A Cascade Of Dunbar Numbers

Back when I used to play at micronations, we had grand dreams of creating a fully-fleshed out simulated country with hundreds of citizens and dozens of Byzantine government departments and industries. For a while, these dreams seemed attainable. Most of the people we advertised to were really interested, and we went from one or two founding members to about a dozen people very quickly.

Then we stopped. For the past fifteen years or so, the population of Shireroth remained fixed at between ten to twenty people. It wasn’t always the same ten or twenty – some new people entered, some old people left – but it always stayed within that window.

Ten years ago we got really tired of this and spent a decent amount of money on an advertising campaign. It was a good campaign, and we got maybe thirty or forty people to apply for citizenship at once. We thought this was going to be so awesome – the size of the simulation quintupling within a week. Instead, the new people just kind of wandered around aimlessly, bumping into things, never quite figuring out what was going on, never quite taking an active role in anything or showing any initiative, finally showing up less and less until two months later the population was back down to about a dozen people.

Our Ministry of Immigration and Naturalization, which was in charge of recruiting new people, kept beating its head up against the same sort of idea with about the same results. I thought this was dumb, so I manuevered my way into the Ministry and shifted all of our resources from recruitment to retention – setting up mentoring programs for new citizens, writing a bunch of guides for what they were supposed to do, trying to befriend them and integrate them into our social scene. The result was that our new immigrants maybe stayed three months instead of two months before evaporating.

We were only ever able to find one successful method of recruiting new micronationalists: wait for random people around the world to learn about micronations, start their own projects without any impact or influence from us, integrate those projects into the larger community, and then if those projects failed after a year or two sometimes their members would join us instead. This sort of worked, but never enough to get us more than ten or twenty people. And those other projects, the ones they made, very rarely had more than ten or twenty people either, despite hordes of people who said they thought micronations were interesting and agreed to participate for a few months. It was a weird problem, and one I was never able to solve before I left the hobby.

I was thinking about this recently because of some people’s complaints about the Bay Area rationalist community.

I have always had universally positive experiences with the Bay Area rationalist community. I arrived in about 2012 when it was still smallish, and I got to know the group involved and hang out with fascinating people like Eliezer Yudkowsky and Anna Salamon and Michael Vassar and learn a lot from them.

The complaints I heard were something like this. A lot of people join rationalist communities in Missouri or Italy or Australia or some far-flung place like that and really like them and feel like they have a great tight-knit community. They decide that if a small rationalist community is good, a ten-times-larger rationalist community will be ten times as good. Then they get here and everyone’s already doing their own thing, and it turns out Eliezer and Anna and Michael do not have enough time to personally interact with every single person in San Francisco on an intimate basis, and they kind of hang around the edges of the community not really knowing what to do or being connected to its general rhythm. It doesn’t help that a lot of people (DESPITE MY BEST EFFORTS) keep thinking that rationality will turn them into supermen and get upset that they’ve been in the Bay Area like an entire month and are not yet Elon Musk.

And I was thinking of both of these things when I met Samo in Oakland and he told me about his research. It’s kind of complicated, but when I asked him “Is this basically hypothesizing a cascade of Dunbar numbers?” he told me I wasn’t far off.

Dunbar’s number, remember, is the theory of anthropoligst Robin Dunbar that humans have a hard-coded optimal group size. He found that most primates’ brain size correlates with the average size of their social groups. If you extrapolate the connection to humans, you would expect humans to have a social group of about a hundred fifty people (other estimates find anywhere between 100 and 250). Various researchers and amateurs have claimed to corroborate this in every context from hunter-gatherer bands to corporations to MMORPG guilds (1, 2), sometimes more believably than others.

Samo said his own research had found several of these discontinuities – I think he mentioned about 12 people, about 150 people, about 1,000 people, and about 90,000 people as the first few. He tried to tie this in to government forms like “family”, “clan”, “tribe”, and “city-state”. I think the idea was supposed to be that we’re naturally very good at dealing with groups of family size, our brains will grudgingly allow us to deal with groups of up to clan size, with the low-hanging fruit in social technology we can stay together in groups of tribe size, with some higher-hanging fruit in groups of up to city-state size, and so on. When you get to the largest organizations in the world (China and the Catholic Church, if you’re wondering), you have to pull out all the stops – a pyramid of subdivisions of subdivisions of subdivisions, each choosing representatives to the next-highest level which then chooses its own representatives and so on until you reach one guy at the top. He suggested that if my old micronation or the Bay Area rationalist community want to get past their growth limitations, they should try something similar – which is good advice, except that in fifteen years of trying to simulate a country it did occur to us to have provinces and that didn’t seem to help much.

I am not sure how much I buy this theory. Having a systematizing disposition, it is really tempting for me to start carving up reality at these lines – startups with < 12 people, small businesses with < 150, et cetera - or to start noting that most of the college clubs I've seen have like a dozen core members, no matter how big the college is. But I find myself engaging in certain thought patterns that I have tagged as "trying too hard to fit data to a theory". At University College Cork the debating club often got twenty or thirty people; and if someone told me the natural way to slice businesess was <5 people, <30 people, et cetera, I could kind of see that too. And also there are so many other factors - whether a group is online vs. meatspace, whether it's living together versus meeting once every few weeks, whether it's strongly enforced (like a company or military unit) or voluntary (like a social club), whether we expect its members to also have other relationships somewhere else (members of MMORPG guilds presumably have real-life friends and coworkers as well, but members of hunter-gatherer tribes might know literally nobody except their cotribesmen). Even if there were interesting limits to the size of cohesive social groups, I would expect these factors to make them different in different domains - the size restrictions governing the growth of online micronations should be different from the size restrictions governing the growth of startups, even though in practice they're often around the same size. But I do notice that the Bay Area rationalist community is probably around 150 people and having trouble growing, and the size of my micronation was around 12 people and had trouble growing. And that after ten years in micronations, the rule that no matter what happened we would never get more than a dozen or so people seemed like a law of nature. And that I just checked after two years outside the community, and it looks like one new person came in to replace me and no one else, and they’re still stuck around twelve. So there does seem to be this very real community failure mode where everyone is socially saturated with each other and new people feel like they can’t break in. I’m less confident it always happens at a specific number or set of numbers across all domains of community – but who knows?

This entry was posted in Uncategorized. Bookmark the permalink.

234 Responses to A Cascade Of Dunbar Numbers

  1. hawkice says:

    The explanation I have heard for ~12 is roughly the following (which at the very least confirms that this meme is common): you can only know N people well, let’s say ~150. But in socially dense communities, you don’t want to just know the people, you want to know all the relationships between the people, or roughly N^2 relationships. 12 ^ 2 = ~150.

    I think the notion with the bigger groups is that you only need to know [something something] about people in a tribe or city-state. Not know them deeply as people, certainly. So they get a memory discount, as instead of knowing a character you know [garbage-man Kennedy, the one dimensional helping person] or something.

    • Scott Alexander says:

      The square number point is interesting, but I don’t expect knowing a relationship to require the same resource as knowing a person, and if you took that seriously then you could know those twelve people and no one else. Even when I was *really* into micronations I also had other friends.

      • OTOH, a complex diplomacy game seems exactly like the sort of thing that would be require people to model not only each other, but everyone else’s interactions with everyone else.

      • Irrelevant says:

        The Handshake Problem is (N)(N-1)/2, not N^2, which would suggest close-knit groups of 18. But I’d guess that’s a red herring and we’re actually dealing with a different limiting factor: the complexity of each of our individual model humans. At the family/squad level of collective relationship, there’s not a whole/part distinction or formal leadership, you understand the group by understanding everyone in it. But you can only understand each individual person so deeply, and around about 12 the group interactions are complex enough that you know each person in the group only as they relate to the group, and above that the collective relationship breaks down. (This is my impression based on having gone through several of those groups, and noticing the changes in my individual relationships with members as the collective relationship waxed and waned.)

        At the clan level, we give that up and instate some sort of formal structure we can relate to, and just deal with the members individually.

        • Peter says:

          Well, if you’re talking about a squad, then in the US army you’re talking about a leader plus two fireteams of four men each, each fireteam with it’s own leader, for a total of nine. And the next level up is the platoon, four squads plus headquarters (6 men), total 42 men – no Dunbar number there. The next level up, the company, seems to be near the magic 150 number.

          • William Newman says:

            Some organizations can have groupings that don’t have much to do with the ‘Dunbar number’ idea of too many people to keep track of, and instead have to do with too much technical specialization to keep track of. Militaries have this in e.g. artillery specialization vs. cavalry specialization vs. medic specialization vs. pilot specialization, and only some of those end up grouped together in units (e.g., a unit of pilots) at any given level. The IETF seemed to have a lot of this grouping by specialization when I did some work with them late in the last century: their most publicly conspicuous groupings tended to be arbitrarily large numbers of people grouped by interest and specialized knowledge (e.g. internet security, or mobile ad hoc networking), analogous to groupings of artillerists or cavalrymen, not analogous to squads or regiments. Tighter-knit Dunbar-number-ish working units mattered too (e.g., people working together to make interoperable prototypes of a new protocol) but they didn’t seem to be the face that the organization presented to most people.

          • CatCube says:

            @William Newman:

            The other limiting factor is “span of control.” It’s diffcult for a leader to keep track of more than 3-5 “subordinates” so, depending on the type of organization this will become a factor as well. For infantry, where individual people are your warfighting “pieces”, you have companies of around 150, so everybody knows everybody else, but they’re divided up into 4 platoons for command and control.

            Compare to an armor company, where your warfighting piece is the tank. You have platoons of 4 tanks each (span of control for the lieutenant), with 4 platoons (span of control for the captain.) There’s no natural grouping to get close to ~150, so they don’t really try.

        • imuli says:

          Agreed that it is probably a red herring, but this isn’t the handshake problem. Your model of Alice’s model of Bob is different than your model of Bob’s model of Alice. So that gives, n(n-1), at least. (Assuming we don’t get into my model of Alice’s model of Bob’s model of my relationship with…)

      • Deiseach says:

        The “initial twelve” number suddenly made me think of the Twelve Disciples, so this must be a long-established rule 🙂

        (How that went was 12 disciples, 72 apostles, unspecified number of converts during/after Pentecost, and big bust-up at Council of Jerusalem with St Paul who’d been off doing his own thing with the Gentiles eventually resolved in favour of St Paul after St Peter had a vision from God. But even during Jesus’ lifetime, there were people who dropped away because this is a hard saying, who can bear it?)

        With your micronation, I wonder if one barrier is that when the new people come in, they may be bursting with ideas about what they want to do in and with a micronation, but by now Shireroth is established and the Founding Fathers are more or less still around, so they (a) have the common newcomer problem of “I don’t know anyone here, I don’t get the in-jokes, I don’t know how things work, and when I try to change things, I bump up against Tom who gives the nod to Dick who has a word with Harry who closes me down” and (b) they may not be interested in getting into the level of politicking and involvement to change things, so they drop out and create their own worlds or give it up.

        It’s a similar problem with evangelisation 🙂

        • Deiseach says:

          This is why you shouldn’t try to quote Scripture first thing in the morning when suffering from an upper respiratory tract infection and your head feels like a balloon 🙂

          It’s the Twelve Apostles and the 72 Disciples, not the other way round.

          • peltast says:

            It depends. In the Eastern Church the Seventy (Two) are referred to as Apostles, so it looked right to me the first time.

          • Eric Rall says:

            I’m Episcopalian and have also poked at Catholic and Lutheran theology. My understanding was that you were right the first time, at least about the Disciples: I almost always hear that referring specifically to the 12 Disciples who followed Jesus during his lifetime.

            “Apostles” is an overloaded term. I’ve heard it used interchangeably with “Disciples”, and I’ve also used it to refer to the first generation of the senior Church leaders following Jesus’s death (the original 12 Disciples, minus Judas, plus Matthias and Paul; usually counted as “12 plus Paul”). And then there’s the concept of Apostolic Succession, adhered to by the Catholic, Orthodox, and Episcopal/Anglican churches, which holds that modern Bishops are successors to the original Apostles.

      • I could make up some bullshit reasoning about the number of people you can sit around tables, but that doesn’t make it true. We need some falsifiable claims here

      • Mr. Breakfast says:

        ” …if you took that seriously then you could know those twelve people and no one else. Even when I was *really* into micronations I also had other friends.”

        What if these various group size thresholds refer not to the number of people you can know overall, but to your maximum social frame size? Like you can only load and manipulate enough information to model so many people or relationships in your mind at one time, but you can store the information to create more than one such model as long as the modeled spaces don’t intersect too much?

        I know that before the advent of Facebook, I had parallel social existences for “going out friends”, people from school, people from work, family, etc. When I was in one setting or another, I would be one version of me or another, and I would be immediately aware of the personalities, relationships, and interpersonal histories of the people in that social frame. If my mother mentioned my great aunt, I would immediately picture the woman and her significance, history, etc. But if you had approached me at work and asked me about my great aunt, I would have been flummoxed by the very question and probably struggled to recollect the most basic things about her.

        How often did your micro-nations friends interact with your IRL friends? How often did a member of your family invite you to engage in dinner table gossip about them?

        I remember early Facebook and the rest of the “real name” internet being like a bomb exploding in my social life: Suddenly, my work network was invading my regular social life which was bleeding over into my extended family. My ability to cope with it all collapsed, and my social existence ever since has been radically depopulated and simplified. Now I basically have a work network and a non-work network. I suspect this has happened to a lot of people. Maybe throw in a interest-based online community (pseudonymity helps to fend off Google), and everyone else in the world has to stay on the “cardboard cutout” level of modeling just to stay sane.

    • Unless the relationships are two-sided (you treat A having a relationship to B differently from the relationship of B to A) you should only have around N^2/2 .

      • ShardPhoenix says:

        >Unless the relationships are two-sided (you treat A having a relationship to B differently from the relationship of B to A)

        Seems true to me in general.

      • HeelBearCub says:

        The wisdom of J. Geils captured it:
        “He loves her. But she loves him. And he loves somebody else. You just can’t win.”

        Knowing how I feel about 11 other people is not the same as knowing how they feel about me.

    • Gerry Quinn says:

      This idea reminds me of ‘House of Bones’ by Robert Silverberg, in which a time traveler from the present day becomes part of a tribe of people in Northern Europe around 15000 years ago. What reminded me of it is that people have names, but not unique names – everyone has their own name for everyone else! Which means, I suppose – at least in principle – that you have to know everybody by everybody else’s name for them, at leat if you want to engage in gossip…

    • dipitty do says:

      I propose basically reversing this. Yes, with 12 people, you will probably have some vague notions of “A and B like each other, but B isn’t so keen on C,” which probably isn’t so hard to keep track of. But in managing large collections of people, say, all the characters in War and Peace, remembering their relationships to each other is probably even more important than remembering exactly who they are. If you occasionally encounter Bob, it may be less important to remember exactly what sort of person Bob is as to remember that Bob is married to Carol, and Carol is nasty and you don’t like her. Everyone in Carol’s orbit therefore gets tagged “Carol” and treated as extensions of Carol, rather than as individuals. This might be a dozen people reduced to a single relationshipal attribute, allowing you to manage life among many people.

  2. Emeriss says:

    Would kinda be careful. Larger gardens require more gardeners, and even then there are more cracks, crannies, and shadows. Not sure whether increasing group size is strategic in the bay area context above.

    Eta words and speeling

  3. Herpaderp says:

    Use of the phrase “failure mode” is interesting, because having read a large portion of your writings, I’m still unclear as to what the goal the group would be “failing” at is.

    • I was wondering that too. I feel weird about the idea of intentionally trying to maximize the number of members. People already complain about the quality of SSC comments going downhill. I’ve never been to a meetup before, so maybe my opinion shouldn’t count for much, but I like the idea of just remaining open enough for people to discover the community organically.

      • Devilbunny says:

        When a group is small, it is possible to follow the entire conversation without a large expenditure of effort or time. When there are 500 comments on every post, keeping up with what’s going on becomes much more time-consuming, and socializing newcomers becomes much more difficult – they can’t reasonably be expected to read the entire archive, and yet the presence of large numbers of them is very disruptive to the established members’ conversation. I, for one, would happily attend Deiseach’s suggested catechism class rather than spend a year lurking to learn the ropes.

        • Okay, so who is in for setting up a catechism that new commenters have to go through before they can comment? Seems like a thing that wouldn’t be too hard to do.

          • Jaskologist says:

            I’ve long been a fan of the quest Kingdom of Loathing requires you to complete before you can access chat.

            The ghost of the English language looks up at you.
            “You have already proven yourself literate! Go forth, then, and avenge my death!”

      • blacktrance says:

        Data point: I went to the recent SSC meetup in Berkeley, and there were too many people there, and it got better as people started leaving. I’m relatively new to the meatspace rationalist community, though, and I probably would’ve enjoyed it more if I knew more of the people there.

  4. Now I want to see a survey of rationalist communities, to get data like how big they are, how many core members vs. peripheral members, etc. This seems useful to know.

  5. Levi Aul says:

    A fun anecdote: I wrote a piece of group-chat software once. Instead of showing user names or avatars, each user would “own” a given color (an arbitrary RRGGBB code they entered) and their messages would appear with this color as the background color.

    The naive thought is that there are enough colours for everyone. In reality, though, people will avoid colours that might get them mistaken for another already-well-known user. So the second-level thought is that it was only a matter of time before we ran out of colors and had to start giving users the option of mixing in subtle patterns or something.

    But, as it turns out, each group that used this thing ended up floating just below the “color cap”: there was always only one obvious color left-over, which new users would constantly end up with for a while, decide the community was too hard to penetrate into, and then abandon for the next person to pick up.

    I should make it clear: nobody told these people to keep picking the “guest color”, nor was there a color picker showing unused colors, or a histograph of color usage, or anything like that. There was just the observation of current chat room activity, which would naturally push them toward the color.

    I think the “workable palette” ended up being ~28 colors, including this guest color, so there’s another number for you.

    • Scott Alexander says:

      You think that the number of colors limited group size, or that you got lucky and optimum group size was just below number of colors?

      • Irrelevant says:

        Implausibly convenient possibility: we can distinguish roughly the same number of colors and personalities?

  6. Ialdabaoth says:

    I’ve observed the same thing. In my experience, ten times-or-divided-by-two people is the viable size range for ‘task force’ / ‘hunting pack’ / ‘familial creche’. Then you go up a factor of 10, and a hundred times-or-divided-by-two people is the viable size range for ‘tribe’ / ‘clan’ / ‘business’. And then going up another factor of 10 is the viable size range for ‘conglomerate’ / ‘fief’ / ‘small town’. I would not be surprised at all if this factor-of-tennish trend continued all the way up.

  7. I’m kinda surprised that I’m a rationalist. maybe there’s room for conservatives in this movement, or maybe there’s more conservatives and libertarians in this movement than previously thought (they’re in the closet).

    • I’ve been gradually becoming more conservative since shortly before my son was born. Still an anarchist though.

      • “Still an anarchist though.”

        Is there some inconsistency in being a conservative anarchist?

        Arguably the earliest legal system is decentralized and privately enforced—what I think of as feud law. The basic rule is that if you wrong me I threaten to harm you if not compensated. The primary requirement is some mechanism that makes the threat more believable if you really have wronged me than if you haven’t, some reason why right makes might.

        Stateless so anarchy, and it’s hard to get much more conservative than the very earliest institutions for rights enforcement. I suppose you could argue that “reactionary” is more accurate than “conservative.”

    • Vulture says:

      I’ve been trying to avoid having political opinions, but I have noticed my worldview gradually drifting conservative since I started getting more involved in the online rationalist community.

      • Shenpen says:

        It really depends how you define conservative. For example, LW is making me more accepting of transgender, transsexual stuff, previously I thought it is yet another liberal hedonist bullshit, narcissistic people who have it so good that they can have luxury problems like that, and I used to think all they need is to do some real work like dig ditches for a year and they will rethink what matters, but now it seems something more real and profound to me.

        With LW I am less and less likely to simply dismiss other people’s untraditional views and experiences and this is making me less conservative.

        I am worried. Will one day I will not even want a hierarchical society dominated and ruled by my class, non-poor white straight men? Sounds like class treason to me but I may end up there.

        • Zorgon says:

          I’ve been shunting rightwards for the last couple of years. But I maintain that I’m mostly flinching away from the horrible stink produced by the authoritarian SJ front and their suspiciously-upper-middle-class cheerleaders; and that’s an immensely left-wing position to take.

          Still, I’m finding I’m getting on better with right-wingers than I used to. The enemy of my enemy is, if not my friend, at least significantly less hypocritical than my enemy.

        • LW is mostly progressive in its values (as far as I see; I’m not a regular there by any means), but it crystallized ideas like Chesterton’s Fence and Schelling Points, which are essentially conservative cognitive tools. Someone with these in their analytical toolkit is going to trend conservative relative to the more naive progressives, even if their terminal values are to the left.

      • blacktrance says:

        I’ve become more anti-left (particularly anti-SJ), but I’m as anti-right as I ever was. In the past couple of years, I’ve continued my progression into increasingly radical culturally liberal libertarianism.

        • loki says:

          See, I’m inclined to suggest that ‘SJ’, or at least, a lot of working definitions of SJ or SJW used by people who do not define themselves as SJ, does not actually refer to a political position but more a particular style of wording and presenting things, within a category of political thought.

          That is to say that an intersectional feminist or someone who would simply define themselves as socially liberal and at least moderately economically left wing (both political identities that predate ‘SJ’ as a term by a long way) may agree with an SJW on every important point except a few details of terminology, but not be ‘Social Justice’ because the latter is actually understood to mean things like ‘using Tumblr’ and ‘being uncompromisingly hostile to certain groups of people who disagree with them’.

          If SJ is, as I believe, a style of writing and presenting rather than a distinct political ideology then anti-SJ isn’t a political stance either, it’s an opinion on how people should express their opinions.

          @Zorgon: this means that, in my model, being against the values of SJs because of the way they express them isn’t in any way rational.

          I’d be interested to hear your thoughts on this but tbh I feel it would be more productive if ‘SJ’ and ‘SJW’ were tabooed when people are trying to disscuss actual positions.

          (For the record, I agree with SJWs about almost everything, politically speaking, but I do not identify as an SJ because I choose to distance myself from a lot of the way in which they express the opinions we share. I’m also a lot less certain of things than the prototypical SJW and view, for instance, ‘the kyriarchy’ as a useful model with flaws that has definite potential to be replaced by a more useful model).

          I would take exception at how SJs are apparently all white and middle class but due to aforementioned distancing from the label I kind of have an online social circle of nontypical SJs and SJ-adjacent people. It’s entirely possible that the ethnic and class diversity there isn’t a representative sample.

          Oh and alternative model, which also applies to a lot of the time I hear the term used: ‘SJW’ actually means ‘someone in a particular category of political opinions (such as feminist, trans-positive, LGBT-positive, anti-racist) who expresses their opinions in a way I, the person labelling them, dislike’

      • Ghatanathoah says:

        I’ve become much less conservative since I started reading Less Wrong and SSC. This is mostly because I realized that a lot of my more conservative positions were based on the Non-Central Fallacy (i.e. “Taxation is Theft”).

        I have always hated SJW type leftism though. I remember encountering a lot of it during my first year of college (2005) and going through reading binge of anti-postmodernist literature (i.e. the Sokal Hoax). When I started reading about Internet SJWs recently I was like “Good God, it’s spread.”

    • Wrong Species says:

      Isn’t there a decent number of libertarians in the rationalist community? After all, they both have a sort of “nerdy, tech guy living in San Francisco” reputation.

      • aerdeap says:

        nerdy bay area libertarians aren’t usually all that much like red tribe/tea party type libertarians, although there’s obviously overlap, and libertarian is probably also a safe way of signalling more conservative views without losing social status, in that type of enviroment.

    • stillnotking says:

      I’m sometimes mistaken for a conservative in online discussions, which is distressing on several levels. Ideas I have always regarded as core liberal principles are slowly migrating into, if not actual conservatism, at least a suspiciously pinkish no-man’s-land. (Examples: “People should be free to express their opinions, no matter how odious,” “Ideas should be evaluated independently of the person proposing them,” and “Laws designed for selective enforcement are bad laws, however noble their goal.” None of these are popular theories in contemporary liberalism.)

      I don’t know if I’m drifting to the right, if the Overton window is drifting to the left, or if the whole thing is a lot more complicated than it looks. The older I get, the more I incline toward the latter theory — I don’t doubt that “conservative” and “liberal” are basic categories, but the actual ideological content of those categories is more variable than it seems. Reading primary sources, as opposed to the modernized versions of political history that show up in history books, has proven very educational.

      • Jaskologist says:

        You’re a classical liberal. American conservatives are also classical liberals. American liberals, however, are increasingly not classical liberals, and might be better called “leftists.” It is indeed confusing how the terminology has evolved.

        • stillnotking says:

          I probably am a classical liberal, but I don’t think American conservatives are, by and large. When I read conservatives like Ross Douthat or Thomas Sowell, I find more points of disagreement than agreement. Admittedly, they piss me off a lot less than Ezra Klein does, but that’s probably bias. Dogshit in your own yard smells worse than dogshit in your neighbor’s.

          But the conservative POV *is* feeling less “diametrically opposed” and more “orthogonal” these days — I just don’t care about the same things they care about; the political warfare isn’t happening in my trench.

          • Anthony says:

            Arnold Kling hypothesizes that liberals, conservatives, and libertarians hold mutually orthogonal primary political concerns:

            My hypothesis is that progressives, conservatives, and libertarians view politics along three different axes. For progressives, the main axis has oppressors at one end and the oppressed at the other. For conservatives, the main axis has civilization at one end and barbarism at the other. For libertarians, the main axis has coercion at one end and free choice at the other.

          • Jaskologist says:

            Kling’s hypothesis sounds highly plausible to me, certainly more than thrive/survive.

          • Here’s a link to Kling’s ebook, The Three Languages of Politics, wherein he lays out his theory. As I’ve mentioned in a couple previous threads, it was a major influence on my own political perspective and I highly recommend it.

            Note that I don’t consider this incompatible with Scott’s “A Thrive/Survive Theory of the Political Spectrum”, which I’m also a fan of. In my view, the two theories are not competing to answer the same question. The former aims to identify and explain a key ideological thread that runs through the politics of all human civilizations throughout history; the latter aims to identify and explain the divisions of modern American politics at the national level, and the particular cultural tribes associated with them.

      • CJB says:

        I started using a two axis metric “Liberal” and “Conservative” standing for social positions, “Statist” and “anarchist” being the other two points.

        So you can be a Liberal Statist- and be stalin. or a Conservative Statist- and be Hitler. or a Liberal Anarchist” and be…..Occupy protestors, maybe? or a Conservative Anarchist and scream angry invective about how the government wants to replace your guns with chemtrails, presumably.

        Yes, the examples are all purposely stupid.

        I started out liberal statist, went to liberal anarchist, went to conservative anarchist, and am drifting towards conservative statist. Honestly, a number of socially conservative views make sense to me although I’m primarily economically conservative.

        Abortion? My main problem with the liberal line on abortion is…well, it’s pretty fucking glib, and ultimately boils down to what seems like a “push the fat guy on the train tracks” moral dillemma- do I have any obligation to let my body be used to save anothers life? The fact that liberals just go “Psssh yeah, baby! Your body, rock on, take no prisoners!” is disconcerting. There may be no right answer, in the end to “X harm to me vs Fetus”. But I think that question deserves a lot more moral consideration and at least an acknowledgement of it’s moral consideration.

        • How would you feel about a legal requirement to give blood? To give a kidney?

          • Anthony says:

            How would you feel about a legal requirement to give two years of your life?

          • jaimeastorga2000 says:

            @Anthony: The draft?

          • Anthony says:

            @jaimeastorga2000 – child support works out to about that, in terms of time. 18+ years, up to 40% of one’s actual income (possibly more if the state “imputes” income) out of a 2080-hour (or more) working year.

          • CJB says:

            I’d be opposed to it. I’d feel pretty much the same way about it that I do about outlawing abortion- the fundamental moral precept of bodily autonomy trumps.

            The Draft is generally a matter of military necessity, historically. Part of the societal deal is being prepared to defend your civilization.

            As for child support- well, I’m against most abortion in general from moral precepts, in that I’d advise against it in anyone I knew (barring extraordinary circumstances, obviously). Don’t wanna pay, don’t play- again, extraordinary circumstances apply (Cf. underaged guy that was raped and then forced to pay child support.

    • Doesn't know where to start says:

      The rationalist community started in a primarily liberal area. But as long as it is not actually connected to being liberal, you should expect to see the amount of conservatives rise as the rationalist community expands to other areas.

      As an independent, I think that tying the community to a political party is big bad mind killing idea.

      • MichaelM says:

        This is wise, mostly because political parties are interest groups that attract ideologies, rather than themselves ideological groups. Rationalism aspires to be an ideological group, so tying it to a political party will eventually mean subordinating the ideology to the interest.

      • A significant number of comments on this topic appear to be “I sort of like political ideology X though I have been drifting more towards Y”. That’s a good sign – people are adapting their view as they investigate new evidence, and not getting to hung up on being part of a particular group. Perhaps we could all go one step further and agree to adopt the idea of investigative politics.

    • Chevalier Mal Fet says:

      I’d say that conservative comes pretty close to describing my political beliefs (I’d say libertarian, but then everyone mistakes you for a huge Rand devotee, or I’d say liberal, but that term has been co-opted by progressives). I’m not precisely in the closet but I don’t really post much on here either. I’m sort of lurking in the shadows near the closet door, if you will, prepared to duck inside at a moment’s notice.

    • Emile says:

      I probably qualify as “mildly conservative”, or at least, not allergic to conservative ideas, but that could be a combination of contrarianism (I’m in a mostly liberal environment, be it LW, family, or work environment) and being interested in all kinds of ideas regardless of their tag.

      I generally don’t care much for political labels and try to avoid attaching to any of them (this is made easier by the vast amount of idiots that can be found on any side).

  8. Highly Effective People says:

    As appealing as the Dunbar idea is there is probably something else at work here.

    One possibility that springs immediately to mind is the one Angry DM mentioned in his article ‘Dear WotC: Why Do You Suck at Selling Games?’ (unfortunately I can’t link to it because the guy is screwing with his website):

    A lot of activities, like playing tabletop rpgs or being a member of a phyg*, require a special tier of members to initiate new members and prevent the group from collapsing. No matter how many people might want to be a part of the hobby the rate limiting step is how many of those DM-tier members you have because without them there’s no game.

    Notice the descriptions of new Bay Area Rationalists and Shireloth Micronationalists(?) gormlessly walking around looking for guidance. It’s the opposite of the adage “too many chiefs and not enough indians,” where the chiefs can only hold down a few new indians each and the rest drift off to other tribes. And since chiefs have their own rates of attrition and replacement the size of the group remains constant.

    Maybe the better solution would be to figure out what kind of skills / aptitudes your chiefs have and try to either poach chiefs from other tribes or see if there are any nascent chiefs in the newest crop of indians.

    *Best euphemism ever btw.

    • Scott Alexander says:

      Doesn’t this just pass the recursive buck to the question “Why can’t we effectively recruit more chiefs?”

      • Highly Effective People says:

        Edit: Yes, it basically does do that.

        But it also offers new avenues of attack on the problem (who are the nodes in your network? what do they have in common? which traits are important for node-iness?) and seems fairly plausible.

      • Cheers says:

        In the case of the Bay Area rationalists, “lack of trying” seems like as good an answer as any. As you say, all of the high status people who would be natural leaders/meetup draws/etc. are prioritizing other stuff (not saying that’s a bad thing; just saying). And it feels presumptious, perhaps, for someone of lesser status to take a stronger organizing role? Or perhaps the Berkeley meetups have been such a weak Schelling point for such a long time that improving things is an uphill battle? Or perhaps LWers are too introverted to make large group formation happen easily?

        The cynical explanation is that people join IRL groups as a means of getting laid, and joining the LW group to get laid doesn’t work well if you’re a heterosexual male, which is the majority of the community.

        • I think there is a good deal of truth to your cynical version, although “getting laid” is too narrow. Part of the way that activities without a direct payoff, such as political movements, work is by providing participants with the side benefit of meeting people who share their values and interests. Objectives include sex, long term relationships, non-sexual friendships, … .

          One of things that impressed me about Students for Liberty, a relatively new libertarian organization, is that it seems to be at least a third female. Not typical of other libertarian organizations I have observed.

        • Possible solution: niche meetings, so less famous people don’t feel presumptuous organizing, and attendees are more likely to be interested in the topic itself and less disappointed by lack of famous attendees.

      • drethelin says:

        That seems to be a much more actionable problem, because “how do we attract and keep people” Doesn’t point very well to specific things you can compare to each other. You can say “This group has Matt Elder, and thus it does better, let’s look and see at what he does to make a group awesome, compared to group y, which doesn’t have him.”

        • Deiseach says:

          Could it have something to do with “I got interested in this because of Big Name or Person I’ve Heard Of”, and then people either don’t get to meet with Big Name, or their interaction is very much curtailed, and they drop out because they’re not getting what they think they’re going to get, e.g. “I came here because I love Mike McGrath’s books, and in the eight months I’ve been attending, I haven’t even seen him from afar, much less heard him speak on the topics I thought I was going to get to hear him speak on!”

          In other words, your Matt Elder problem may be that Group A does better because Matt Elder is there, and since all the new people want to meet and interact with Matt Elder, but they can’t all be in Group A, then they fall away from Groups B, C and D after a while?

          • drethelin says:

            No, I specifically used Matt as an example because he’s a great organizer and good at running fun meetups but isn’t actually a Lesswrong Name in the way Eliezer is.

      • Held in Escrow says:

        Being a Chief requires both a level of commitment and skill. Only 1 in X people will have it in them to become a Chief. A each Chief can recruit a certain number of people, and there is turnover both in the number of Indians and Chiefs. Therefore any Chief that does not recruit X people will not hit the replacement rate for Chiefs, which means you slowly lose recruiters, and therefore stagnate or even contract.

        There’s also the fact that the initial group is going to be made up of Chiefs; people dedicated enough to get the group off the ground and expand (for any group that isn’t will cease to exist). So you’ll have a decent sized group which all has their own personal connections, but as you branch out from there (friends of a friend) you have people will less attachment to the group and thus are less likely to have Chief attributes.

      • Anonymous says:

        Well, hey — maybe that’s the question you need to answer!

      • Ptoliporthos says:

        Perhaps you should look at how the army identifies potential sergeants? Or how they manage indoctrination of new recruits, especially when they need to expand rapidly (in a WWI or WWII-size draft)? I suspect this is a problem with known solutions.

        If you don’t want a D&D/micronations/rationalist boot-camp, perhaps the rapid expansion and cultural homogeneity of Starbucks provides a less threatening example?

      • Eli says:

        Doesn’t this just pass the recursive buck to the question “Why can’t we effectively recruit more chiefs?”

        A) Lack of trying.

        B) Different people who would otherwise lead the same cult grate on each-other’s nerves.

    • grort says:

      In my mental model of communities, the thing about the “chiefs” (high-status members? cool people?) is that they don’t want to “hold down a few indians each”. What they want is to hang out with N other people, and if more than N people are available to hang out with then they want to hang out with the N most awesome available people.

      N is determined here by the group’s most common form of interaction, which might be “participating in a forum thread” or “having a conversation at a house party”.

      In this scenario, adding more awesome people actually doesn’t change your group’s maximum size, it just increases the rate at which awesome people get to hang out with other awesome people.

      • grort says:

        (Clarification: clearly it’s wrong to suggest that none of the new people trying to join are awesome. There’s probably some composite factor based on intrinsic awesomeness and on how well you know the group members already?)

      • Highly Effective People says:

        This is a case where confusion can be cleared up by going back to the original D&D metaphor.

        Your focus on ‘cool people’ is similar to the old idea of the “sausage party.” Nerds want to meet girls, so if membership is flagging get more girls into the group and bam problem solved! Except attracting new players isn’t the rate limiting step. That would be the number of poor dopes willing to put in an extra ten hours outside the game drawing out dungeon maps and making sure everyone knows the gameshop is closed saturday but we can all still play at my place dammit.

        Chiefs may well be cool and high status but that’s not why they’re important. They’re important because they are the ones who roll their sleeves up and make sure that the hunt next week is going to actually happen.

      • HeelBearCub says:

        I have noticed that many online forums have a dynamic that goes something like this:

        * Well Know Poster 1 observes X
        ** Well know Poster 2 responds Y
        *** Poster 1 replies to Poster 2
        ** Unknown Poster 3 responds Z
        *** [crickets]

        Poster 1 and Poster 2 make a habit of noting each others comments and replying, which extends to many other “old timers”. They only have so much bandwidth to reply, so even if they find observation Z well thought out, they may not reply, instead choosing to further comment on something some other WK Poster has provided.

  9. emr says:

    Because people can belong to multiple disjoint (EDIT: Uh. Loosely connected?) groups, there is a difference between the limited connections of an individual and the limited connectedness of a group. If an individual can only form X relationships, then the more connections a person forms in one group, the fewer they can form in other groups.

    As a consequence, you should be careful about accidentally misallocating your mental relationship space. In particular, don’t form too many one-sided emotional relationships with e.g. famous people you’ll never meet. Maybe these aren’t a full “single relationship equivalent”, but I suspect they draw on the same limited resources used by two-way relationships.

    It can be interesting to model people as dots and relationships as lines connecting dots: You can even prove some elementary facts about e.g. the max size of a group where everyone knows everyone else, or the number of hierarchical “levels” required for a group to contain N people and still have everyone connected along some chain of edges. And consider the impact that individual variation can have on the possible shape of groups: Like being able to place person capable of having more relationships than normal at the top level of a hierarchy.

  10. Susebron says:

    There are communities that are fairly similar to your micronation example but significantly larger. With the main example I have in mind, which is a few orders of magnitude larger, one way it works is that it provides a space for people to do non-micronationish things, which lets it attract more people some of whom want to be involved in the micronationish stuff. I think that this is probably somewhat analogous to the online rationalist community – people are attracted to different aspects of it, and some of them will then want to get involved in the offline community, but it lets them choose about where they want to enter and what level of involvement they want to have.

  11. Nestor says:

    I first noticed this kind of static number running a yahoo group years ago. Signups were constant until the group settled at a smidgen under 1000 members and stuck there, even though people were constantly signing up, they were also constantly signing off so the dynamic equilibrium reached a stable point.

    I remember also being fascinated when I was more active running websites how a given link from a given site would day in, day out always send the same amount of visitors, more or less.

    It’s weird to think of all this aggregate of individual human agents as “traffic” but that’s how they behave, like a liquid flow. If you want the new dynamic equilibrium to stay where you’ve raised it, you need to keep that ad campaign going permanently. That’s why cocal cola still has ads despite being an instantly recognizeable planetary brand name, I guess.

  12. ishaan says:

    I wonder how exactly the formation of heirarchy makes larger group size (on the order of, say, a nation) easier?

    If you think of, for example, a typical school – 150 students, 12 teachers – it’s not like the 12 teachers are divvying up the students among themselves, and the teachers know each other better than they know the students. However, the 12 teachers are doing broad-coordination processes to manage the 150 students.

    Are there ~12 people doing broad coordination among bay-area rationalists? Does 150 represent the number of people a tight knit group of 12 can realistically coordinate, or is the 150 group made of un-coordinated fission-fusion groups?

  13. Sonata Green says:

    Makes me think of structuring groups into covens of 3-13 members, heirarchically if more people are to be accommodated. When the group starts to get big (~10-12 members), it’s time to mitosize.

    • tomlx says:

      Hm, reminds me of the Christian cell group concept. The community consists of many groups not larger than around 12 people (cell group), who meet weekly and are usually organized by one person in the group allocating tasks (meetup place, food, moderation, music, sermon, …) to the rest of the group in a round-robin fashion or by innate talent. Newcomers have it easier integrating in a small group than being lost in a big group. When a cell grows too big it is split to maintain the structure. Intergroup communication is maintained by bigger regular meetings and by regular meetings of smaller groups with common interest (sport, music, …), both of which also function as an attractor for new people, who are then invited to join the cell groups.

  14. suntzuanime says:

    Don’t groups just tend to naturally find their optimal size, where the gain from a marginal member joining the group is less than the cost of maintaining the extra network complexity? I find it much more believable that a certain group has a natural size given certain incentives/structures/”social technologies” than that this size is the same for all groups, or that you can slot group types into a few bins or whatever. Maybe 12 is just the right number of micronationalists.

    What’s the right number of psychiatry blog commentors?

    • J.P. says:

      This was my first thought on reading the post, that it was not so much Dunbar’s Number as Dunbar’s Curve of group size against social-technology level.

    • Shenpen says:

      “What’s the right number of psychiatry blog commentors?”

      We aren’t a group. It is just a thing I occasionally do. It is not the same people commenting under each post.

    • Harald K says:

      What’s the right number of psychiatry blog commentors?

      I don’t know, but I think we’re over it. That also shows a weakness in your argument: just because a group is over its optimal size, doesn’t mean people will want to be the one leaving it. Quite possibly they instead wish that other people would leave it, so there became more elbow room.

      On groups below optimal size it becomes even more obvious. I mean, the ideal size of our Hex club is surely 100-200 people, of which maybe 10-20 drop in to play each day. But in reality, it’s more like 6-8 people who hold a little tournament every two years or so.

      • suntzuanime says:

        Well, it still leads to a particular non-varying size based on the incentives etc. in play, it just might not be the “optimal” size in the sense of satisfying the values of group members. Which is a problem for sure, but if we’re just curious about why a particular group ends up always being 12 people, it’s not relevant to that point.

        EDIT: To be clear, you make a good point, and I was being wry because I felt my position had a weakness and hoped to draw it out. But now that you’ve drawn it out for me, I don’t feel it’s fatal.

    • Jaskologist says:

      I’m involved in a variety of local hobby groups, all of which have more than 12 people involved.

      However, that’s only counting membership/average meeting attendance. In most groups of this type, you’re going to have your core group of heavily involved people upon whom the club really relies, and then a bunch of less-committed people who may attend/pay dues, but not much beyond that. I doubt the core group exceeds 12 in any of these cases, but I don’t have real numbers. I don’t think we’re accounting for levels of involvement in a group.

      Some groups allow for varying levels of commitment, letting them grow much larger. Micronations are probably high-commitment, therefore low group size. My clubs meet once monthly, at most. Can a micronationalist really be a part of things with only a few hours a month?

    • Scott Alexander says:

      I agree this is a better way of putting it.

      I think Samo’s point is that this works in the form of a couple of chokepoints, where 12 is the right number of micronationalists, but if we really wanted a larger micronation and were very smart we could have discovered some social technological advance that transforms the right number to 100.

  15. Izaak Weiss says:

    See, the difference between the two communities you’re talking about is that it seems like the Bay Area community isn’t trying to get more people, but that other people are trying to come in but can’t; whereas Shireroth wanted more people, but other people weren’t really interested.

  16. Bagricula says:

    Evidence of one:

    I’ve only interacted once with the Bay Area rationalist community (at the Google SSC meet-up this week), and even from that brief hour, it is very evident that like all groups it has a strong set of unspoken customs, a hierarchy of ideas and modes of discussion, accepted and suppressed topics, etc. I imagine for someone less comfortable entering new societies, or adroit at navigating fluid and uncertain rules, this could be daunting.

    More than that, I felt much more keenly the sifting and ranking functions of the group, than the evangelizing functions. This kind of reminds me of the difference between my experience of Church groups and nerd groups. The Church groups I’ve interacted with are almost frighteningly interested in engaging me in a non-confrontational and re-assuring way. The nerd groups on the other hand have far more shibboleths and highly prize forms of interaction that lead to status changes (debate; argument) as opposed to interactions that muddle status differences (communal singing, communal prayer).

    Also, Berkeley is far from San Francisco and the BART doesn’t run 24 hours like a civilized system.

    • Bagricula says:

      Further, having constructed micronations of sorts of my own in years past: I’ve never seen such simulations get much beyond 20 or so people.

      My sense is that this is like novels…everyone playing wants to read the most interesting storyline, but after a certain number of characters people become relegated to side-plots or drop off altogether. My micronations were in great part experiments in collective writing and as such there really was rarely more than one “most important plot” happening at any one time even when the games spanned 20+ people with beautifully and individually fleshed out societies.

      • Susebron says:

        It depends. I’ve seen micronation-like things with hundreds of people, but they weren’t very in-depth in terms of story AFAIK. So there’s probably a tradeoff between depth and number of people.

        • Scott Alexander says:

          I’ve seen micronations with hundreds of people, but when I investigate them further they are very often micronations with twelve active people and a bunch of people who were coerced into joining, showed up once or twice, and were never taken off the citizen rolls.

          • Susebron says:

            I’ve seen sort-of-micronations with around a thousand people, and I was assuming that about the majority of those weren’t active. The largest one I’ve seen has over 3000 people, but at least 2/3 of them haven’t even logged on in the past month and for various reasons I think it’s reasonable to assume that probably 80% of the rest don’t do much, which comes out to around 200. It’s not quite a micronation, though, so the differences in general nature probably have a lot to do with the different size.

      • Nestor says:

        This is why most mmorpgs are structured like theme parks where everyone queues up to do the same quests and defeat the same bosses, everyone wants to be the hero

    • RCF says:

      If you’re looking for a SF LW group, there is one meeting on Mondays.

    • Deiseach says:

      like all groups it has a strong set of unspoken customs, a hierarchy of ideas and modes of discussion, accepted and suppressed topics, etc.

      Oh, yes. I’ve seen this in fandom, where bright-eyed young newcomer has a brilliant idea about “Hey, why don’t we/what about this thing?” and gets smote from on high with fire and brimstone from the old-timers; not because they’re stupid or offensive but because they simply don’t know that Topic X or Wonderful Suggestion Y has been done to death already and that your patience runs out the 500th time someone says “Why don’t we try this?” when you’ve already explained 499 times why yes, we tried that, it didn’t work, this is why and what happened or yes, we’re doing that already, look here and see.

    • Hemid says:

      Right. Dunbar-like limitations are probably inherent in everything they can inhere in, but every limitation isn’t Dunbar-like.

      I’ve never had anything to do with rationalism-as-community, but I’ve heard a lot of defeated would-be joiners walk away from it singing variations of your song. Defensive/dismissive responses invariably signal the rightness and health of the outsider’s decision to stay out—and they signal it with no self-awareness, no idea of their (in every sense) repugnance. So I believe them.

      The recruitment/retention limiter for rationalism-themed groups and maybe “nerd” groups more generally doesn’t look—from out here—like some interesting, universal, calculably emergent, hidden-from-all-but-mathematics-and-eternal-intuitions whatzis at all, but like something utterly boring: false advertising. The thing is not what it says on the box.

      Of possible interest: The thing in the box has no idea what it is, and it may be about not knowing what it is.

    • Scott Alexander says:

      I know a lot of people who think of debate as an unpleasant status move, and a lot of other people who think of debate as pretty much the only tolerable form of social interaction. I think among many of my friends debate and argument are kind of phatic forms of social grooming. This has led to some of my relationships being kind of complicated until it’s pointed out.

      Anyway, whether or not your points are correct, the specific complaint I heard was that fully-socialized members of the rationalist community from other cities and countries had trouble integrating into the Bay Area community. From my own experience the rationalist community has pretty similar norms and customs everywhere, so I don’t think it’s about that.

      • Nornagest says:

        I think debate norms have changed somewhat over the last while too, at least on the online side of the community. It used to be that I could expect to make relatively uncontroversial declarative statements without backing them up with high-quality peer-reviewed sources, unless someone else had a specific positive reason to disbelieve me — whereas these days if I post something like that on LW, there’s roughly even odds that I’ll be met with a demand for a source (or a better source) or something along the lines of “You actually believe that? #hashtagrofl”.

        I find this aversive. (confidence: high; sample size: me)

        • Matthew says:

          This could be a function of group size, too, though. When Lesswrong was smaller*, participants had a better sense of each other’s trustworthiness.

          *I’m not actually sure if the active community is larger, because a lot of older participants have moved elsewhere. But that replaces longtime participants with newcomers, so I think the point stands.

  17. James Picone says:

    If dropout rate is proportional to group size, and induction rate is proportional to group size but of a lower order than dropout rate’s proportionality, the group will grow until the rates are roughly equal and then stagnate.

    Different social technology/group properties change those proportions, so different groups have different intersection points.

    That is, the observation that different groups appear to have different stable sizes doesn’t necessarily imply several Dunbar-ish thresholds, it could just be different relationships between group size and dropout/induction rates.

    If you graph stable-size against number-of-groups-with-that-size, I would expect Dunbar thresholds to produce a stair-step structure, and differing rates to produce some curve that doesn’t fall off significantly at several different points.

    • RCF says:

      I’m not sure what you mean by “lower order”. If dropout rate = (c1)*size and induction = (c2)*size, then the net chance will be (c2-c1)*size. If this isn’t equal to zero for one size, then it won’t ever be equal to zero. The group will either grow exponentially or decrease exponentially.

      • Boris says:

        I assume James means something like dropout rate being c1*size^N and induction rate being c2*size^M, where N > M, such that for small size the induction rate is larger but for large size the dropout rate is larger, with equilibrium somewhere in between.

        • Peter says:

          And thus we see the difference between linear dynamics and nonlinear dynamics.

          • Jesse says:

            Pendant moment:
            These are still both linear…
            The first example is linear of order 1, the other higher order. They are both linear in that there is a set of constant C’s multiplied by x’s to some power. Terms like 2^(CX) or Log(CX) are non-linear, where a change in the value of the constant changes the term in a nonlinear way, and you can’t consider each term independently anymore.

          • Princess Stargirl says:

            There are people who use “linear” to describe C*n^2,C*n^3, etc!

            I am not sure I have heard this before. In R I have only heard linear used to mean ax or ax+b (this is affine). And in higher dimensions Ax+b where A is a matrices ad b is a vector.

            (and in infinite dim f(x+y) = f(x) + f(y) and f(xc) = cf(x) ).

          • Peter says:

            Linear in the constants, nonlinear in the variables. Seeing as nonlinear dynamics is all about the variables…

            At least that’s my understanding. I did some playing around with these things and reading a textbook or two last year. I thought that x^2 or xy terms were the things people pointed out when describing the nonlinearity of the Lorenz system or the logisitic map.

          • Jesse says:

            In control theory (like PID controllers) a system is linear as long as it is linear in the parameters. All the terms have superposition, meaning you can consider each term independently, and then multiply them by the scaling parameter to add them up, but the shape of each term is independent of the weighting. It makes the math nice that way. Real nonlinear is HARD.

          • Douglas Knight says:

            Princess, if by “R” you mean “statistics,” you should figure out what part of linear regression is linear and what part is not.

          • Peter says:

            I wasn’t talking about control theory, I was talking about nonlinear dynamics, as in, the field called “nonlinear dynamics”.

          • Princess Stargirl says:

            By R I mean the real numbers. Or in one dimension.

      • James Picone says:

        I’m very-loosely referring to a concept like .

  18. Chevalier Mal Fet says:

    Missouri’s not THAT far-flung. <_<

    Well, that was my initial reaction. Then I thought about how many people here in Kirksville (pop. 10,000, not counting college students) or back home in Kansas City I know of in the "rationalist community" and didn't come up with much. Er, are there any other Missourians here?

    • BD Sixsmith says:

      I suspect somebody could have all the fun they might have with people in the Less Wrong-o-sphere with people who have never heard of rationalism. Sure, the number of Americans – never mind humans – who have heard of Eliezer Yudkowsky is small but a far bigger number have heard of critical thinking, future technologies, political philosophy, science fiction fandom et cetera – and far more would be interested to hear about them.

      But, whatever, good luck tracking down like-minded souls.

  19. gattsuru says:

    The obvious critique, at least to my sense, is that Dunbar numbers are calculated averages, not strict rules, in the same sense that the Magic Number Seven for working memory has an oft unspoken +/- 2 after that. Facebook’s researchers, for example, report that the median user to have 99 friends, but the range is massive and a sizable group cluster around /20/ users.

    I have a very hard time interacting with more than about even half of Dunbar’s number on a regular basis: I’d not be surprised if my personal Dunbar’s number is a half or a tenth the expected number. This is freakishly low — I am not a social animal, this is smaller than most extended families including my own — but it’s not so far from the normal that I have difficulty operating in normal society. I don’t think it’s as rare as most people would expect. If you go into MMORPG guilds you’ll find that a lot of people only really interact with a small subset of the guild; if you look into companies or parts of the military or neighborhood divisions, you’ll see folk that talk with /everyone/, and then another group of folk that might only talk with close neighbors.

    This isn’t really a cascade of Dunbars, but it results in similar behaviors despite each individual person having a single ideal group size.

    I’d expect the bimodal distribution results from this part of human psychology. Very small groups can operate through purely direct results because even the loners must interact with every corresponding member of the group and individual success/blame can be readily distributed and attributed. As a group goes between this size and the typical Dunbar Number, it must develop social technology to close the gap of expectation between the loners and the social people. After it grows beyond the Dunbar number, it needs to develop /different/ social technology to handle expectations of supersocial people and to handle things like reputation that even the normally social can no longer handle automatically.

    EDIT:For bonus frustration, much of the ability to implement and even develop social technologies are dependent on the size of the group, as well. It is very hard for a group of one hundred to have an impartial mediator.

    On the other hand, sometimes it helps to have folk from outside your monkeysphere to work with. There’s no drama like small group drama, and it helps to not know your sausage-maker.

    • Peter says:

      The thing about Facebook etc. is that you don’t have a strict “group” structure, in that I expect I have lots more friends-of-friends than friends. Also I expect that I interact with some on Facebook a lot more than others. So I wouldn’t expect Dunbar’s number to transfer very well.

      • gattsuru says:

        To an extent, but Marlow’s definition of “friend” is somewhat more restrict than than the Facebook common definition, counting only cases where people friend each other, generally have a mutual friendships, and some other studies not on arxiv use a further restricted definition based on number of messages to each other yet get similar results.

    • Anthony says:

      If you go into MMORPG guilds you’ll find that a lot of people only really interact with a small subset of the guild; if you look into companies or parts of the military or neighborhood divisions, you’ll see folk that talk with /everyone/, and then another group of folk that might only talk with close neighbors.

      This isn’t really a cascade of Dunbars, but it results in similar behaviors despite each individual person having a single ideal group size.

      In modern society, most people have multiple different “social” groups that they interact with. In the EEA, the 150 people in your band were the only 150 people you ever interact with on a human level. These days, there’s 150 people in your company (or department), a bunch of old school friends, people who live in your neighborhood, the parents of the kids your kids go to school with, your church, etc., etc.

      And even in the EEA, you aren’t going to have the same level of interaction with all 150 members of your band. The Dunbar Number reflects the problems in coordinating the multiple overlapping smaller social groups that exist when there are more than 150 people in one social pool.

    • Scott Alexander says:

      I think Facebook is likely confounded by different behavior in using the network. Some people friend everybody they’ve ever met, others only friend the people they are closest to.

  20. Ilya Shpitser says:

    If you are right about the reason the Bay Area rationalist community has trouble growing, perhaps they should try to transition from being friends to being colleagues.

  21. Vladimir Slepnev says:

    Scott, it seems to me that the problem is not related to group size. To take an example that’s very familiar to me, most open source projects have only one contributor who is the original author. Most of the remaining ones have only a handful and are struggling to get more, and only a few projects become really popular.

    The tiny projects could blame Dunbar limitations for their lack of popularity, but they would be wrong. They could put in all the affordances for growth that they wanted (welcoming atmosphere! training for new members!) and would still stay unpopular.

    The actual reason is that they don’t appeal to people as much. It’s the want/like distinction from your article “Are Wireheads Happy?” all over again. You attract people by targeting the “want” mechanism, not the “like” mechanism. Adding more things for people to “like” won’t solve the problem.

    • vV_Vv says:

      The actual reason is that they don’t appeal to people as much.

      If the problem is a lack of appeal to outsiders, then the group size will invariably dwindle as original members leave and little or no new members join.
      This may be the case with many open source projects, but it is definitely not the case with the type of groups Scott is talking about, which have high turnover rates but still somehow hover around the same size for many years.
      This suggests that these groups do have an appeal to outsiders, but there is some sort of capacity effect that disincentives people from joining or pushes existing members out when the group grows above a critical size.

    • Scott Alexander says:

      In Shireroth, we consistently got people joining who said “This looks really exciting, I want to participate!” and then never did.

      In the Bay Area, from what I hear the problem is that many rationalists from other communities are so inspired by it that they move here, but then don’t know what to do once they’re here.

      Neither of those sound like communities that just aren’t appealing to people.

      Even if it were merely that only one in a thousand people who thinks that they will like micronations has the right personality to actually like micronations, I would still expect micronations to grow over time if the founding members aren’t leaving.

  22. Shenpen says:


    For some reason, if the names of places or suchlike in a fantasy world suck, I am very strongly turned off. In Mount and Blade Warband I cannot play the original game, only mods because I think city names like Praven and Uxkhal are disgusting. I think the “correct” way to do it is to use largely understandable words like in A Song Of Ice And Fire, maybe somewhat mangled, like Blackfyre, Winterfell, Riverrun. Not necessarily English, although if you are writing in that language it is useful, but I accidentally played Skyrim in French which I don’t speak at all, and yet it was mostly okay to me to call Winterhold as Fortdhiver. If you invest the effort to make a new language really detailed and cool sounding, like Tolkien did, you can pull off entirely fantasy names, but if not, then their made-upness just comes accross clearly. If it begins with Shire- probably you have a mainly Anglo culture in mind, so just call it Redshire. If you are more into Mediterrean culture, La Contea Rosso or assume a bit of historical mangling and just Conterosso. At any rate, either use real languages or make a new one, but don’t just make up names please, really please, use names that mean something.

    • Benkern says:

      The micronationalist community, Shireroth in particular in fact, has a fixation on retaining things – from terrible place names to unworkable political structures. It’s a response to the challenges posed by creating and maintaining a community identity based on simulation in an ephemeral online world. Shireroth has lasted where others haven’t, even though their conlangs, backstories, culture etc. are more original and brilliant, because they lack the social capital that (in our tiny corner of the internet) ‘Shireroth’ retains. It is an interesting feature of the high turnover of citizens in Shireroth, rather reminiscient of the ecological side of the Chicago School, that new members assume and maintain the values of previous members long after they leave.

    • Deiseach says:

      I think the “correct” way to do it is to use largely understandable words like in A Song Of Ice And Fire

      That’s odd, because I have the opposite reaction: Uxhal seems perfectly fine to me, but the mangled Ye Olde Englishe of “A Song of Ice and Fire” drives me spare. Mainly, I suppose, because I recognise the concepts he’s swiping building upon, and I keep going You’re doing it wrong/That’s not how that name is spelled/No, that role doesn’t function that way.

      Though Winterfell is a perfectly fine name. Riverrun, on the other hand, gives me Joycean flashbacks:

      riverrun, past Eve and Adam’s, from swerve of shore to bend of bay, brings us by a commodius vicus of recirculation back to Howth Castle and Environs.

      • Anthony says:

        Doing “Ye Olde English” is easier to get right because you’re hearkening back to the days before spelling was invented, but there are still right and wrong ways to do it. I don’t suffer Joycean flashbacks nearly as easily, especially from single words, because on my first initial exposure to Joyce, I recoiled strongly enough to avoid him so much that I don’t recognize the patterns.

        • TheAncientGeek says:

          Disagree. Evidence: lots of writers doing it badly.

        • Nornagest says:

          Martin isn’t really doing Olde English; he borrows a few patterns from older (mostly Elizabethan or later) dialects, but the actual structure is something closer to what Anthony Burgess does with his “Nadsat” slang. Lots of substitutions and period idioms to give flavor, but on a modern English backbone.

          I don’t blame him. Actual archaic dialects are really hard to get right, and tend to make things much harder to read even if you do. A lot of fantasy writers in the pre-Tolkien era wrote in a more-or-less good archaic idiom (usually something close to what you’d find in 17th- and 18th-century fairytales), and reading them tends to be a slog because of it: Tolkien’s “translation from the Elvish” convention was actually a stylistic innovation over the likes of for example The Worm Ouroboros or anything by Lord Dunsany. Although he was almost certainly trying to suggest something like the translations from Old Norse and Anglo-Saxon English that he would have been familiar with through his work, rather than to make things easier on the reader per se.

          • Douglas Knight says:

            Isn’t that exactly what Deiseach’s phrase “Ye Olde Englishe” evokes? Isn’t “Ye Olde” an example of what you’re talking about?

          • Nornagest says:

            “Olde English”, to me, evokes something closer to the Eddison or Dunsany examples. but more naive and less complete; you’re a writer, you have a vague idea that people in fantasy books should speak in an old-timey way, so you reach for “old-timey” and grasp a 21st-century imitation of a 19th-century revision of a 17th-century folktale that you read in your grandma’s house when you were ten years old.

            The results bear a passing resemblance to Martin’s approach, but only a passing one; it’s pretty obvious when someone’s mixing modern and archaic deliberately, for a particular effect. (HBO’s Rome takes a similar tack.)

    • vV_Vv says:

      A Song Of Ice And Fire is overtly set in an fantasy version of medieval Britain, hence a somewhat mangled Ye Olde Englishe may sound appropriate.

      In other works the authors want to maintain a generic fantasy setting that can’t be pinpointed to any specific country, hence they try to avoid names that sound plausible in any real language.
      I suppose that this easily results in abnormal names that don’t obey consistent morphophonological rules, but I doubt that many people people other than linguists can notice it. 🙂

      • But seriously, how hard is it to just create a naming language with a consistent phonology? The hard part, I think, is just knowing that this is a thing which you can do; once you have been exposed to the concept, the execution is trivial.

      • John Schilling says:

        If you’re translating the rest of the work into English, shouldn’t you do the same for the names?

        The names that actual people come up with aren’t arbitrary conglomerations of syllables, and they are often functional descriptions. Just grabbing the nearest map of England, I see a whole bunch of -mouths including the doubly-descriptive Portsmouth, plenty of -fords and -bridges, some -gates and -fields and -woods, and of course -tons and -hams, -boroughs and -shires. Similar patterns emerge in other countries, even if you aren’t fluent in the language.

        These not only feel like organic names, they are really useful for people who are trying to keep track of a fictional geography that they just learned an hour ago and really don’t need to burn into long-term memory through long study. So I’d say either Anglicize at least the functional parts of the names, or set up consistent rules for that part of your fantastic linguistics and geography and include obvious pointers in the text.

        • Irrelevant says:

          If you’re translating the rest of the work into English, shouldn’t you do the same for the names?

          In fiction or in real life? Because in real life, the reason we don’t do this is that people are really awful at naming things. I mean, look at how we named all our deserts. When you translate those names, most of them are just “[The Desert] Desert” or “[A Place Without Water] Desert.” On the slightly more creative side you have Great Desert, Eastern Desert, High Desert, and my personal favorite because it reflects the difference in degree of precision about grazing conditions between languages, Mongolia’s Partly-Desert Desert. None of these work as unique signifiers once you’re in a culture that cares about global geography.

    • Nornagest says:

      This is one of the things I love about Gene Wolfe as a fantasy writer. He fills his settings with strange words, but if you’re armed with a decent grasp of etymology and a sufficiently stump-like dictionary (hoary, cumbersome, suitable for use as an improvised stepladder), they’re always intelligible: either as obscure English words or close derivations therefrom, as novel assemblages from recognizable Latin or Germanic roots, or, in one case, as a printer’s error.

      One trick of his, for example, is to use the names of extinct animals to simultaneously suggest a setting and lend an air of strangeness. The Book of the New Sun features e.g. merychips and teratorns, which the context reveals as beasts of burden and raptors respectively, but one familiar with the paleontology of the Americas will recognize Merychippus and Teratornis.

      I think George R. R. Martin was going for a similar effect, but he isn’t as good at it.

    • Irrelevant says:

      My favorite solution to the fantasy naming problem is that seen in The Black Company, where in order to facilitate future reading and translation of his book, the polyglot narrator translates all the place names into (whatever language he’s currently writing in, represented to the reader as) simple English. Solves the problem of things having weird names handily.

      • Anonymous says:

        Does he translate people’s names, or just place names?

        Looking at a couple of wikipedia pages, it looks like a lot of people have meaningful names, but many so meaningful that they were probably not given at birth, but are nicknames. But I also see fabricated names of places and people: Bomanz, Hagop, Shukrat, Tamarask.

        • Irrelevant says:

          Company members primarily have nicknames or literalizations for the same reason as the place names. Non-Company people often have real names.

          Also the convention changes a bit in the Books of the South, to my slight disappointment though with sensible internal justification.

    • Scott Alexander says:

      Yes, the name was made by a high schooler 15 years ago. When I can get away with it, I try to refer to it by its autonym, Sxiro, but I could never get anyone else to agree to the change.

    • Agronomous says:

      Praven: brand of male contraceptive
      Uxkhal: former Albanian finance minister
      Riverrun: twee; I’m sure somewhere in suburbia I can find a strip mall dubbed “The Shoppes at Riverrun”
      Winterfell: tweer (tweeer?); female protagonist in mediocre fanfic
      Blackfyre: twees- tweeest- fuck it, MOST twee; almost certainly the name of Winterfell’s steed. (Though I’d be fine with “Blackfrye”, for some reason.)
      Skyrim: sexual euphemism (you don’t want to know; it may involve the above two)
      Winterhold: homeopathic cold remedy/immune booster
      Fortdhiver: moderately-successful ’90s metal-turned-prog-rock band; I forget where the umlaut’s supposed to go.
      Redshire: alternative name for Cambridge, Massachusetts, Cambridge, England, or Ithaca, New York
      Conterosso: brand of sparkling wine not so bad you can’t drink enough to get drunk on it, but not so pricey you feel bad about doing so
      Shireroth: the image of a four-foot-tall male humanoid with long blond hair, cavorting around stage in tight pantaloons swinging (and swigging) a half-full bottle of Buck’s Hall brandy, pestering the lute player, and executing jumping splits unavoidably springs to mind. Also, he sings.

      • Deiseach says:

        You got it backwards: Blackfyre would be the female fanfic protagonist and Winterfell would be her steed 🙂

        Fort d’Hiver I do like, but then I’m a sucker for French and Romance language names, so Conte Rosso also gets past without pinging my lámatyávë alarms.

    • loki says:

      I think with standards that strict you’d be highly disappointed with a lot of actual place names.

      ~ born within 10 miles of Rastrick, Scholes, Wyke, Tong, Hanging Heaton, Heckmondwike, Farnley Tyas and Slaithwaite (which is pronnounced ‘Slow-ett’).

      The made up names from M&B actually sound very like real place names from the areas of Europe those places approximate.

  23. Shenpen says:

    I think the idea of the Dunbar Number is that people outside the tribe we tend to dehumanize and not engage in an empathic basis with. A classic example is political fury. And that is usually called tribalism. So, we avoid tribalism by staying tribal: tribes of 150 people are less tribal than tribes of a million.

    • The two explanations I’ve heard for Dunbar’s Number is that esprit de corps evaporates at above about 150 people because, at that point, it’s too easy for slackers to find niches where they can hide.

      Alternatively, 150 is about the largest number of people whose skills most people can keep track of so as to know who to go to for help.

      Now that I’m thinking about them, I suspect there’s something about sub-groups hidden in both explanations. Maybe the slackers fall between subgroups so that getting them to work doesn’t seem like anyone’s responsibility.

      I don’t believe that most people can keep track of the skills (even for a particular project) of 150 people (is this just a reflection of my limits?), but they can keep track of who would know who to get in touch with.

  24. Steve Sailer says:

    Militaries have probably spent the most time wondering how big to make units. Here’s a table from Wikipedia:

    Typical Units Typical numbers Typical Commander
    fireteam 3 or 4 corporal
    squad/section 8 or more sergeant
    platoon 15-30 lieutenant
    company 80-150 captain/ major
    battalion 300–800 lieutenant colonel
    regiment/brigade 2,000–4,000 colonel/ brigadier general
    division 10,000–15,000 major general
    corps 20,000–40,000 lieutenant general
    field army over 80,000 general
    army group several field armies field marshal

    • Shenpen says:

      A fireteam is probably a technical innovation and we can ignore it. As for squads, it is interesting how stories about very strong bonds and friendships and camaraderie and all that are all about squadmates. I think terms like platoonmates or companymates are not even used. Your brothers in arms are the squadmates.

      Platoons are not special IMHO, similarly technical to fireteams.

      I think the company is the highest level that has some kind of personal linkage and identity.

      Above it, the whole thing is just kept together by symbols and traditions and ceremonies, but less personally.

      So the squad and the company are the most important from this angle, the first representing the really strong bonds, the second the level where personal bonds exist at all.

    • Tarrou says:

      Shen has a point. Higher organization may be dictated by technology, especially communications, but all militaries throughout history have organized small groups of about what we would call a squad (10 has been popular, nice round number) and a larger group of between 100 and 150 (a century, a zuun, a company, a lochos). Everyone from the Mongols and Egyptians to the Greeks and Prussians have followed this pattern.

      • Chevalier Mal Fet says:

        The reason both armies and navies have the rank “captain,” but at different levels of command, stems from this.

        Originally a “captain” was the rank that could command the largest fighting unit that could still be handled by one man (ie, without delegating to subordinates). In naval terms, that’s a single ship. On land, that comes out to…yep, about 150 people, usually.

        You still see echoes of the old usage as simply “leader of men” in places like the Bible or Tolkien (Faramir, “Captain of Gondor,” the Captains of the West, etc.).

        So yeah. Just wanted to reinforce the data point “150 is about the largest we can get away with” and share what I’ve always thought was an interesting bit of etymology.

  25. Murphy says:

    I just realised we were probably at UCC at the same time. Greetings.

    • Scott Alexander says:

      2008 to 2012?

      Are you the one rationalist from Cork whose existence I know about, initials R. D. ?

      • Murphy says:

        Right date range but I’m not R.D.

        I wasn’t into lesswrong or the rationalist movement back then, it’s just kinda weird sometimes when I realise that someone I assumed was in utterly disconnected social circles on the far side of the globe turns out to be a couple of nodes away since I had a few friends in med and philosoph.

  26. Autolykos says:

    The Pirate Parties have run into the same problem and solved it with the “Swarm” concept: Instead of a pyramid-like hierarchy, you have lots of small groups, some organized by region, some by topic, some by function. You can be member of multiple groups (in practice, no more than two or three), and groups will split when they reach eight members.
    That way, communication overhead is minimal, there’s always room for new members, and the groups are small enough to make total consensus possible (it’s even required). There’s also the “Three Pirate Rule”: If at least three members think something is a good idea, they can just go ahead and do it in the name of the party – effectively founding a new group for the task.
    Here’s a more in-depth description of the concept:

    (I really like this blog, btw. I’m just reading it front-to-back.)

    • Corwin says:



    • John Schilling says:

      “Groups will split when they reach eight members”

      Given that, for most members of most groups, the ability to socialize with their cool groupmates will be at least as important as the nominal cause or ideology the group is fighting for, this might not work out as well in practice as it does in theory. I’d expect groups reluctant to recruit the inherently divisive eighth member, six-two splits with the two becoming disaffected and useless, and various dodges to pretend that a dozen people are “really” two groups of six, or one of seven and five irrelevant outsiders.

      I didn’t see in my brief skim how “Swarmwise” proposes to avoid this, and I’m not willing to do an in-depth read of an entire book just to find the secret recipe.

  27. chaosmage says:

    Many groups that continue to grow over time have permanent spaces reserved for new members. Parties and churches have youth organisations, civil society has schools. Here new people can make between themselves the strong bonds that most seasoned members don’t have the time to establish with them. And unlike CFAR, these spaces don’t have great cost to enter.

    If neither Bay Area rationalists nor Shireroth has such a space, maybe permanent spaces reserved for new members aren’t just a sufficient condition for growing beyond a certain size, but a necessary one.

    • Deiseach says:

      Maybe rituals for newcomers. I don’t know what the equivalent of a rationalist catechumenate would be, but if everyone knows that “everyone else in the group is new and ignorant and nobody is going to laugh at us for not knowing the ropes yet”, there are mentors or people who you know you can go to with your questions, plus there’s a way of advancing that you understand and is fairly explicit (rather than the “no, we don’t have a hierarchy here, everyone pitches in and we treat everyone the same!” official line but you know damn well in practice that Some Animals Are More Equal Than Others), that may help with the “numbers recruited = numbers leaving by natural wastage and attrition or lack of interest” retention problem.

  28. brainiac256 says:

    When I was young, my parents’ church was in the same place of worrying about stagnation. Since there is a clear commandment to Christians to always be trying to grow the community of Christians, they tried a number of things, which I was made part of because I was not allowed to not be a Christian.

    All bitterness about my own freedom aside, some things did stick with me from the presentations and conferences. One of the major points seemed to be that growth is facilitated by rings of about 4-8 people. For example, a small church might have a youth program with about 8 kids in the younger division and 8 kids in the older division; once you start to grow, you should consider making more small divisions rather than allowing the two groups to increase in size willy-nilly, because keeping the number of “in” people small will allow newcomers to more easily fit in to the new community.

    Breaking up these groups into rings doesn’t mean you end the relationships between people who end up in different rings — you leverage that into a ring of its own. So the texture of a strong, healthy, growing community looks like chain mail, with lots of interconnected rings. This sounds pretty similar to Autolykos’s comment about the Pirate Party’s Swarm.

  29. onyomi says:

    So does anybody who believes in Dunbar’s Number still think one “representative” for every 600,000 people is a good form of democracy?

    • Peter says:

      Thinking about it… sizes of parliaments and other bodies, and parties within them. There should be an interesting comparison in the USA between the Senate (100 people) and House of Representatives (535 people). A majority in the House of Representatives will be above Dunbar’s number in size; the current Democratic minority is above the number in size, but a party that’s on the ropes should be below it.

      The UK Parliament is even bigger than the House of Representatives, and has two main parties, so the same considerations should apply.

    • Scott Alexander says:

      Maybe some of the historians here will tell us whether there’s ever been a Dunbarocracy – a country with units of 150 that each elect a representative to the next highest level of 150 people.

      Or if this is even desirable compared to Samo’s sizes of 150, 1000, 90000, etc.

      • Peter says:

        I remember hearing somewhere that Gadaffi-era Libya had a system that was ostensibly like that, but the number might have been higher than 150. That said, Gadaffi, so I think that Dunbar’s number is the least of the issues.

    • So would it follow then to introduce more levels, so that you know the person you vote for personally, then they mix in a Dunbar sized group, then vote for someone at the next level. So you get groups of say 120, 14400, 172800 and 2073600 (maybe should just go metric). The advantage seems to be that you can actually know whether the candidate is a scumbag or liar or not, but on the other hand multiple levels could lead to obfuscation and less accountability because the people making the decisions are no longer directly accountable to regular folks. Another problem is that it becomes harder to vote on principle and easier to vote on personality (which may be disconnected to final policy).

      I guess a mixed/weighted system might be interesting. So you vote once on personality at a local level, and once on policy at a national level, probably using proportional voting, and then the votes are combined somehow.

      • onyomi says:

        I just think nation-states should be much, much smaller. At a certain level, so long as you don’t have world government, you always reach a state of anarchy (among the sovereign states of the world). I think almost everyone would get more effective, responsive government if they reached that level much sooner, both by reducing the ratio of representation and by reducing the number of levels between the individual and sovereign power. Very small nations and city states are usually nicer places to live, partially because they have to be, because it’s too easy to move away.

        • I do intuitively like the idea of local control, but some problems, like pollution being sent across borders, or foreign tax havens, seem to not be amenable to simple mutually beneficial agreements, and require international coordination or at least regional cooperation. Wouldn’t it be important to have some democratic mechanism to try to make sure such cooperation is fairly implemented, and not just some folks pushing around others?

          • onyomi says:

            I like tax havens because they limit how much governments can tax people before people start hiding their money. A dream for some, world government, or intense cooperation among governments for such purposes as taxation is a nightmare to me, as it removes all barrier to abuse when there’s nowhere to escape.

            Realistically, competition among governments is the only way to keep them honest, just as competition among business is the only way to keep them honest. And the only real way to express discontent with a government is not by voting but by moving away. This is why the most evil governments prevent people leaving. A good government doesn’t have to erect exit barriers, nor worry about tax havens.

            If it is ever necessary to break up, say, Standard Oil, whose customers are voluntary, how much more should we break up the United States federal government, whose 300,000 million “customers” have few alternatives but moving to Canada or Mexico?

          • I think those are good points and don’t disagree with the mechanisms you describe. Unrestrained governments can turn bad. I do think your solution ignores the serious problems of coordination that I mentioned though. Naturally I am also concerned about those that take or consume and don’t give back, especially when they are financially capable. You’ve probably heard the arguments before – some level of public services are needed to ensure social mobility and meritocracy. If you don’t have that then government just becomes a tool for others anyway.

            Generally speaking I think the Western world has traditionally used the idea that you use markets where there is a natural diversity, and you use a strong democracy (voting, but also well-informed non-propagandized citizens) to retain control when there is a natural monopoly or need for centralization. Sadly the democracy side of that coin seems to slowly failing to apathy and corruption, but having no government seems like a non-option. There will definitely be other powers that will be filling that vacuum. I hope we can improve the functioning of our democracy somehow :-/

          • onyomi says:

            The democracies in smaller countries like Singapore also tend to be less corrupt, more functional, and more responsive.

            Though I understand what you mean about centralized vs. free market solutions, the problem is, which omniscient angels get to decide which problems are to fall under central control? If you put a group of politicians in charge of deciding whether x does or does not need to be centrally regulated, guess what solution they will come up with?

            As David Friedman has conceded many times, an anarcho-capitalist society will tend to be sub-optimal with regard to a few collective problems like global warming, though it’s also not at all clear governments do a good job with those.

            But I’m not even calling for an ancap society, though I think it’s a worthy goal to move towards; I’m just saying countries need to be smaller, as I think that’s the only real way to incentivize governments to do a better job.

            Though this is also not super likely to happen in the near future, I think it’s more plausible and ultimately more fair (because I’m not trying to force everyone to accept libertarianism, just to get them to let libertarians break away from governments they don’t support) than trying to get a libertarian elected everywhere.

            Therefore I will always strongly support almost any secession movement for just about any reason. I was very disappointed, for example, that Scotland did not vote for independence, even though they were breaking away to become more socialistic, so far as I can tell–the principle of more governmental competition producing better results in the long run is more important, imo, than demanding any one particular govmt be run more in accord with my personal preferences.

          • I think that the balance ought to be set by a democratic public debate informed by non political, moderate economists. But I also agree with your sentiment that honest, left or right, is always better than dishonest/corrupt, left or right. I can see an argument for smaller divisions in some cases for this reason.

            I might not agree with moderate libertarians but I find them personally very respectable in their convinctions.

          • Cauê says:

            “I think that the balance ought to be set by a democratic public debate informed by non political, moderate economists.”

            And thus the focus of tribal politics changes to the fight over who are the non-political, moderate economists.

  30. Jack says:

    Scott, in order to battle the steady decline in Shirerithians, you could write another blog post on how awesome it is… Seems like your last one at least had some effect in countering the negative immigration rate in the Imperial Republic. 😉

  31. Carl says:

    We miss you in micronations 🙁

  32. Jordan D. says:

    I think Bagricula, above, offers a very good explaination for the small size which inheres to a micronation- everyone who joins the micronation envisions themselves as the conworld’s next Great Man. If I joined a forum and found that every nation had a hundred members and it would take me years to join the ranks of The Secret Council Of Great Rulers, I might wonder why this was an improvement over joining a borough council. There’s a heroic narrative which I hope for when I’m joining a new community.

    …but while this explaination feels really good, it doesn’t quite match my experiences in, say, online guilds. When I first joined a guild in WoW, we had roughly 80 members and the reason we had roughly 80 members was so we could have the requisite 40 people online at a time in order to do raids. The guild master and his council of raidleaders didn’t personally know everyone there, and there WAS some increased milling in-and-out among the less-involved members, but the size was stable and we weren’t literally losing and gaining the faceless masses every day.

    Then, of course, WoW decided to shift the size of raids down to 25-man raids, and the size of the guild dropped to roughly 45 people. And we KNEW it would, although I don’t think anyone knew precisely why it would. It just seemed obvious. It wasn’t a jarring transition for me at the time; all the people I knew remained and the people who vanished were the ones I didn’t know.

    So then 10-man raids became a thing (though not a replacement for 25-mans in the same way), and our guild organized itself into roughly two 10-man groups and a single 25-man group and… the number of people fell to about 30!

    And come to think of it, I guess this DOES sort of match the heroic theory above. In a micronation, the ‘goal’ is to shape a collaborative metanarrative to the shape you want, so the useful number of people for that goal is ‘enough to appriciate my genius but not enough to render it meaningless’, a number which naturally increases as you move up the ranks. In WoW, our ‘goal’ was to conquer the Gigantic Structure and the Unfathomable Doom at the end so that we could have the most impressive swords and everyone else would wish they were us. When the number of people statutorially required to conquer the Gigantic Structure was reduced, entropy ate away the people with the weakest bonds to the organization.

    I think that’s what’s going on. Obviously, Dunbar’s number isn’t the optimal number of people for every project, and most groups are formed with a utility function that doesn’t match up with ‘everything I need from other people’. My guild wasn’t 150 people because it was organized to do one or two raids a week, which wouldn’t have been enough to satisfy that many people. Shireoth wasn’t 150 people because nobody who joined that conworld community wanted to place themselves in charge of writing about the municipal projects of Province 84, and 150 warring would-be kings would tear the nation apart.

    (I once got known as a fun GM on an IRC network and had 20 people sign up for a game. I tried to run it. Nobody was happy.)

    I don’t live in the Bay Area and so I don’t feel all that qualified to judge exactly what utility the rationalist community there fulfills. If I were to irresponsably speculate, though, I would guess that Scott is right-on when he offhandedly notes that EY and co. cannot meet ‘everyone’. The usual criticism of EY is that he is too comfortable in the role of a guru, and that too many people on LW want him to be one. I don’t know if that’s true, but it could be a sort of explaination for why people would go to California and then leave unhappy. If I joined the Catholic Church and read the Bible and fell in love with the words of Jesus and then discovered that Jesus actually lived in Rome and still taught there, I would obviously hop on a plane immediately. If I got to Rome and found that Jesus didn’t actually meet with every pilgrim who showed up (in this story, Jesus’ miraculous powers do not extend to warping time), it would be a huge bummer.

    Likewise, if I found LW and became such a huge rationalist that I was willing to move across the country just to live with my own people, it would probably hurt to find out that the great luminaries didn’t have time for me. On the other hand, that experience itself sounds like a lesson in rationality.

  33. vV_Vv says:

    Samo said his own research had found several of these discontinuities – I think he mentioned about 12 people, about 150 people, about 1,000 people, and about 90,000 people as the first few. He tried to tie this in to government forms like “family”, “clan”, “tribe”, and “city-state”.

    Is there any specific evidence of natural group sizes about Dunbar’s number? Sovereign states populations range between the 839 people of Vatican City to the 1.37 billion of China. Similarly, the size of religion organizations ranges from ~12 people cults to the 1.2 billion of the Catholic Church (funny how Catholicism is at the opposite ends of both scales).

    But I do notice that the Bay Area rationalist community is probably around 150 people and having trouble growing, and the size of my micronation was around 12 people and had trouble growing.

    I suppose that Shireroth is a community of peers, by the nature of the game there are strong incentives against recognizing anybody as a stable authority figure.
    The “rationalist” (I can’t help but always cringe at the term) community, on the other hand, has recognized elders: you mention Yudkowsky, Salamon and Vassar, and you could also add yourself and a few others, probably totaling up to about 12.

    Maybe ~12 is the magic number above which communities can’t function well without establishing at least an informal but stable hierarchy among its members.

  34. Sophronius says:

    Apologies if this has been said already, but the answer seems to me to be twofold:

    1) If you have a large group of people centred around a movement or idea, the group will increase until fallout from alienation > new membership attracted by idea/word of mouth. Alienation doesn’t really set in until a group reaches about 150 in size, so it has difficulty growing much bigger than that unless there is constant advertising.
    2) If you have a tight-knit group of friends where everyone knows each other, members remain part of the organisation just for the sake of spending time with friends and so it does not shrink any further than ~12.

    Anecdatum for this explanation: When I was part of a large multinational strategy gaming campaign, the members we gathered hovered around 150. When the campaign ended and there was nothing left to sustain the membership, the number dropped to around 12 and then very slowly finally went down to 0.

  35. Baby Beluga says:

    This is a cute theory, but I share your concern that the specific numbers (12, 150, 1000, 90000) are basically made up, and any other series of exponentially increasing numbers would have appeared equally plausible.

    In particular, the way one should test this is by asking people the size of their community *before* you tell them about this theory. Once you’ve told them, they’re hopelessly anchored–and if they like the theory, they’ll always be able to convince themselves that the size of their community is one of these four numbers, by appropriately setting the threshold for “true membership” to admit exactly the right number of people.

  36. Murphy says:

    On the other end of the scale I’m part of a non-profit hackspace with 1200 paying members, we’ve had steady growth for over 5 years, we long ago breezed past dunbars number without noticing and kept going (even when we didn’t want to) . We long ago passed the point where even the founders were saying it was far too large for one space at about 400 people.

    The trustees actively try to help other hackspaces to get started with equipment, support, advice and even funding to take some of the overflow/pressure but we keep growing by a few percent per month.

    Some notes:
    Our central unit of organization is the mailing list rather than chat, a forum or in-person meetings: whoever wants to can keep up passively with events without needing to actively take part.
    Our central repository of information is a wiki.
    Our bureaucracy is automated wherever possible to an astounding degree.
    Very flat organization, almost no hierarchy. (hierarchy takes effort and introduces single points of failure)
    We do have a lot of loose subgroups who maintain particular machines, training or parts of the space.
    We run on a “pay what you feel it’s worth” subscriptions model with a minimum of £5. (recommended £25)

    (Anyone who happens to be in london feel free to drop by, we have an open night ever tuesday)

    There is of course an in-group who are especially active but it’s far more than 12 people.

    So there may be something about people investing even a small bit of money personally, there may be something about *non* hierarchical organizations or it may be something about the forms of communication used , 1-1 communication vs mailing lists and wikis.

    • Scott Alexander says:

      I think there might be a distinction between a useful organization and a social organization.

      The hackspace is fulfilling the very specific need of people who want to hack stuff. That doesn’t seem any more Dunbar-limited than the observation that McDonalds has a hundred million (or whatever) customers.

      Micronations, Bay Area rationalists, and hunter-gatherer tribes seem a lot more like community for the sake of community. Even if micronations fulfills some important role-playing urge, it’s important that you’re fulfilling it with those specific people.

      • Murphy says:

        I see your point but I believe for a lot of members, including myself, it’s a social organization that happens to have tools and there’s a definite ingroup feel to it with group traditions, identity, stories/injokes which you don’t get with other customers at mcdonalds.

        It’s not just a business with a lot of customers, it’s got a strong community. It started a lot like some of those other groups you talked about with little more than a few friends meeting weekly in a pub for drinks and discussing projects.

        Do the rationalists communities have no purpose other than social?

  37. Squirrel of Doom says:

    If these numbers arise from brain capacity, you can infer how smart various groups are from them.

    If the kayaker club consistently has 14 members, and the poetry club 11, kayakers probably have bigger brain capacity than poets.

    Sounds like rationalists are just about average.

    • Humans’ brain capacities, in the sense that Dunbar was talking about, don’t much differ from one another. Humans in general are capable of doing a lot of things that chimpanzees can’t do, like maintaining interpersonal relationships with 150 people at once, and hypothetical minds with greater capacities than our own would be able to do things that we can’t do, like maintaining interpersonal relationships with millions of people at once. I’m not aware of any evidence that social group size is a workable proxy for IQ.

  38. anonymous says:

    I am much more interested in joining an irl rationalist community now that I know that Yudkowsky doesn’t hang out with everyone.

  39. thirqual says:

    A similar dynamic exists in (some martial arts) dojos. Above a certain number of members the assigned space is too small, the socialization more difficult and the in-group feeling that helps people come regularly and improve diminished.

    The solution that was taken were I practice is to push a portion of the always-present and very motivated black belts out, so that they will form and develop other spaces.

    • onyomi says:

      Having been a member of many martial arts schools in my life, I can say that they are a rich source of anthropological material re. group formation. They tend to be surprisingly drama-heavy, despite all the emphasis on self-improvement, etc., some places encourage the black belts to leave and start their own schools, some actively discourage that, many actively discourage someone simultaneously practicing at two different schools, some teachers intentionally teach at an extremely slow pace for fear of losing students, etc. etc.

      • thirqual says:

        Yeah, I had the luck to practice in places where financial incentives where low to non-existent, so was shielded from a lot of the drama. I heard horror stories about teaching pace management and progress sabotage.

  40. Emile says:

    Then they get here and everyone’s already doing their own thing, and it turns out Eliezer and Anna and Michael do not have enough time to personally interact with every single person in San Francisco on an intimate basis, and they kind of hang around the edges of the community not really knowing what to do or being connected to its general rhythm.

    I had a somewhat similar experience in that I was in the Bay Area and San Francisco for a week (for work), so I fired an email at the mailing list asking if there was a chance of meeting anybody or attending a meetup … and got zero response. Considering that I had helped organize meetup in Paris when we had people from the US visiting (Scott, and two other times when someone from the bay area was visiting, including one were she didn’t even show up and we ended up waiting all evening), I felt a bit miffed 😛

    (I did meet Alexei, but had contacted him separately)

    And in addition to that, I’M STILL NOT ELON MUSK!

  41. Princess Stargirl says:

    The people in the rationalist community core (Michael Bloom,Vassar, Anna nevermind Elizier) do not need more people. In fact too many people want to be friends with them. So they have minimal incentive (and therefore desire) to be warm and inviting to general members. Of course they were overjoyed to see Scott move to California. And I am sure David Friedman is being greeted with open arms too. But socially most of the leadership doesn’t really benefit from more people moving to California.

    If someone wants to grow the rationalist community they need to be someone who deeply wants to grow the rationalist community. Someone who would be overjoyed to see a talented but normal lesswronger move to california (say anyone with 145 IQ, decently knowlegeable about lesswrong and with ok career prospects). They cannot be someone who is only happy when an exceptional person (like scott) moves to Cali. And honestly not jsut the leadership has to want growth. So does the rank and file. I just do not think most rationalists actually have a deep desire to see the Bay Area rationalist community grow.

    *I see nothing odd about treating Scott more warmly than most people would be treated. This is just how people work.

    • Irrelevant says:

      I love that the bar for normal is IQ 145.

      • jaimeastorga2000 says:

        The actual LW average IQ is roughly 140 (it was 146 in 2009 and has declined all the way down to 138 as of 2014).

        • Irrelevant says:

          I don’t doubt it.

        • Edward Scizorhands says:

          I supposedly have a 150 IQ based on my SAT. This is silly.

          • jaimeastorga2000 says:

            If you doubt the IQ numbers, I refer you to the “2013 Survey Results”. Specifically:

            Can we finally resolve this IQ controversy that comes up every year?

            The story so far – our first survey in 2009 found an average IQ of 146. Everyone said this was stupid, no community could possibly have that high an average IQ, it was just people lying and/or reporting results from horrible Internet IQ tests.

            Although IQ fell somewhat the next few years – to 140 in 2011 and 139 in 2012 – people continued to complain. So in 2012 we started asking for SAT and ACT scores, which are known to correlate well with IQ and are much harder to get wrong. These scores confirmed the 139 IQ result on the 2012 test. But people still objected that something must be up.

            This year our IQ has fallen further to 138 (no Flynn Effect for us!) but for the first time we asked people to describe the IQ test they used to get the number. So I took a subset of the people with the most unimpeachable IQ tests – ones taken after the age of 15 (when IQ is more stable), and from a seemingly reputable source. I counted a source as reputable either if it name-dropped a specific scientifically validated IQ test (like WAIS or Raven’s Progressive Matrices), if it was performed by a reputable institution (a school, a hospital, or a psychologist), or if it was a Mensa exam proctored by a Mensa official.

            This subgroup of 101 people with very reputable IQ tests had an average IQ of 139 – exactly the same as the average among survey respondents as a whole.

            I don’t know for sure that Mensa is on the level, so I tried again deleting everyone who took a Mensa test – leaving just the people who could name-drop a well-known test or who knew it was administered by a psychologist in an official setting. This caused a precipitous drop all the way down to 138.

            The IQ numbers have time and time again answered every challenge raised against them and should be presumed accurate.

          • Edward Scizorhands says:

            Huh, for some reason I hadn’t internalized that an IQ of >150 is had by 1-in-2330 people. There’s 128K people in the US in that range. Okay, that doesn’t sound so unreasonable.

          • Wouldn’t a group composed of a minority of people with authentically high IQs, plus a bunch of others exaggerating their IQ, look basically like this? Because ~120-130 IQ people would probably be smart enough to know their IQ isn’t as high as it should be for this group, but are smart enough to take a flexible, unreliable test so they can semi-honestly signal group membership? Or I guess there could be just straight out deception. Perhaps I am just too cynical.

          • Ali’ve the average IQ.Don’t underestimate the effects of self selection: people can’t report a high iq by a reputable test unless they have taken a test, and they are more likely to take a test voluntarily if they suspect it will be flattering.

            (Involuntary tests amount to much same thing in context, since they are more likely to be given to someone who a professional suspects is at one extreme or the other).

            Witness, if you will, this faraggo, where the BBC decided that the UK is a giant Lake Wobegon where the average IQ is
            above the average IQ.


    • onyomi says:

      It is obvious to me that the IQ of the average LW/SSCer is much higher than average, though whether 30 or 40 points higher I have no idea. But I wonder if such people don’t tend to suffer from an “all chiefs and no indians” effect when they try to form largish groups?

      • Tarrou says:

        I think IQ doesn’t have much to do with leadership, and in fact may be negatively correlated. The problem with nerd-fests I’ve seen is more often too few chiefs, too many indians, and much too little social lubrication in terms of the skills it takes to form and exist in hierarchies.

        • onyomi says:

          Interesting. Good point. As someone who is strongly opinionated but not much of a “leader” or “organizer,” I can sympathize.

          • Tarrou says:

            As a corollary, good groups have a certain diversity of function or role. You have the responsible guy, the irresponsible one, the funny one, the scary one, the alpha etc. Everyone has a role. When groups are over-selected one can get a group in which few people feel they have their own niche, because everyone is good at the same things.

        • Joeleee says:

          I think if high IQ is negatively correlated with leadership skills, it means that people with strong leadership skills find it more difficult to become leaders. That is, they have the skills to lead, but not the skills to rise to a position of leadership (i.e. they lack the IQ that would make them high enough status to lead).

          Interestingly, I see this problem in engineering firms. The smartest engineers are often promoted, but their leadership/organisational skills aren’t actually well suited to higher roles, so their team/division stagnates. Other team members may have been better leaders and helped the group grow, but they didn’t have the necessary engineering skills to rise to a place where they could become a leader.

  42. Czernilabut says:

    On an unrelated tangent, I’ve always wondered if those who dismiss Malthus’s critiques have a solution to overcoming the limits of human socialization capacity. Despite twitter and all these other new technologies that facilitate social grooming, there still seems to be hard caps on the number of interactions facilitated ( I don’t think this can be something that can be handwaved as easily as “Well, technology will help us overcome this challenge”, especially given that large institutions, like the RCC and PROC exist in a pretty fragile state given how historically they’ve split off or existed as separate entities.

    • Irrelevant says:

      In the future, Google Now will tell us all of our optimal friendships?

      • I can’t wait! That would be a fantastic service if it actually worked.

        Whenever I move to a new city, it takes me several years to put together a community of close-knit friends. In the meantime there is a lot of socializing with people with whom, for whatever reason, I just don’t “click”, and we drift apart.

  43. Albatross says:

    I managed groups numbering between four and twelve people for nine years and there were between five and twelve other managers doing the same.

    There is a plateau at twelve. When my group, or any other group in the department, numbered twelve the manager was fully consumed just managing the team. Hiring, training, developing, mentoring, juggling everyone’s vacation schedules, moving work around when they were sick. A team with four people managers did lots of projects, managers with eight people did about half as many projects, and managers with twelve people had to delegate core duties to be able to attend to projects.

    It doesn’t take very long for the vacations, sick kids, funeral leaves, and broken cars of a dozen people to consume most of a leader’s time. And good luck finding a time when everyone can go to lunch, happy hour or have a meeting.

    I don’t know if I even buy 150 as the next plateau. I see more of cluster around 4 to 8, meaning 8^2 =64. I also note that communes and direct democracy tend to have a max of about 30,000 members… the number that could conceivably meet in a stadium.

    150 and 90,000 strike me more as the level, like twelve employees, where things start to get unmanageable.

  44. Albatross says:

    I’m an assistant instructor with a community ed Kung Fu program and I’ve noticed the head instructor simply stops teaching students who get to a certain level and don’t help teach the lower level students. There are a dozen locations and lots of room, even an advanced class just for those who teach. But he just doesn’t spend any time with black belts who attend class and help with lower belts.

  45. Maia says:

    This hypothesis lines up with my experience coming to the Bay Area, in that it seems like all the people you might want to befriend just don’t have the time or incentive to interact with new people and bring them into the group.

    The SF meetups seem to be starting from scratch, though, in a good way: we seem to be attracting people who are all having this problem, which there seem to be more than enough of. It’s a lot easier to befriend people who aren’t already 100% socially saturated.

    Of course, there’s also the problem that there aren’t really any good places where a newcomer can just show up and expect to be able to participate outside of the South Bay. The Berkeley meetups exist, but they’re not very well attended. So that’s another factor. I suppose that might interact with social saturation: it takes time to organize these venues, and why bother maintaining your friend-creation engine when you personally already have an extremely satisfying social life?

    Anyway, I have reasonably high hopes for what we’re building. It’s basically just another tribe.

  46. Tarrou says:

    This may be speaking to a small part of the OP, but it occurs to me based on two of my formative experiences that a sort of social feudalism is how you grow a group.

    In the church I grew up in, it was something of a cult, and there was an interesting way in which they absorbed new people. The church was divided into “small groups”, of ~7-15 people (there’s your 12 person group) and they’d meet once a week outside church times to socialize and discuss theology. New people would be hooked up with a group that met at a time and place congenial to them, and when a group reached critical size, it split and formed two new groups. Each group leader was expected to have a protege who would take over the new group when it came time. Everyone would come to the main services on Wednesday and Sunday, for gatherings of several thousand people, but the small groups were the core of the church.

    The military, on the other hand, already has all their small groups formed already (the squad is ~8-15 men, the team is 4-5), and each squad leader has an Alpha Team Leader who is second in command and being groomed to become a squad leader in his own right. New privates are absorbed into a team which has lost a member due to transfer, exit or promotion. If a squad gets more than twenty guys, they split off some troopers, promote a team leader and form a new squad.

    It would seem to me that extrapolating from this and things like multilevel marketing, political organizing, and large WoW guilds, the key is to constantly be generating new leaders and providing them with the opportunity to start their own little fiefdom within the whole. You can build the aggregate size by adding new groups, and around twelve people would seem to be optimal.

  47. Harvey says:

    Micronations definately has the unique problem of being a hobby based on rejecting authority. People start nations in order to create something new. These people aren’t necessarily looking for others that share a common ideology. It’s not a surprise that there’s a very low functional population cap even on the most successful nations – it has a natural “if you don’t like it, rebel and do your own thing” vibe built straight into its own existence.

    • Scott Alexander says:

      Yes, but it’s weird that rebelling doesn’t *help*. When our original Audentior group rebelled into AIN + AFTP, it didn’t create two nations that then grew to the size of original Audentior, it just created two nations that were half as big. When AFTP split into AFTP + GPR + FIOJ, they mostly drew from existing citizens and got even smaller.

      If we could split Shireroth into Shireroth 1 and Shireroth 2 and attract 6 new micronationalists to each, our problems would be over.

      Maybe the problem is that the community is too integrated. In fact, I wonder if this was a hidden danger of Bastion. Shireroth and Gralus were no longer independent entities, so we reached a stage at which the sum of Shirerithians and Gralans was twelve, rather than allowing twelve each.

      • Harvey says:

        I’m not sure. I worry that you’re mixing in other factors like most nations’ chronic inability to gain new citizens period. 12 may be a cap for Shireroth, but the par for most nations seems to be 5 or less active and dedicated citizens that do more than just show up once in a while. That suggests that there’s something about the model that can be improved upon: since Shireroth is more successful than many nations, why not a more successful Shireroth?

        But ultimately I think its more about my first point – that there’s no point in joining someone else’s dream. To most outsiders wanting to create a new society, Shireroth isn’t that much different than the USA – its got a long history they didn’t participate in and is run by people they don’t know. Shireroth skirts the issues a bit with subdivisions, but even THOSE have a decade and a half behind them now.

        After this long I’m pretty convinced its a hobby without a serious future. Like you said, we were so sure that we were going to grow indefinitely, but it’s just not realistic.

        • Benkern says:

          The ancient texts, carved in AIM tablets, state that people have been saying “this place has no future” since the year 2001. I’d leave this kind of nonsense to Bastion, but no wild Harveys spotted there for a while.

  48. Faradn says:

    I’ve never been involved with micronations, but there is this world-building GMless RPG called Microscope that has similar appeal.

  49. Tarrou says:

    An evil thought has occurred to me. I was thinking about relatively informal groupings and what seems to bind them together aside from the standard stuff (wierd religion, charismatic leader, outside pressure/competition).

    Many of the more successful at propogating themselves in a counterintuitive move raise barriers to membership. Basically, it costs something to join. This can be financial, but more often is some sort of hazing ritual. Frat pledges get all sorts of shit, military members are routinely “kicked in” to units, everyone from ashrams to the Skull and Bones do lemon sessions, etc, etc. Hell, even serious raiding guilds often put new applicants through a probationary period where they farm all the flask mats for the guild’s raids, and are randomly given difficult tasks to perform (out-DPS our top X player on Y fight!).

    I think there might be something to this as a builder of group identification. It could be as simple as cognitive dissonance, “if I put up with this, I must really want to be in this group!” I’d guess there’s more to it than that, shared hardship, connection with fellow group members etc. Hazing of some sort might be an important part of group formation.

    • Anthony says:

      This hypothesis, which isn’t original to you, can explain a lot of Occupy Whatever’s behavior.

      • Tarrou says:

        Actually, I’d argue Occupy as a counter-example. There were no costs to joining, people just showed up. As a result, they petered out and were unable to develop a coherent hierarchy, set of goals, or means of perpetuating their struggle.

        • onyomi says:

          Yeah, Occupy strikes me as having been too open to accepting disgruntled people of all stripes. As a result, no one could clearly articulate what it was about.

          Seems like a combination of clearly articulated goals and a sense of momentum are key to growth. The Free State Project has the former, but not the latter, which I think is what’s holding it back.

          If one really wanted a bigger community of rationalists in a particular area it seems like there would need to be a single, clearly articulated goal other than “wouldn’t it be cool to have all these smart people hanging out together.” Even WoW has the quests motivating the participants, even if they become ultimately a kind of excuse for the character building and social interaction.

    • onyomi says:

      I think you are right about this, but I don’t think it’s so much cognitive dissonance as the desire to feel special and the bonding power of shared adversity.

      It is generally hard to get into elite groups like “top 100 Chess players” in the world. Groups like fraternities seem to replicate the exclusivity and the difficulty without necessarily requiring any particular ability. It sounds like in the WoW case you describe test of ability is mixed with recreation of a social hierarchy that makes the new users feel like (junior) members of a family of sorts.

      People like Navy SEALs surely feel closer to their teammates as a result of knowing they’ve all gone through the same difficult training, but I think even when you’re being hazed by your senior members there’s a weird way in which even mock abuse is a bonding experience in that it indicates some kind of emotional investment.

      And in certain cases the exclusivity itself is a positive, maybe because it allows a kind of intimacy that is hard to achieve with > Dunbar numbers, and also because it implies a superiority: the “I liked that band before it was cool” phenomenon.

    • In case you didn’t realise, this is actually quite a well known effect. At many universities they teach it to first or second year psychology students. It’s related to the fact that people tendancy to assume if something was hard for them to get, it is very valuable. Of course, like all human stuff, it’s not quite that simple, but it’s definitely a factor in many groups.

  50. Pingback: Outside in - Involvements with reality » Blog Archive » Chaos Patch (#52)


    Also, is there anything to be learned from how communities (some of them pretty large and stable) form in Second Life?

  52. Baisius says:

    “Then they get here and everyone’s already doing their own thing, and it turns out Eliezer and Anna and Michael do not have enough time to personally interact with every single person in San Francisco on an intimate basis, and they kind of hang around the edges of the community not really knowing what to do or being connected to its general rhythm.”

    This is most of the reason I haven’t moved to San Francisco.

  53. mico says:

    I think this absolutely fascinating because I was a member of another, slightly different, conworld site and we actually encountered Micras. Some guys from there joined our site and seemed (to us) to be there mainly to try to poach members to their own site or ideally get all of us to just throw away our thing and do their thing instead.

    We made clear that we weren’t having it, being forced to remove ad links from their signatures and so forth, but not having actually banned them, they too got bored and drifted away after a week or two.

    Our site had exactly the same dynamic though – 10-20 active contributors the precise identity of which varied over time without affecting the numbers – and we made similar efforts to try to improve the situation with no success.

    If you haven’t already, though, take a look at NationStates. This is a very very large conworld site, which seems to have divided itself up into chunks of about that size. Not all of the Regions are so small, some are in hundreds or thousands of members, but I’d bet none have more than say 50 real active members.