of people, and you are subtly (possibly subconsciously) inclined to notice the negative information about black people more, and then you pass it on to others including your employees, etc. As James Madison famously put it, "As long as the connection subsists between [man's] reason and his self-love, his opinions and his passions will have a reciprocal influence on each other". Which is to say, I think the mechanisms here aren't as simple as the historical sketches above (and which can get downright silly if you interpret them literally rather than as schemas), but the economic motives provide a subtle pressure to notice, spread, and even develop the ideas which are in some ways useful to you (which is like what Soctt is saying), but also that psychological usefulness is at work too (which is another element).
White workers may well have benefitted, if employers had to hire them in preference to blacks in some cases. Thomas Sowell said white workers in South Africa pushed for apartheid to protect them from black competition. But in this case it's employers whose interests were against racist policy: segregation essentially functioned as a white workers' union/cartel; a union works best for the interests of its members if it excludes some people who could do the job, and can have them prevented from competing with it.
Meanwhile, if employers could hire black people freely, even if separately, then workers (both white and black) might have made somewhat less money than if they worked together, but white and black people would have got a similar salary for equivalent work.
If black people made much less for equivalent work, that points to "my" version (the first paragraph above); if they got a similar pay for similar quality work in a similar position, that points to yours.
> One common answer is that racism provided what W. E. B. DuBois famously called a "psychological wage": they got to feel superior to other people.
I'd really like some way to highlight this brilliant observation.
The buzzword version of this is called "last place aversion", apparently.
https://www.scientificamerican.com/article/occupy-wall-street-psychology/
There is a clear downside to joining Daniel's backscratchers club: you have to favor other backscratchers over better qualified alternatives.
So while the club is getting started, a startup founder is a member and is trying to hire a software developer. A homeless person comes in and asks for the job.
"Do you have programming experience?"
"No. But I'm a fellow backscratcher, so you have to hire me."
"Fine. I quit the backscratchers."
The founder's fiance is then approached by their ex.
"I'm also a backscratcher, marry me instead."
"No way. I quit."
And so on.
Well yes, but then the founder is ostracized in a "2 minutes of hate" that is calculated to be enough of an overreaction to deter anyone else from resigning for the next 18 months. All backscratchers participate, especially any who worry that their loyalty might be questionable.
And the less said about what happens to the fiance, the better. Eventually he makes a tearful confession, and admits to subversive anti-backscratcher behavior, and submits himself to a re-education camp for recovering apostates. His physical health never fully recovers, and he occasionally breaks down sobbing for no apparent reason, but he insists that he's *happy* now.
Except that Daniel's backscratchers club does not have a 2 minutes of hate as a founding rule and if it did, it would have to be big enough that the 2 minutes of hate would be meaningful and difficult to escape.
I'm confident that these things would evolve.
Perhaps. But wouldn't the existence of the two minutes of hate provide an even bigger downside for joining in the first place?
I'm pushing back on the claim Scott made that joining Daniel's backscratchers has no downsides.
Ah, OK. I was pushing back on the idea that leaving Daniel's backscratchers would free oneself from the downsides. :-)
Well, this is why real mutual aid societies vet their members (or did). These days they'll take most anyone who isn't obviously a basket case, but at one time you needed to actually be at least somewhat successful and reputable...
it's giving old ssc (being kinda dumb)
The Shelling point in the array of integers was displayed as a hyperlink in my browser. I clicked on it and my iPhone asked if I wanted to call the number. 938 is an area code in Huntsville Alabama.
This whole situation could have been avoided if, upon Albert yelling "let's kill and eat Bob," everyone else replied "OMG no, that's horribly unfair to Bob! Let's draw lots."
The golden rule, "don't do to others what you wouldn't want them to do unto you," has been a staple of human morality for centuries for a good reason. If I were on an ACX boat, I certainly would not want anyone to yell "let's kill and eat drosophilist!" and everyone else to go "yup, sure sounds like a swell idea!" Nor would I want people to coalesce around the Schelling point of "drosophilist is a woman, all the rest of us are men (not an unlikely scenario given the gender distribution of ACX readers), so let's kill and eat her!" I value my life. And because I wouldn't want others to do it to me, I ought not to do it to others!
I gotta say, the utter lack of compassion behind all these calculations of "let's see, if I draw lots I have a 10% chance of dying, but if I throw another person under the bus I have a 0% chance of dying (for the moment, until someone else has the bright idea of shouting MY name on the next round)" is chilling.
All this reminds me of something I read a while back:
Bob is sitting at home when there's a knock on the door. He opens the door and sees an ominous-looking stranger with a box with a lever on it. The stranger says, "If you pull this lever, I will give you $10,000, and someone will die."
"Who will die?" asks Bob.
"Someone far away," replies the stranger. "Nobody you know."
The stranger leaves the box with Bob, who agonizes for a while, but then rationalizes that people die every day, he doesn't owe some distant random person anything, and he's really not doing so well financially and $10,000 would really help him. So, finally, he pulls the lever.
There's another knock on the door. The stranger is back. He gives Bob $10,000 and asks for his box back. Bob hands him the box and asks, "What are you going to do now?"
The stranger replies, "I'm going to make the same offer to someone else. Someone far away. Nobody you know."
> I value my life. And because I wouldn't want others to do it to me, I ought not to do it to others!
Yep. The Torah while standing on one foot.
Exactly!
Yes, thank you, I'd be happy to share a lifeboat with you. (Sort of. You know what I mean.)
Right back at you. You come across as very thoughtful in your comments.
Yes, I know what you mean. 😊
That was a Twilight Zone episode, actually.
https://en.wikipedia.org/wiki/Button,_Button_(The_Twilight_Zone)
What does the stranger do after Bob gives him the box back without having pulled the lever?
Presumably, he rewards Bob's ethics by not choosing him to be the person to be killed when someone else pulls the lever? I would like to think so. But I don't actually know.
The resolution to all of these paradoxes is that at no point in these stories do any of the characters demonstrate morality. They're all optimizing for outcomes like a rational creature that has no beliefs and is only self-interested. In real life people do have morality and do not only do self-interested things (except perhaps in states of extreme insecurity). Now you can cast that as another form of security: like after a point, one's safety is better increased by living in a fair and just and kind society than by getting any more personal material benefits. Maybe that's all morality is. But in any case none of your characters care about that at all, hence they are inhuman or at least very dehumanized.
Well, for starters, I shouldn't have to define morality, because people have been talking about it for at least 2000 years and it's a concept most people know well. So it is not an empty point: I'm answering the question by pointing at "a giant thing that you are completely familiar with" and saying "what about that?"
Second, it works for almost any definition you want. I am not trying to solve the problem of "explain this paradox in terms of a rational model of humans as optimizing agents". I am trying to solve the problem "explain why this paradox feels like a paradox when held up against my human intuition for how reality works". After all the point of all paradoxes is that they feel wrong, but it's hard to put into words why. So if you're Scott Alexander and you write this post and you think: "wow this is interesting why does no one talk about this?" it's because everyone else is thinking about moral humans, not this horrible and repuugant thing.
Specifically, he wrote "Something about this surprises me. It’s weird that there’s another solution which is more stable than the fair one of drawing lots. It’s strange that by shouting an obvious suggestion - one that adds no more information - Albert can save his own life with certainty. Still, that’s how it goes." What's surprising is that when you model humans *incorrectly* like this, the results seem wrong, but you can't easily tell how you modeled them incorrectly. The resolution is that actual humans do not work like this. Actual humans can perceive the game theory of the situation from the start and especially the unjustness of the solutions that fuck over one person arbitrarily, and they prefer a situation in which they are not at the whims of being fucked over by game theory, hence they have a moral intuition for the general rules that everyone should follow (and everyone should *signal* that they will follow) and they will strongly adhere to those rules---because they are *not* self-interested in this way.
>Well, for starters, I shouldn't have to define morality, because people have been talking about it for at least 2000 years and it's a concept most people know well
We don't agree on object level morality , which is why we still need politics. We don't even agree which kind of equality is the right kind.
Fine, but my point isn't that you need consider *my* morality because that's what resolves these paradoxes, it's that you need consider morality *at all* because it's an entirely missing dimension here.
"What about that?"
What about it? Scott pointed out that his hypothetical situations merely demonstrate the incentive structures behind real-life repugnant behavior.
If you're going to claim that an allegory for discrimination is unrealistic because the people in it aren't behaving morally, you need to contend with the fact that non-hypothetical people have practiced discrimination for our entire existence, despite having every reason to behave morally. The people on the lifeboat are actually going to starve; in real life, we don't even need that much pressure to start getting ugly.
I guess you believe that's how people work. I, for the most part, do not, and I certainly don't want to take part in an intellectual exercise that doesn't content with the *idea* of people behaving morally, since that just reinforces the idea that people don't. Better to talk about how to be moral than how to be a perfectly cynical backstabber.
strangely the last paragraph of my reply is just not showing up? It was
(Although again I will add the disclaimer that in an actual stranded lifeboat situation, maybe they are psychopathic like that; I've never been there and I imagine stress can diminish all this moral stuff. Hopefully they're strong enough not to. In a way that's what strength. Regardless, in a *society* they are not. Although probably a few people are anyway, and that's why the signaling is so important: because everyone wants to know that they are going to be safe from everyone else.)
Very good point. That's what I was trying to get at with my comment.
I think in real life a network of relationships would have sprung up among the people on the boat. There would be alliances and partial alliances and people or groups who disliked or distrusted other people or groups. There would be one or more people seen as potential leaders, or one person whom the majority saw as the leader. And it seems to me the leader would have a lot of influence over how the cannibalism possibility was handled,
If I were on the boat I would be drawn to someone fair-minded and kind who was seen as a potential leader. I wouldn't be drawn towards them because of my values, I'd be drawn because I'd see having them as leader as my best chance of having a decent outcome to the situation. A fair-minded and kind leader would probably stand up for using a lots system(after maybe first asking for a volunteer, because ya never know, there might be someone who would volunteer if given a chance). But it's also possible that the person seen as the most dangerous could end up as the leader, because people feared opposing them. Maybe Violent Guy or one of his allies would kill me in my sleep if I opposed him. Violent guy, if leader, would probably choose his strongest opponent as the first to be eaten.
Overall, it seems to me that the lifeboat, even as a toy example, doesn't do a good job of capturing how groups function. Most of our relationships with other people are valenced, so groups quickly develop a structure.
"I think in real life a network of relationships would have sprung up among the people on the boat."
We have one modern-day example of this, and a movie was recently released about it:
https://en.wikipedia.org/wiki/Uruguayan_Air_Force_Flight_571
"Uruguayan Air Force Flight 571 was the chartered flight of a Fairchild FH-227D from Montevideo, Uruguay, to Santiago, Chile, that crashed in the Andes mountains on 13 October 1972. The accident and subsequent survival became known as the Andes flight disaster (Tragedia de los Andes, literally Tragedy of the Andes) and the Miracle of the Andes (Milagro de los Andes).
...Three crew members and nine passengers died immediately and several more died soon after due to the frigid temperatures and the severity of their injuries. ...Search and rescue aircraft overflew the crash site several times during the following days, but failed to see the white fuselage against the snow. Search efforts were called off after eight days of searching.
During the 72 days following the crash, the survivors suffered from extreme hardships, including sub-zero temperatures, exposure, starvation, and an avalanche, which led to the deaths of 13 more passengers. The remaining passengers resorted to eating the flesh of those who died in order to survive. Convinced they would die if they did not seek help, two survivors, Nando Parrado and Roberto Canessa, set out across the mountains on 12 December. Using only materials found in the aircraft wreck, they climbed 839 metres (2,753 ft) from the crash site up 30-to-60 degree slopes to a 4,503-metre (14,774 ft) ridge to the west of the summit of Mount Seler. From there they trekked 53.9 kilometres (33.5 mi) for 10 days into Chile before finding help. On 22 and 23 December, 1972, two and a half months after the crash, the remaining 14 survivors were rescued. Their survival made worldwide news."
https://en.wikipedia.org/wiki/Society_of_the_Snow
"On October 13, 1972, Uruguayan Air Force Flight 571, chartered by an Uruguayan rugby football team and their supporters to take them to a game in Santiago, Chile, crashes into a glacier in the heart of the Andes mountains. Of the 45 passengers on board, 29 survive the initial crash, although more would die from injury, disease, and an avalanche over the following weeks. Trapped in one of the most inaccessible and hostile environments on the planet, the survivors are forced to resort to survival cannibalism of those who had already died in order to stay alive. However, rather than turn against each other, the survivors draw upon the cooperative teamwork they learned through rugby, and spiritual faith, in order to escape the mountains."
The reality seems to have been somewhat more complicated than that, but the survivors did establish a way of dealing with the necessity, and it probably did help that they didn't have to kill anyone in order to obtain corpses.
I think in practice most of the people would volunteer to sacrifice themselves because people are mostly good and moral, and they'd end up agreeing to draw lots. Maybe not, though, because as I understand it morality breaks down when people are super unsafe (cf what starving people do for food). But yeah I think the cynical self-interested response you describe is basically psychopathic. Personally, barring some extreme psychological change, I'd be drawn to a fair-minded and kind leader because of my values.
"If I were on the boat I would be drawn to someone fair-minded and kind who was seen as a potential leader."
Seriously!
ACX is a fascinating place and great for thought experiments, but the amount of amoral calculation focused on "how do I survive/improve my status, and everyone else can go pound sand" gets disturbing at times. (What makes it really paradoxical is that we know Scott is a good person! He gave up his own kidney to a stranger for crying out loud!)
If I'm ever on a lifeboat, I'm teaming up with you, Alex, and Moon Moth. Team Kindness & Fairness FTW!
Kindness for the Kindness God! Fairness for the Fairness Throne!
😁
I don't think Scott endorses any of this as being morally right--he says it's obvious that this is what people will do, not that it's good--and I read it as following his previous statements about how people respond to their incentives (or get out-competed by those who don't).
If you're going to make the argument that these situations aren't realistic because the people in them aren't behaving morally, you've got to contend that they're all allegories demonstrating the incentive structures which have led to many real-life atrocities (committed by people who had plenty of reason to behave morally).
I think it's good, as an intellectual exercise, to speculate about what would happen in the absence of morality.
Because in real life, while morality does exist, it appears that most people are going to choose self-interest over morality most of the time, at least if the self-interest is relatively large. So our speculations about what happens in the total absence of morality are likely to be a reasonable approximation for what happens in real life. And it's instructive to take the models describing an absence of morality and check how well or not well they correspond to real world phenomena.
By the time we got to Greg, I wanted someone to say "let's just go kill Erica and have her money."
In all the Backscratcher Clubs, the obligation to favor your fellow club members imposes a cost that I don't think you accounted for, that can sometimes be quite high. For example if you are seen doing favors for someone with a bad reputation, that can hurt your reputation as well. Or if you're a politician and one of your fellows asks you to throw your weight behind their stupid pet cause, it's just squandering your political capital. It actually pays to be choosy in joining clubs of mutual obligation.
This is how Tammany Hall worked, though, and very successfully for a while. There wasn't absolute obligation to do favours for people with bad reputations or crazy notions, so you could say "I can't help you out on this, but I can do that for you".
https://en.wikipedia.org/wiki/Tammany_Hall
"It became the main local political machine of the Democratic Party and played a major role in controlling New York City and New York State politics. It helped immigrants, most notably the Irish, rise in American politics from the 1850s into the 1960s. Tammany usually controlled Democratic nominations and political patronage in Manhattan for over 100 years following the mayoral victory of Fernando Wood in 1854, and used its patronage resources to build a loyal, well-rewarded core of district and precinct leaders; after 1850, the vast majority were Irish Catholics due to mass immigration from Ireland during and after the Irish Famine of the late 1840s.
After 1854, it expanded its political control even further by earning the loyalty of the city's rapidly expanding immigrant community, which functioned as its base of political capital. The business community appreciated its readiness, at moderate cost, to cut through regulatory and legislative mazes to facilitate rapid economic growth.
...Throughout the 1830s and 1840s, the Society expanded its political control even further by earning the loyalty of the city's ever-expanding immigrant community, which functioned as a base of political capital. During the 1840s, hundreds of thousands of Irish immigrants arrived in New York City to escape the Great Famine and Tammany saw its power grow greatly.
Tammany Hall's electoral base lay predominantly with New York's burgeoning immigrant constituency, which often exchanged political support for Tammany Hall's patronage. In pre-New Deal America, the extralegal services that Tammany and other urban political machines provided often served as a rudimentary public welfare system."
So by providing services from A to Z for the voters, and making sure that you take good care of local issues and local problems as they arise, you consolidate power. Be powerful enough, and you can afford to turn down doing favours (openly) for bad apples or pet causes, because the majority of your clients know that you will continue to look after them.
Reading that article is a treasure trove of early political history in the USA. Voter fraud and voter suppression is no new thing:
"In the 1830s the Loco-Focos, an anti-monopoly and pro-labor faction of the Democratic Party, became Tammany's main rival for votes by appealing to workingmen. However, Tammany's political opponent remained the Whigs. During the 1834 New York City mayoral governor election, the first city election in which the popular vote elected the mayor, both Tammany Hall and the Whig party, from their headquarters at the Masonic Hall, battled in the streets for votes and protected polling locations in their respective regions from known opposition voters. During the 1838 state election for governor, the rival Whig party imported voters from Philadelphia, paying $22 a head for votes in addition to paying for votes at their polling places. Tammany Hall operatives continued their practice of paying prisoners of the almshouses for votes and also paying for votes at their polling places."
Someone on The Motte recommended the book "Plunkitt of Tammany Hall" about/by a former leader of the organisation, and former New York State Senatoar, George Washington Plunkitt, who describes the methods by which he operated - what he called "honest graft" as opposed to "dishonest graft":
https://www.gutenberg.org/cache/epub/2810/pg2810-images.html#link2HCH0006
"EVERYBODY is talkin' these days about Tammany men growin' rich on graft, but nobody thinks of drawin' the distinction between honest graft and dishonest graft. There's all the difference in the world between the two. Yes, many of our men have grown rich in politics. I have myself. I've made a big fortune out of the game, and I'm gettin' richer every day, but I've not gone in for dishonest graft—blackmailin' gamblers, saloonkeepers, disorderly people, etc.—and neither has any of the men who have made big fortunes in politics.
There's an honest graft, and I'm an example of how it works. I might sum up the whole thing by sayin': "I seen my opportunities and I took 'em."
Just let me explain by examples. My party's in power in the city, and it's goin' to undertake a lot of public improvements. Well, I'm tipped off, say, that they're going to lay out a new park at a certain place.
I see my opportunity and I take it. I go to that place and I buy up all the land I can in the neighborhood. Then the board of this or that makes its plan public, and there is a rush to get my land, which nobody cared particular for before.
Ain't it perfectly honest to charge a good price and make a profit on my investment and foresight? Of course, it is. Well, that's honest graft.
Or supposin' it's a new bridge they're goin' to build. I get tipped off and I buy as much property as I can that has to be taken for approaches. I sell at my own price later on and drop some more money in the bank.
Wouldn't you? It's just like lookin' ahead in Wall Street or in the coffee or cotton market. It's honest graft, and I'm lookin' for it every day in the year. I will tell you frankly that I've got a good lot of it, too.
I'll tell you of one case. They were goin' to fix up a big park, no matter where. I got on to it, and went lookin' about for land in that neighborhood.
I could get nothin' at a bargain but a big piece of swamp, but I took it fast enough and held on to it. What turned out was just what I counted on. They couldn't make the park complete without Plunkitt's swamp, and they had to pay a good price for it. Anything dishonest in that?
Up in the watershed I made some money, too. I bought up several bits of land there some years ago and made a pretty good guess that they would be bought up for water purposes later by the city.
Somehow, I always guessed about right, and shouldn't I enjoy the profit of my foresight? It was rather amusin' when the condemnation commissioners came along and found piece after piece of the land in the name of George Plunkitt of the Fifteenth Assembly District, New York City. They wondered how I knew just what to buy. The answer is—I seen my opportunity and I took it. I haven't confined myself to land; anything that pays is in my line.
For instance, the city is repavin' a street and has several hundred thousand old granite blocks to sell. I am on hand to buy, and I know just what they are worth.
How? Never mind that. I had a sort of monopoly of this business for a while, but once a newspaper tried to do me. It got some outside men to come over from Brooklyn and New Jersey to bid against me.
Was I done? Not much. I went to each of the men and said: "How many of these 250,000 stories do you want?" One said 20,000, and another wanted 15,000, and other wanted 10,000. I said: "All right, let me bid for the lot, and I'll give each of you all you want for nothin'."
They agreed, of course. Then the auctioneer yelled: "How much am I bid for these 250,000 fine pavin' stones?"
"Two dollars and fifty cents," says I.
"Two dollars and fifty cents!" screamed the auctioneer. "Oh, that's a joke! Give me a real bid."
He found the bid was real enough. My rivals stood silent. I got the lot for $2.50 and gave them their share. That's how the attempt to do Plunkitt ended, and that's how all such attempts end.
I've told you how I got rich by honest graft. Now, let me tell you that most politicians who are accused of robbin' the city get rich the same way.
They didn't steal a dollar from the city treasury. They just seen their opportunities and took them. That is why, when a reform administration comes in and spends a half million dollars in tryin' to find the public robberies they talked about in the campaign, they don't find them.
The books are always all right. The money in the city treasury is all right. Everything is all right. All they can show is that the Tammany heads of departments looked after their friends, within the law, and gave them what opportunities they could to make honest graft. Now, let me tell you that's never goin' to hurt Tammany with the people. Every good man looks after his friends, and any man who doesn't isn't likely to be popular. If I have a good thing to hand out in private life, I give it to a friend—Why shouldn't I do the same in public life?
Another kind of honest graft. Tammany has raised a good many salaries. There was an awful howl by the reformers, but don't you know that Tammany gains ten votes for every one it lost by salary raisin'?
The Wall Street banker thinks it shameful to raise a department clerk's salary from $1500 to $1800 a year, but every man who draws a salary himself says: "That's all right. I wish it was me." And he feels very much like votin' the Tammany ticket on election day, just out of sympathy.
Tammany was beat in 1901 because the people were deceived into believin' that it worked dishonest graft. They didn't draw a distinction between dishonest and honest graft, but they saw that some Tammany men grew rich, and supposed they had been robbin' the city treasury or levyin' blackmail on disorderly houses, or workin' in with the gamblers and lawbreakers.
As a matter of policy, if nothing else, why should the Tammany leaders go into such dirty business, when there is so much honest graft lyin' around when they are in power? Did you ever consider that?
Now, in conclusion, I want to say that I don't own a dishonest dollar. If my worst enemy was given the job of writin' my epitaph when I'm gone, he couldn't do more than write:
"George W. Plunkitt. He Seen His Opportunities, and He Took 'Em."
Also, if you want votes from the opposition or the moderate middle or the squishy, get to the kids first and bribe them with candy 😁
"Chapter 6. To Hold Your District: Study Human Nature and Act Accordin'
There's only one way to hold a district: you must study human nature and act accordin'. You can't study human nature in books. Books is a hindrance more than anything else. If you have been to college, so much the worse for you. You'll have to unlearn all you learned before you can get right down to human nature, and unlearnin' takes a lot of time. Some men can never forget what they learned at college. Such men may get to be district leaders by a fluke, but they never last.
To learn real human nature you have to go among the people, see them and be seen..1 know every man, woman, and child in the Fifteenth District, except them that's been born this summer—and I know some of them, too. I know what they like and what they don't like, what they are strong at and what they are weak in, and I reach them by approachin' at the right side.
For instance, here's how I gather in the young men. I hear of a young feller that's proud of his voice, thinks that he can sing fine. I ask him to come around to Washington Hall and join our Glee Club. He comes and sings, and he's a follower of Plunkitt for life. Another young feller gains a reputation as a baseball player in a vacant lot. I bring him into our baseball dub. That fixes him. You'll find him workin' for my ticket at the polls next election day. Then there's the feller that likes rowin' on the river, the young feller that makes a name as a waltzer on his block, the young feller that's handy with his dukes—I rope them all in by givin' them opportunities to show themselves off. I don't trouble them with political arguments. I just study human nature and act accordin'.
But you may say this game won't work with the high-toned fellers, the fellers that go through college and then join the Citizens' Union. Of course it wouldn't work. I have a special treatment for them. I ain't like the patent medicine man that gives the same medicine for all diseases. The Citizens' Union kind of a young man! I love him! He's the daintiest morsel of the lot, and he don't often escape me.
Before telling you how I catch him, let me mention that before the election last year, the Citizens' Union said they had four hundred or five hundred enrolled voters in my district. They had a lovely headquarters, too, beautiful roll-top desks and the cutest rugs in the world. If I was accused of havin' contributed to fix up the nest for them, I wouldn't deny it under oath. What do I mean by that? Never mind. You can guess from the sequel, if you're sharp.
Well, election day came. The Citizens' Union's candidate for Senator, who ran against me, just polled five votes in the district, while I polled something more than 14,000 votes. What became of the 400 or 500 Citizens' Union enrolled voters in my district? Some people guessed that many of them were good Plunkitt men all along and worked with the Cits just to bring them into the Plunkitt camp by election day. You can guess that way, too, if you want to. I never contradict stories about me, especially in hot weather. I just call your attention to the fact that on last election day 395 Citizens' Union enrolled voters in my district were missin' and unaccounted for.
I tell you frankly, though, how I have captured some of the Citizens' Union's young men. I have a plan that never fails. I watch the City Record to see when there's civil service examinations for good things. Then I take my young Cit in hand, tell him all about the good thing and get him worked up till he goes and takes an examination. I don't bother about him any more. It's a cinch that he comes back to me in a few days and asks to join Tammany Hall. Come over to Washington Hall some night and I'll show you a list of names on our roll' marked "C.S." which means, "bucked up against civil service."
As to the older voters, I reach them, too. No, I don't send them campaign literature. That's rot. People can get all the political stuff they want to read—and a good deal more, too—in the papers. Who reads speeches, nowadays, anyhow? It's bad enough to listen to them. You ain't goin' to gain any votes by stuffin' the letter boxes with campaign documents. Like as not you'll lose votes for there's nothin' a man hates more than to hear the letter carrier ring his bell and go to the letter box ex pectin' to find a letter he was lookin' for, and find only a lot of printed politics. I met a man this very mornin' who told me he voted the Democratic State ticket last year just because the Republicans kept crammin' his letter box with campaign documents.
What tells in holdin' your grip on your district is to go right down among the poor families and help them in the different ways they need help. I've got a regular system for this. If there's a fire in Ninth, Tenth, or Eleventh Avenue, for example, any hour of the day or night, I'm usually there with some of my election district captains as soon as the fire engines. If a family is burned out I don't ask whether they are Republicans or Democrats, and I don't refer them to the Charity Organization Society, which would investigate their case in a month or two and decide they were worthy of help about the time they are dead from starvation. I just get quarters for them, buy clothes for them if their clothes were burned up, and fix them up till they get things runnin' again. It's philanthropy, but it's politics, too—mighty good politics. Who can tell how many votes one of these fires bring me? The poor are the most grateful people in the world, and, let me tell you, they have more friends in their neighborhoods than the rich have in theirs.
If there's a family in my district in want I know it before the charitable societies do, and me and my men are first on the ground. I have a special corps to look up such cases. The consequence is that the poor look up to George W. Plunkitt as a father, come to him in trouble—and don't forget him on election day.
Another thing, I can always get a job for a deservin' man. I make it a point to keep on the track of jobs, and it seldom happens that I don't have a few up my sleeve ready for use. I know every big employer in the district and in the whole city, for that matter, and they ain't in the habit of sayin' no to me when I ask them for a job.
And the children — the little roses of the district! Do I forget them? Oh, no! They know me, every one of them, and they know that a sight of Uncle George and candy means the same thing. Some of them are the best kind of vote-getters. I'll tell you a case. Last year a little Eleventh Avenue rosebud, whose father is a Republican, caught hold of his whiskers on election day and said she wouldn't let go till he'd promise to vote for me. And she didn't."
There may be a lot of howling over Donald Trump, but I feel he would have been right at home in Tammany Hall era Democratic party of New York!
Scott Alexander, the place to go for people who want to understand and mitigate racism out of genuine curiosity rather than guilt.
(social media is where guilt-motivated people go, with academia now mainly adding an intellectual spin to norms that emerged online rather than leading the curve like the late 2000s).
Regarding the lifeboat games, a natural Schelling point is that "rulebreaking is bad". When Albert shouts to eat Bob, Juanita pops up and says "It looks like we can't trust you to keep agreements, but fortunately it sounds like you just volunteered. There's nothing special about Bob, but you, Albert, we all know exactly what you'd do if you had the chance."
Which would would you prefer to live in? You get to choose!
Oh maybe they eat Juanita instead.
The kind of people who'd rather eat Juanita than Albert would probably kill the rest of us in our sleep. They're moral monsters and must be eliminated. No mercy. (Mostly kidding. I don't think I'd actually make that argument.)
There's actually a book on Judaism that reminds me of this. Judaism Straight Up by Moshe Koppel. It's a great read, even if you are not Jewish.
It seems to me like a subsequent "obvious" outcome is that, being granted so much status by the backscratchers, actual orphans co-opt the club and turn it into a lifeboat game serving their own purposes. Non-orphan members are incentivized to stay in the club, since membership is still a requirement for access to orphan-club-controlled institutions, but they also pay higher dues that actually benefit the orphans rather than just enough to act as free-rider deterents. A hallmark of this process would be increasing punishments for dissenters/non-members to further incentivize non-orphans to maintain membership despite its rising cost. The reason why these high costs aren't rejected, like in the adoption example, is that the orphans themselves do not suffer from increasing membership costs. Rather, they benefit from it since it's now lifeboat game (aka don't stick out and get eaten) and they are Albert in that they can leverage the fact that their favor serves as the defining in-group signal in order to enforce compliance.
If you are a non-orphan who was previously benefiting from orphan club membership, what can you do? The most authoritarian and self-serving members of the very group whose promotion defines your club are the ones you need to demote to reestablish symbiosis, but the second you do, you get Alberted. In a cruel inversion, the original members find themselves abandoned by the organization that once favored and protected them.
Orphanship might not actually be susceptible to this sort of comendeering since what orphans probably most want is to not be orphans, aka get adopted, at which point they lose their priviledged orphan status (POS). Coopted backscratcher clubs targeting individuals with the permanent, highly identifiable traits that serve well in lifeboat games would be much more durable.
People could try to disown their family and get themselves legally declared to be orphans?
That's an interesting countermove. The viability of identifying as a member of the club-defining group as a route to regain its benefits probably depends on how socially malleable the identified trait is. How socially malleable the trait is is itself socially malleable, so I'd expect most of the discourse to take place on this meta-level.
In any case, it seems inevitable that dead-parent orphans would vehemently resist their power being diluted by potentially opportunisic legal orphans, creating space for a counter-club defined by its opposition to legal-disownment exclusionary radical orphans (LEROs).
> legal-disownment exclusionary radical orphans
(We need to figure out a way to add a 'Y' at the end, so we can call them "leroys".)
That aside, looking at the rest of the post... I notice that you have described a problem without proposing a solution? ;-)
One would think that Albert, by proposing everyone eat Bob, would self identify as the best candidate to be eaten. Then when time comes to eat some one else, no one would propose a candidate (except possibly himself if he wished to signal his own non-edible worth) instead of drawing lots.
> Something about this surprises me. It’s weird that there’s another solution which is more stable than the fair one of drawing lots. It’s strange that by shouting an obvious suggestion - one that adds no more information - Albert can save his own life with certainty. Still, that’s how it goes.
No, the callee would immediately suggest choosing Albert instead. Now the others have two options, both suggesting the other, and they would have no reason to choose the original callee over Albert. Albert has broken the initial N-way tie by volunteering himself as the most deserving of being eaten, providing the others a better justification to society for their actions than siding against a random victim. Knowing this outcome, no one would shout out a suggested victim in the first place. Even under the axiom that future judgement doesn't matter, only survival, there is no reason the callee wouldn't suggest the caller, so the caller would at best make themselves one of two possible victims, so no one would choose to do that.
Yes, Albert has revealed himself to be an amoral schmuck who's willing to throw an innocent person under the bus to increase his chances of survival. Why would you want to keep him around in a life-or-death situation?
This is a big weakness in Scott's piece.
That was my reaction as well, although I think the more general principle still holds in cases where proposing a Schelling Point is less clearly villainous. Scott used a very similar lifeboat scenario many years ago as an illustration of dynamics at work in a diplomacy game where the players were five SAIA Visiting Fellows, plus me (*) and another one of Scott's online friends who was local to the area and got pulled in to round out the group of seven. Once the game started, the five SAIA Fellows mostly followed through on a prior agreement to team up against the two outgroup players and then declare a draw. Unbeknownst to Scott (and rather strengthening his interpretation of the dynamics at work), me and the other non-SAIA player (let's call them Alex) also teamed up on the basis of ingroup bias, since Alex and I were and are long-time friends. One of the five SAIA players wavered somewhat from their agreement, but came back around once it became clear that Alex and I were going to lose.
Scott described running that diplomacy game as a mistake and the game as being one of the worst he's ever seem. I respectfully disagree, as I had a great time. Alex and I were both much better tactical diplomacy players than the SAIA crew, so we were able have an enjoyable time going down swinging, and I had fun trying to split people off the coalition even if I didn't succeed. Moreover, it's my attitude that if you can't have fun despite losing for bullshit reasons, you have no business playing Diplomacy.
(*) If the character of Erica was named as a shout-out to me because of this, I appreciate it.
After Albert shouts: "Let's eat Bob!" shouldn't Monty say: "You still have no idea how Bob will taste! I'm now going to offer everyone a taste of Charlotte, just a morsel of her flesh, not enough to kill her. Pretty gamey, isn't she? Now that you've tasted Charlotte, do you eat Bob or choose someone else? Keep in mind that someone on this lifeboat will be delicious."
One thing that troubled me reading this is that while the lifeboat context is one where gains are at the expense of the excluded, a zero-sum game, it's not obvious that it's also the case with the backscratching context. To the degree that membership is an informative signal and backscratching is a positive sum game, nobody has to pay a price for the gains from collaboration. The benefits could be redistributed back to the excluded. Backscratching clubs are not necessarily evil.
I think that if you have to start figuring out the world from some empirical question, a good place to start is how often are we in a zero-sum interaction and how often are we in a nonzero sum game. Let's call it P(zerosum). My P(zerosum) is somewhere around 1%. That's what makes me a relatively nice person. If I saw evidence that it's higher than that it would change who I am. Any evidence that moves it up will make me more evil. Any that moves it down will make me more good. This is where morality and epistemics are conjoined twins. Conjoined twins who are screaming in agony.
>If there were an obvious self-recommending Schelling point, like one of them was a twelve-foot-tall green Martian
This would probably also yield more meat than killing a normal sized person but I'm no expert in Martian physiology
And they'd be weaker in Earth's gravity, so they'd be mostly helpless!
And their meat would be nice and tender, like wagyu beef 🤤
On the other hand I'm less confident that Green Martians aren't poisonous.
I'd always sort of assumed that they and the Red Martians were compatible that way, so that scavenger animals would eat both. It's been a long time, but did any of the Green hordes ever engage in cannibalism?
For the people debating which Schelling point would take precedence in the lifeboat example - I think the point is not to show that one particular Schelling point is a more likely candidate than the rest, but rather to show that some Schelling point will almost certainly be accepted within the group.
My problem with your hypothetical is that it rests on the assumption that everyone on that boat is trying to minimize their chances of being eaten (or maximize the odds of surviving) when in fact some (I really believe most) are more concerned about avoiding responsibility for killing another human being. Even drawing lots might be considered participation in the execution of another. I think historically, people don't draw lots, they wait to see who dies first, which absolves everyone of guilt.
In terms of sociology, it can be described by Bourdieu: identities and beliefs are symbolic capital invented/articulated by intellectuals as a tool to create a new group within a field. This new group then uses the tool as cultural capital to struggle for status with the other groups already there, some of whom may take the ideas as part of their own toolkit.
Wow, it's been over 2 decades since I read any Bourdieu, but that still made perfect sense. :-)
Having watched a dozen seasons of survivor over the past couple years, I have to make a point about the very first example (It doesn't undermine the main gist of the post). When Albert shouts "Wait, let's all kill and eat Bob!", the other most likely outcome is that Albert gets eaten. Has painted himself as a target. Even if he wins the first vote (this can be accomplished by shouting the least popular person's name), he will then become the biggest target on the second vote. This is the primary incentive not to try something like this.
Yeah, I've played a fair amount of Mafia and this dynamic is very real among inexperienced players. If you aggressively advocate for lynching someone day one you often succeed, but also put yourself in real danger day two.
Now, this may still be worth it because your goal in Mafia isn't merely to survive but to catch the mafia, so among higher level players the dynamic changes. But at the base level of people being put in a new and unfamiliar situation, speaking up and drawing attention to yourself is very risky.
First of all, people are adaptation executors, not utility maximizers, and nobody goes around consciously inventing such strategies (because that would make them evil). So we probably should look for stuff that worked in tribes with less than 150 people that existed for a million or so years of human history.
In such tribes you'd have a rich preexisting network of social relationships, like I want to borrow some flint spearheads from the uncle of my wife's niece, so I support his wife's father as the Chief of the Hunt. And if we win, that uncle will say a word for me and I'll also get some choice antelope haunches. That would be more than enough to kickstart an otherwise arbitrary factionalism, without any explicit or implicit rules about backscratching, purple ribbons, or whatnot.
Second, in our societies it works markedly worse, but I suspect that this, and not the median voter theorem, causes elections to be surprisingly close. Or maybe it is a more truthful restatement of the median voter theorem: in a tribe under no threat of extinction, people's main competition is fellow tribesmen, so the goal of any political movement is to invent some polarizing issue such that it alienates a slight minority of the people, then your faction kills and eats them and gets to propagate your politically adept genes.
The entire point of constantly inventing euphemisms for orphans and banning previous euphemisms is to alienate some people and kill and eat them. And some of us seem to be naturally adept at it, which should not be surprising because a million or so of years of evolution since language first appeared.
Finally, pursuing that angle makes a proper complement to https://slatestarcodex.com/2013/03/04/a-thrivesurvive-theory-of-the-political-spectrum/ which did an excellent job explaining how rightwing attitudes are optimized for tribes under a constant threat of extinction, but then kind of floundered and concluded that leftists must be, like, chill hippies? They are not, they are much more vicious actually, understanding the difference in terms of PvE vs PvP societies explains what we can see with our own eyes much better.
I'm not convinced that leftism (of whatever variety) is inherently PvP. It seems like the ideologies ignore PvP entirely, which may be why it often winds up covering up a lot of PvP - it's like the distraction in a magic trick. If everyone's looking over there, then you won't notice that I've killed and eaten Bob.
Of note is that the old fraternal societies were often a substitute for modern insurance. They would vet members to make sure that they were at least somewhat unlikely to be a useless drain on the organization, and then had a system for taking care of those that had suffered some misfortune.
There were of course also other angles (political organization, MLM robe and paraphernalia sales), but this mutual aid aspect was probably the most universal.
I was once a member of the local Grange (rather a moribund outfit these days, as many fraternal societies are), and they still ask at the end of every meeting - "Is any member sick or in distress?"
Mutual aid was a big purpose of fraternal organizations. Another purpose they served was providing what we'd now call "third spaces" that were owned and managed by the organization for the benefit and use of their members, kinda like a middle/working class counterpart to upper-class Gentlemen's Clubs back before the latter term became a euphemism for strip joints. Compared to their main competition in this role, bars and saloons, fraternal organizations would provide the benefits of a curated group of people to socialize with. Saloons also generally weren't an ideal place for non-drinkers to hang out, and their business model inclined them towards encouraging people to clear off and make room for other patrons when they're done buying drinks.
Although I may be overstating this. Fraternal organizations as substitutes for saloons seems like it would predict that fraternal organizations in the US should have boomed during the Temperance movement and Prohibition. But I just looked it up, and it sounds like they were biggest in the late 19th century and were well into decline when Prohibition went national. On the third hand, I also found while looking it up that at least a couple largish fraternal organizations (the Good Templars and the Rechabites) were heavily involved in the 19th century Temperance movement.
I have to say, I suspect there was a bit of racism involved in them, too. :-/
My family was apparently part of the Elks' club for a while, but I literally cannot remember any use we made of that other than getting access to a nearby outdoor pool. It was very well-maintained, and filled with nice people, as I recall.
Backscratcher's clubs in reality suffer from the Groucho Marx problem ("I wouldn't join any club that would have me as a member"). You don't want to join a backscratching club full of people less powerful than yourself (you'll just wind up doing them a whole bunch of favours and getting little in return).
A few ways out of this:
1. Instead of forming a backscratching club that pretends to be a sailing/ club, you form an actual sailing club where people bond over a shared love of sailing. Maybe a little bit of backscratching does occur but it isn't that major a part of the club's activities.
2. You enforce really strict requirements so that everyone is a roughly-equivalent level of elite. This is the model of, say, a London club. But of course it's tricky to start something like this from scratch.
3. The Freemasonry model -- you are super secretive about your club, and allow people to *think* that there's a whole lot of really elite people in it. Once an elite person joins and figures out that everyone else in his local chapter is just some random schmo, you tell them "oh yes, but if you keep going to our meetings then eventually you'll get invited to the secret elite circle where all the really elite people hang out".
All three of these map on nicely to the medieval guild system, which handles these issues through an internal ranking that rewards skill and effort.
1. There's bonding over specific professional crafts that were internally generally cooperative while competing with other guilds
2. There are strict requirements *for each level within the guild.* You'd start by being an apprentice, move on to being journeyman, then finally graduate to being a master member. In the best cases, advancing was fairly meritocratic.
3. They were elite and secretive, particularly at the highest level, which still gives some proto-prestige to the apprentices (it's relevant that Freemasons evolved from stonemason guilds).
"You don't want to join a backscratching club full of people less powerful than yourself (you'll just wind up doing them a whole bunch of favours and getting little in return)."
On the other hand, this is how you end up being the big fish in the small pond. Everyone owes you favours, so their families and friends and neighbours need to do you favours. So maybe that means you are now the only grocery store, pub, and undertaker in the town. Or you are the guy who employs everyone. Or you and your family are the elite of the small town. You get to wield a lot of power that way.
So you can always get preferential treatment and skip the queue when you want it. You employ Joe and his cousin Bill and their in-law Tom, so now when your mother-in-law needs her house painted, Tom's cousin Mike does it for you at a hefty discount. Things like that.
Isn't a lot of employment advice about the importance of networking? If you're starting out on your career, you don't have a lot of power and influence yourself yet, but if you know someone who knows someone who can get you that interview at Wiggins Widgets, then in turn you help them out with a call to your Uncle Bill and so forth. That's backscratching - why does the 'more powerful' person in this case help out a friend of a friend? Because of the web of mutual obligations and favours that can be called in later.
Isnt "Joining Backscratcher clubs is an advantage" basically the same claim as "Economic protectionism works"? Except without the fancy arguments about industrial policy.
And two backscratcher clubs could negotiate a free-trade agreement so that members of each could scratch each other's backs. Especially if the two groups are already coordinating against another backscratcher club (or another coalition of backscratcher clubs).
I think you would be interested in Laurence Iannaccone's work on economic models of religion, which touch on some of the themes here.
https://scholar.google.com/scholar?hl=en&as_sdt=0%2C40&q=laurence+iannaccone&btnG=
>But it’s easier if everybody comes color-coded so there’s one obvious coalition that occurs to everybody and which they can easily check that they’re a part of. And so on to nationalism, religious conflict, political ideology conflict, and so on.
I don't know, it seems to me like there is a substantial difference between--on the one hand--forming a coalition against short people because most of us are tall(er) and it's convenient to us and they're easy to spot, and--on the other hand--forming a coalition against the As Soon As We Gain State Power Either Electorally Or By Force We Are Going To Literally Kill Everybody Who Is Not A Member of Our Party, Oh And Also We Are Not Accepting New Members So If You're Not In The Party Right Now We Will Kill You Party because if you don't form a coalition against them they will kill you.
I guess you could argue that it's just a difference in degree or something, but It just seems like there are other differences as well. Or is it simply that my second example needs to be categorized as a "later" move in the lifeboat game, one that only takes shape after e.g. Heather and Iolanthe's "let's eat everybody except Heather and Iolanthe" coalition has revealed itself?
Campaign to Eat Bob
https://spartacus.app/campaign/278093/sharable
Or go to spartacus.app and enter code 278093
(Brought to you by ACX Grant 2024 Awardee Spartacus.app, a web app for solving coordination problems)
The Lifeboat Games thing reminded me of this from SMBC back when they did videos: https://www.youtube.com/watch?v=XMEV8DHc6E0
One think I would add about the calculations people make to join backscratchers' clubs is that they are mostly subconscious. People might consciously believe they're doing things for noble reasons, but beneath the surface a lot of human behaviour is driven by a desire for status.
Yes, I keep coming back around to Jonathan Haidt's theory in "The Righteous Mind", that our ability for rational thought is a by-product of our ability for rationalization and persuasion, to make it easier to form coalitions against other humans.
I think most societies have advanced to the point where they develop immunity systems to the Alberts and Backscratchers Clubs of the world. Eg, by voting to kill Alberts first.
To some extent, opposition to (perceived) Backscratchers Clubs can be seen as one of the main stated reasons behind animosity towards merchant minority ethnic groups (Chinese in Southeast Asia, Jews in many places, Igbo in Nigeria, etc).
"Any coalition of 51%+ can do this. But it’s easier if everybody comes color-coded so there’s one obvious coalition that occurs to everybody and which they can easily check that they’re a part of."
If you find this interesting, the book to read is
https://www.anarchonomicon.com/p/cocytarchy
which describes how and more importantly WHY this plays out exactly as it does in the very real life and death stakes of maximum security prisons, and why prison gangs are all race-based.
" But a third reason - linked to the second - is as cover for a backscratchers club."
On the one hand, this is the way we of the 21st C have been primed to interpret everything.
On the other hand, Robert Trivers and _Mother Night_ both have a point...
The best way to fool someone completely is to start by fooling yourself, ie by really really believing your story. And if you really really believe your story, and your story is that you do various good things, then, uh, what exactly is the issue?
I think the 21st C has swung the pendulum way too far on the side of cynicism, to the point where we simply cannot believe that people would ever do anything nice for anyone. This is even seeping into the most banal of entertainment where one of the features I've noticed in recent movies as opposed to say movies of the 70s or earlier, is what "heroes" do is go out and save their families, not strangers. A 60s movie would have us believe that someone might be willing to risk their life for the sake of strangers or an idea; a 2010s movie cannot imagine such a strange possibility.
And I'm not sure this makes us any the wiser about humanity; it's just another form of WEIRD. We haven't become smarter about understanding or modeling the human condition, we've just enforced another version of our culture on the world and claimed that it both is and should be universal.
Relevant real world dynamic: I've known a businessman who was a Liberal party member to say he would always stay a part of the Liberal party regardless of their policy positions, purely because of the business connections he made through being involved in the party.
The Liberal party is not purely a backscratchers club for business people (the most powerful people in the party are dedicated political animals many of whom have never worked in business). But being a pro-business political party is a great Schelling point for it to serve that function along with its other purposes.
I'm just curious, which country's Liberal party is this? Considering you describe them as pro-business, I'm guessing it might be Australia rather than Canada, but too many other countries have Liberal parties at various positions on the political spectrum to be 100 percent sure.
Yep, it's Australia. Sorry, I feel self conscious about going AS AN AUSTRALIAN constantly so I omitted that context. :p
I might suggest “Member of the liberal party, here in Australia”. It’s a bit annoying because Americans don’t have to do that but it’s there world right now.
"the stories in this section seem suspiciously like the Establishment"
I don't know if that's quite right.
I've stated in a few places that the way I think this works is that some genuinely smart person discovers some heuristic that's useful - it makes life more pleasant, it helps with science/technology, it helps you understand people, whatever. Then what?
The inevitable pattern is that various other genuinely smart people take up the heuristic because it is genuinely useful. BUT other midwit people see something else in the heuristic, that it can be weaponized, that people not using for whatever reason can be demonized.
This is *always* possible. Look at the petty hatreds that exist within a group of "technology enthusiasts". Maybe you imagine these people spend their time actually educating each other about how neural nets work, or how to simulate branch predictors. But no, 99% of their interaction is screaming matches about why Apple is better than Microsoft, or RISC-V is superior to x86.
Most people's favorite thing in the world is to hate someone else. And so when any new idea comes along, the first question is: "how can we use it to hate someone?" There's no difference between hating someone because they believe Christ is of one nature when of course we all know he is of two natures; and hating someone because they believe homelessness can be solved by providing housing when of course we all know it's a problem of mental illness. In both cases it's UTTERLY UNIMPORTANT what the underlying reality may be -- there may be no such underlying reality, or something not amenable to empirical testing, or both sides mean something different by various key words. But the point is not a difference in words, it's I have my tribe and you have yours.
Point is, your model is that DEI/Woke succeeds as an ideology because it provides an uncoordinated way for a group ("the elites") to scratch each others's backs. My model is that that's a convenient side effect, but it's not primary. Hatred is primary, the fact that it can be channeled to benefit some of the elites is just a spandrel.
If you look at this in rational terms, DEI/Woke is not improving the overall wealth/capability of the US.
It's not even improving the overall wealth/capability of the DEI/Woke elite as whole, not long term, not once the bills start coming in.
It's not even improving the relative social standing of DEI/Woke elite, since the only social standing they care about is within-group, and that's a fixed sized pie.
But what it IS doing is providing a steady stream of hate-fueled dopamine every day. The Two Minute Hate is one of the more overlooked elements of 1984, but to my eyes one of the elements that rings most true.
So how do you avoid this, the morphing of any good heuristic into a weapon of hate? I have no idea!
One element that may help is having strong restrictions on who can join your club. If you have to demonstrate real mathematical ability to become a physicist, then while physicists as individuals may hate each other, the system as a whole most cares more about physics and truth than about those hatreds. Something like that COULD (maybe?) save Social Science if we were willing to say that the number of Official Social Scientists needs to shrink to 1% of where it is today, and those 1% need to demonstrate serious competence in *something* difficult (math sure, or historical languages, or neuroscience, but *something* beyond "I is interested in peoples").
Another element that may help is having a conservative establishment that is tasked with manufacturing consent, with others not getting much of a word in, not until they join the establishment. That's been part of the traditional story (blown apart in Europe in the Reformation, then in a different version in the US in the 60s). We're supposed to believe that manufacturing consent is a terrible solution. I'm not so sure; as always the comparison is not with some perfect world of your imagination but with the actual real world of everyone empowered to say whatever craziness they like.
Scott's story is that "The Establishment" is doing this to ensure that they have better lives. In the case of the establishment as DEI/Woke, I've already said that I think hae motivation is more hatred of perceived enemies than back-scratching. But for other Establishments, like eg US WASPs, or many nobilities for many centuries, I think there is a genuine element of appreciating just how bad chaos can be in the absence of manufactured consent, and just how valuable it is to ensure that civilization is maintained, by whatever fictions and even occasional draconian punishments are necessary.
I think the conditions you're describing only apply under conditions of material scarcity. When there's enough to go around, greed and novelty-seeking can beat out hatred. Thus, I am optimistic how things will go politically once the renewable energy transition is complete.
Interestingly, you've written about this topic in a complimentary way before, Scott. Through a very different lens:
https://slatestarcodex.com/2017/11/13/book-review-legal-systems-very-different-from-ours/
Scott seems to have independently re-discovered Curtis Yarvin's conception of "The Cathedral" in his description of elite ideology as a decentralized but extremely effective backscratching club.
I'd say that this model has more gears:
https://www.lesswrong.com/posts/B7P97C27rvHPz3s9B/gears-in-understanding
I'm not sure if the Yarvinian Cathedral actually requires any actual backscratching, ie. concrete acts to aid one another.
Very interesting thought exercise.
The lifeboat scenario makes me think about tyranny of the majority, and the need for a “marooned lifeboat passenger bill of rights”.
And the later stories remind me of the virtue signalling that now pervades thinking on both extremes of the political spectrum.
The best solution to the lifeboat problem has to be the nose game.
It has the urgency of shouting random names but it doesn't require choosing anyone else's fate. Albert, who touches his nose first, can secure his survival without making any enemies!
I favor Scenario 3 as a response to Iolanthe's orphan adoption plan, because the plan if successful would exterminate the population of orphans and thus be a horrible genocide.
How arrogant is Iolanthe, anyway, to think she can save orphans when all these important credentialed people have failed to solve the problem for decades?
This situation where if everyone turns on one person, he has to go seems kinda reminiscent of a current political thing...
This lifeboat scenario reminds me of a British quiz show called the Weakest Link. (There are probably similar quiz shows in the US and other countries.)
The quiz starts with a dozen or so contestants standing in an arc, each behind a podium with their name on the front, and a host standing before them fires questions, apparently at random, at individual contestants.
The quiz is in a series of rounds, and at the end of each round, the remaining contestants write with a marker pen on a board the name of one other contestant who they propose to be evicted from the game. After everyone holds up their card to reveal the name they have chosen, the host interrogates one or two of them at random. Some of these questions are about their personal life, but end with enquiring why they voted for who they did. The host then briefly assesses a majority decision and tells the relevant contestant "Goodbye, you are the weakest link!".
Broadly speaking, the strategy adopted by most contestants is that in the early rounds they vote off the genuinely weaker performers. But towards the end, when the number of remaining players has dwindled, they vote strategically to send off the strongest contestants, because these of course would be the most challenging opponents in the final couple of rounds if the voter stays in the game that long. There is also some tit-for-tat revenge voting, as reciprocal payback for previous votes.
The final round is a simple five question play-off, and the single winner at the end receives a token prize. Everyone voted off in earlier rounds receives nothing.
If I remember correctly (it‘s been a long time since I watched this), the prize money was determined by the number of correct answers, and that provided an incentive against voting strong contestants out, resulting in an interesting conflict of incentives. When you say „token prize“, perhaps you mean that the achievable prize money was quite small compared to other quiz shows, like Who wants to be a Millionaire. Anyway, I think it‘s a good example, valuable material for students of Schelling lifeboat situations.
Yes, you are correct. Within each round there is a cumulative money "pot" which starts at zero and increases by some set increments (of the order £100 - I forget the exact figures) for each correct answer. At any time any contestant can shout "bank" to consolidate into the final prize total the amount so far added to this pot and reset the pot amount to zero. It is also reset to zero for any incorrect answer.
Actually, the pot increments may themselves also increase, so there is an incentive to hang on for a while and not simply shout "bank" after each correct answer. But of course, set against that is the risk of leaving it too long and blowing the lot with a wrong answer. Also, I think any money left in the pot when the round ends is discarded. So there is that to take into account as well.
In the early rounds, in which the questions are easier, the total prize money is easier to increase, provided timely "bank" calls are made (and contestants gain kudos for this - It must be easy to forget the banking aspect when concentrating on the questions). But towards the final rounds, the questions become harder and the rounds shorter, due to the smaller number of contestants left. So the pot is harder to accumulate and the amounts smaller, and thus I would have thought the banking in the later stages becomes relatively less important than trying to eliminate stronger contestants.
“This is obvious, right? There’s no downside to being in the club, and the upside is preferential treatment from all existing club members.” Someone has probably already mentioned this, but this isn’t actually true. You are neglecting the cost of having to favor other club members. For example, if you are an employer, having to hire a club member instead of the most meritorious candidate will hurt you and leave you at a disadvantage relative to your competitors for customers’, as long as they aren’t in the club and don’t have to bare this cost themselves*. In fact, since everybody in the club is subsidizing everybody else in the club, if you add the net gain to all members and subtract the net loss, it should come out negative (it’s not 0 due to loss of market efficiency, the employee can’t gain more than the employer loses from being hired, as otherwise they could do a deal where the employee could reimburse the employer for his loss assuming 0 transaction costs, but the employee can gain less than the employer loses)**. Of course, some individuals will gain more from membership than they lose, and some will lose more than they gain, but on net it should come out negative. Of course, this is relative to the counterfactual where there was no club, conditional on their being a club, it may be costlier to stay out than to join it even if you lose more from the club existing than you gain. This is especially true if nearly the entire world is in the club, as than the benefits of staying out are miniscule and the costs are huge, but if the club is small, the costs of joining will be huge, and the benefits miniscule. Of course, if you know everybody will join the club in the immediate future, you would have to be stupid not to join (immediate here meaning the time you spend in net loss is short enough to be outweighed by the future gains of joining early).
“another justification is just “We’re the majority and they’re the minority, and if we all band together to profit at their expense, it probably goes well for us.” Any coalition of 51%+ can do this.” I suspect you know this already, but my inner pedant compels me to mention that being the majority is only relevant in so far as it means you have greater power, and It’s possible to be the minority and still have greater power, e.g nobles in medieval Europe, or whites in Apartheid South Africa. And even if you are the weaker faction, it’s often nevertheless better for you to all band together instead of staying uncoordinated e.g blacks in Apartheid South Africa. And of course you can have several factions fighting for resources and all benefiting from coordination among their members e.g casts in India where I live.
Regarding preexisting models like this, your model is a lot like how I imagined the racial cartel model of discrimination in economics worked, though this was just guessing from the name, as I’m not very familiar with that model.
*The employer employee transaction is just an example, and the same dynamics will apply to any other market transaction where a club member favors another club member.
** this is of course ignoring that assuming 0 transaction costs, if you lose more from being in the club compared to the club not existing, you can just pay other club members to dissolve the club, or not form it in the first place. Though in the real world, that will get you accused of disloyalty and attacked by other club members, since real people aren’t Homo Economicus.
*** in my defense, I do have enough self control not to spew forth on the solution to the oddness of Albert being able to save himself, seemingly without any new information, or how to solve having 8 different shelling points, as I think you know the answer to those, and game theory unconnected from sociology isn’t really the point of your post.
An important consideration I foolishly over looked is that you can coordinate with other club members to extract a larger surplus from non club members. E.g, if you and all your competitors are club members, you can coordinate to form a cartel to force consumers to pay a higher price, as long as you treat non club member consumers worse than consumers who are club members. You can trust your competitors because they would likely get kicked out of the club for betraying a club member in good standing, especially if only one competitor alone betrays you and the other competitors, and might face additional penalties if a club required members to go out of their way to punish traders. So there is no guaranty that the net gain to all members minus net loss to all members from the club existing will be negative. It could be either negative or positive.
The essay is insightful and beautifully written, but I think it misses a crucial point: solidarity.
The emphasis on "backscratching" downplays the more profound concept of mutual support and shared responsibility. When we talk about "restricted circles of extended trust", it's not solely about prioritization but about mutual aid, communal support, and collective well-being. Belonging to the same "club" means sharing a common moral framework and duties, not primarily (if at all) indulging in unfair favoritism.
Just as we naturally treat our families differently from strangers and share a sense of shared destiny, we extend this care to some broader communities. Many of the "hurdles" societies impose on their members are not just about gatekeeping, but ensuring people take these duties and moral codes seriously and deserve the automatic trust group members share.
So, the cynical view that "if everyone joins the backscratchers club it loses its purpose" is flawed, assuming "backscratchers club" stands for a social/ethnic-group. If everyone adopted a universal sense of solidarity and care for all humans, it would create a utopia, not a failed scam.
"the stories in this section seem suspiciously like the Establishment - whether it’s the Catholic Establishment of the Middle Ages, the conservative Establishment of mid-20th-century America, or the progressive Establishment of today. "
These things do not sound very similar to me. The Catholic Establishment (Church plus Divine Right feudalism?) had explicit, legible hierarchies that defined their power. That isn't true of the more modern establishments.
I also think that the modern "Establishments" don't look as much like the OSCs as you believe. Perhaps you're just defining "Progressive Establishment" narrowly around some parts of academia and non-profits. But if you take "the Progressive Establishment" to mean, like, main stream US culture, then I think the concept of OSCs are unnecessary.
What's the distinction between "Corporations and other organizations that want to have mass appeal end up signaling that they like things with mass appeal" and an OSC? Live sporting events regularly feature "Applaud the Military Veteran/Police/Teacher" breaks. Is this just OSC shit? Are those groups all part of The Progressive Establishment? Budweiser will make rainbow cans of beer until it becomes unprofitable to do so. Corporate decision makers will invest in DEI as long as it makes them more profitable by helping them recruit/retain talent more cheaply. Are they all part of an OSC that supports DEI only to leave that OSC and join a different OSC that thinks DEI is bad when the EV calculations change?
The concept of OSCs may be useful in explaining collective action in some circumstances within some organizations, but I think you're generalizing too far.
So, after Daniel founds the first club and everyone joins, what I thought was going to happen--and what I still think should--is this:
1. Someone in Daniel's club (maybe a late joiner and/or a bit of an outcast already) says "screw it, I'm leaving this club and starting my own rival club". In this Other Club, there will be a strict ranking order, based on the order people joined. So this guy, call him Oscar, is king, since he joined first. Except...
2. Does anyone else join? After Oscar makes his announcement, lots of people surely realise this is a golden opportunity to be second or third ranking in the Other Club. But that's meaningless if nobody else joins: you'll just end up being Oscar's personal servant, and ostracised by the rest of the town as well.
3. So everyone considering jumping ship for the Other Club will pause, and look around. If nobody else is moving, and is just waiting and watching as well, then best to stay put. It's even less likely others will join after you do, since the second-ranked position is taken. So you shouldn't join either.
4. But wait. What if you loudly shout "I'm joining the Other Club before you do!"? Then either lots of others make a snap judgement that there's a race to join, and you get the second spot *and* are closely followed by hordes of others who will all be your servants...or hardly anybody follows, and your position is almost worthless. After all, why should anybody take *your* intention to join as reflective of anything but your own weird preferences? It's not like you're some symbol of public opinion.
5. I think there are two solutions to this problem. The first is for Oscar to secretly pay a bunch of people to quickly, one after another, announce their intention to join. Then he creates an illusion of a race to join, everyone else is terrified of being left behind and joins the "race", making it a real race, and Oscar gets a huge populated club where he's king.
6. The second is for some high-profile person who is a natural Schelling Point and a perceived representative of the public to declare that he or she is joining, because he/she knows that everyone else wants to. If lots of people trust this person, or even more importantly if lots of people *think* lots of other people trust this person, then there's a race to join, and again Oscar is king.
The first solution is like an astroturf campaign, a centralised attempt to create the illusion of mass opinion. The second is, it seems to me, basically *the* answer to the question of why politicians' official positions matter so much. Despite the fact that they can and do abandon or reverse them once in office, despite the fact that many of their positions are on things they would have little or no power over in the office they're running for (e.g. presidential candidates taking positions on state laws, on purely cultural questions, and even on hypothetical constitutional amendments) and despite the fact that they may have literally taken the opposite position a year ago...it doesn't matter. The mere act of them publically taking that position, as a widely-perceived symbol of public opinion and the most natural Schelling Point for social coordination, is of *massive* importance, very often decisive, in shaping the direction society goes on that issue. And the people who stand to gain or lose from that position thus have an enormous interest in sacrificing lots to get "mere" lip service from a prominant politician.
(I also hope this description might make rationalists less dismissive of the importance of elections, and less likely to think social changes are either inevitable or random results of meaningless cultural events--like the multiple people blaming the rise of wokeness on teeagers on tumblr (?????)--on the basis of "but politicians have so little actual power".)
Various alumni organizations act as backscratchers’ clubs. The most extreme example is most likely the YC founders alumni group.
The club acts as one of the perks of going through the educational course. Why go to an Ivy League college? Well, one reason is for the networking. Aka backscratchers’ club.
https://timharford.com/2024/06/cautionary-tales-the-revenge-of-the-whales/
Lets say that there is a sub-community of about ten thousand people all of whom are wealthy shareholders of Fortune 500 companies. These people are not all personally acquainted, but they are linked by overlapping cocktail party invitation lists, so networked communication between them is reasonably cost-effective, if a bit slow.
One year the economic growth fails to meet expectations, and share prices start to fall. Everyone is casually acquainted with stories about the Great Depression, and they all remember the last recession, so they begin whispering about each other behind each other's backs. Someone proposes that they pick a major corporation randomly, and short their stocks, so that the rest will still have access to sufficient credit--
No. People in real life do not act this way. What is the lifeboat supposed to represent? If the entire national economy, then within the lifeboat certain people have hoards of food of varying sizes, and use that hoard to "hire" bodyguards. The choice of who to eat will be anything but random. While elites are taken down by other elites every day, that's more or less interpersonal competition, not some sort of coordinated game play.
Or, is the lifeboat the elite community? In that case, they will be surrounded by a large number of smaller boats, and will pick the hapless inhabitants of one of them to play "Andes Survivors" with. A rising tide lifts all boats, but it doesn't change who lives on a yacht and who lives in a rowboat...
And then, what happens when the little people, who outnumber the elites by at least 2 orders of magnitude, decide they've had enough and start digging out the pitchforks? Which computer simulation predicts the October Revolution?
An obvious Schelling point for the second person to eat would be Albert, who established the idea of coordinating around an arbitrary victim. (Ob course, if Albert knows this, it might be in his best interest not to propose the idea in the first place.)
---
I am not a lawyer, but I think that both legally and morally, there should be a vast gulf between people forming a pact to eat a person chosen by lot and people just coordinating a murder to save ones own skin. Just try for murder and hang anyone who ate someone whom they did not have a pact with. What is society for, if not to provide disincentives for behavior with negative externalities? (Of course, the murderers might also coordinate to cover up their crime, but even a slight probability of failure will make the lot-drawing much more attractive by comparison.)
Relevant Wikipedia: https://en.wikipedia.org/wiki/R_v_Dudley_and_Stephens#Legal_background_and_theory
---
I would argue that going along with eating Bob is not the correct solution for everyone (except Bob).
Consider Charlotte: if she is is smart she might notice that
a) one kill will likely not be enough to sustain the survivors until rescue
b) she is next in line
So it might be in her best interest to push for the drawing of lots instead.
Suppose you are n survivors, and you estimate that it will take k killings until you are rescued. Also suppose there a c charismatic persons who are unlikely to be eaten before the rest when people coordinate, and you are part of the rest r=n-c.
If you go along with the coordination once the first name is proposed, the probability that you will be among the survivors of the rest group is:
(r-k)/(r-1)
If every killing is decided by lot, your odds would be
(n-k)/n.
For example, if you estimate that the n=10 survivors will eat three people (k=4), and you know that there are two people who are unlikely to be picked before you (c=2, r=8), but don't know anything about your place in the food chain (other than you not being first), then it would be slightly better to insist on the lot (p=0.6) vs coordination (p=0.57)
One point that the post fails to adress is that there is need for actual justifications a la : "We are the 51% etc" when it comes to racism and so on, that justification is inbuilt. Same goes for the orphan-supporters who discriminate against non-orphan-supporters. Human tribalism is hardwired. Evolution has done the game theory for us. The only purpose for justifications is to rationalize the back-scratching that is already happening anyway. (That hardwired back-scratching is also much more messy and inefficient than purely rational back-scratching would be.)
> “The more complicated versions sound like cults, religions, and ideologies.... I’m sure real sociologists have written about these issues, but it was hard to find them and I figured I might as well write this post without citations. Still, if you know who they are and what their terms are, let me know.”
Wow, something I actually know about!
Rodney Stark is the sociologist of religion who pioneered this the “backscratchers club” model, and brought rational choice theory (basically, Econ 101) to religion. Among other things, he created an analytical model to predict which new religious movements are likely to succeed or fail. It was based originally in the studies he did on the Bay Area cults of the 1960s and ‘70s, but he later extended it to try and explain why both early Christianity and Mormonism took off the way they did.
Probably the best place to understand his model is his 1996 article “Why Religious Movements Succeed or Fail: A Revised General Model”. Especially relevant here are factor #3 (“medium strictness” — too demanding to let in free riders, but not too demanding to keep out all but the weird fanatics) and factor #5 (“a religious labor force” — joining a group that demands a lot of your time can counterintuitively be a great way to be lazy).
The article can be found here: prem-rawat-bio.org/academic/stark1996.html
I’ve (oddly coincidentally) been talking with Eneasz about how religious communities work on the podcast “The Bayesian Conspiracy” recently; he’s been interested in which of their tools we might be able to borrow to cultivate community in the Rationalist space: thebayesianconspiracy.com
My immediate reaction to the first situation was that the second time it happened I'd shout "Let's Eat Albert". That was pure emotion. Albert is a bad person, and I'd expect others to shout it too.
Once we'd got rid of Albert and Bob and I'd had time to think, I'd position myself as someone who's very very in favour of fairness and justice and uncompromising about the principle of actually drawing lots. And then I'd show myself prepared to fight to defend that principle.
That's calculated, but I think it's my emotional reaction too!
Once home, I'd join the various backscratcher clubs if and only if I liked the activities and people involved, at which point I'd become rather dedicated to the club and passionate about it. If I didn't like them, I'd doggedly refuse to join even when it had become obviously a really good idea to join and thus I'd lose out quite badly but keep my arrogance and pride intact.
This is both my emotional reaction and, from a lifetime of observation, how I actually behave in such situations. Calculation tells me to join in earlier, but screw that.
I wonder why nobody cited the lifeboat sketch. Everyone is sick and tired from decades of Monty Python quotes, probably. Anyway, on my lifeboat, I'd found the four-course menu coalition.
I think the answer to how well any of these strategies/clubs will perform often depends on what time depth people are thinking about. This is essentially a Rawlsian point, I guess: if you can get people to think long term, including those future periods when they will not be part of the winning coalition, then they will choose the fairest option. When people think short-term only, unfair options seem more promising to them.
Therefore:
(1) Lifeboat debates are a good way to crystallise this problem, because there's an incentive to think in the short term: you think you'll be rescued reasonably soon.
(2) Humanity now engages in many projects that require timespans longer than a human career/lifetime. Those inevitably create political problems, because relatively few people are going to be good at thinking that long-term; relatively many people will be willing to sacrifice fairness or advantages beyond the term of their own life in favour of short-term gains.
(3) Our ability to think (effectively) about the long term is often very highly constrained by our experience, so teaching history is really important. We won't all experience insidious political unfairness, but we can all learn about it in the classroom.
Interesting story! One question that comes to mind is, is effective altruism a backscratcher’s club?
It seems to me that it definitely wasn’t in the beginning (there were too many calls for sacrifice for that to be plausible), but I could see an argument that the resources of Open Philanthropy and subsequent professionalization have created a kind of Schelling point.
The problem is that (unless I misunderstood something) this presents examples of Backscratchers Clubs that are mostly borne out of simple greed and wish to conspire for mutual benefit, and actual Backscratchers Clubs genuinely seem to mostly come out of other, "higher" pursuits with the backscratching an eventual added emergent behavior.
I.e. at least if we go by the official explanation, Freemasonry really came out of bored gentlemen first getting interested in the Mason guild rites and then self-development, and the whole conspiring for mutual benefit arises after that, and even there the conspiring would probably quite organically develop from things like having a shared space for a lot of people who might indeed derive mutual business benefit from each other and with a belief that they have, indeed, developed themselves to a higher moral level. Same for religions and so on.
The one type of Backscratchers Clubs that genuinely follow the model are criminal conspiracies, but for the rest of them, one might imagine a club where people are genuinely first interested in scratching each other's backs like, a lot - maybe it's a furry thing? - and only after that's been going on for a while they start talking about backscratching deals of other sorts.
This is NRx redux, and since you read a lot of NRx, you must be aware of this (but does your audience know?).
Fairly sure that it's not NRx that innovated the concept that people might belong to religions and ideologies for personal benefit, originally.
Well if that's all the article said, I guess I would not have made that comment.
After reading the book review about prions, it got me thinking about this post and the danger of disease transmission via cannibalism. Of course Scott doesn't go into the weeds of disease or calorie counts because it detracts from the metaphor of the lifeboat to backscratchers clubs, but I find the application interesting.
Suppose everyone is about to draw lots, and Bob shouts out "I have HIV!" Bob has now created a Schelling point that separates him from everyone else in a negative way - nobody would want to eat the guy with a serious transmissible disease. I don't know how much weight this carries for people actually starving to death, but surely it matters when there are nine other people on the lifeboat.
The next step - does everyone else try to claim some sickness that spares them from being eaten? There is a fine line here where the disease has to be bad enough that no one wants to risk being infected but also has to be believable in the absence of obvious symptoms. I don't think anyone else would be convinced if you also claimed to have HIV - what are the odds two out of ten people on a lifeboat have it? Nor would they believe you have Ebola without hemorrhagic fevering - or maybe they do believe you and toss you over the side. Shouting "I occasionally have symptoms of HSV-1 infection!" is not serious enough to prevent you from being eaten. So claiming disease has to be both serious enough that no one wants to eat you, but not serious/infectious enough that everyone wants to kick you off the lifeboat for their own safety. And it also has to be convincing. After Bob does it first, everyone else runs the risk of being seen as an insincere mimic.
I'm not sure how this strategy plays out in the long run. Maybe everyone makes similar claims, and it becomes pointless like the backscratcher club where everyone is a member. Bob might still be safe, because his initial claim carries more weight. Then again, maybe no one believes Bob's conveniently timed confession. Anyway, something to think about the next time you need to strategize how to avoid being cannibalized.
Does this relate to the mystery of why so many dislike EA, when it's just good people doing good things? The answer would that making a noise about how charitable you are is interpreted as "look at the gig leaf, not what's underneath" .
I don’t think you have to look any further than the well-documented phenomenon of do-gooder derogation to explain that; and from my limited understanding, the research on that phenomenon points to the idea that do-gooder derogation arises because do-gooders make others look bad.
The problem with Erica and to a lesser extent the later orgs' models is the lack of a clear enforcement mechanism against internal defectors. Each club member wants to be favored by others, but not have to favor others. Without clear rules on what that means and procedures for removing members that don't follow them, this can't work. Which is maybe cancellation, in the analogy?
The link in "The Ideology Is Not The Movement" is redirecting through google
Lifeboat Games and Backscratchers Clubs sound like interesting concepts! They seem to focus on mutual support and creative ways to help each other succeed, which is fantastic. Just like in the Daman game, where strategy and teamwork can enhance the experience, these clubs could thrive on collaboration and smart planning. Building connections and supporting one another is always a winning strategy, whether in gaming or real life. If these ideas involve sharing resources or skills, they can create a strong community. Best of luck with your Lifeboat Games and Backscratchers Clubs—may they bring success and fun for everyone involved.
Read More:- https://damangame.tech/