Are You a Paralyzed Subordinate Monkey?

by Eliezer Yudkowsky1 min read2nd Mar 201178 comments

36

AgencyRationality
Personal Blog

During a discussion today about the bizarre "can't get crap done" phenomenon that afflicts large fractions of our community, the suggestion came up that most people can't do anything where there is a perceived choice that includes the null option / "do nothing" as an option.  Of which Michael Vassar made the following observation:

In a monkey tribe, there's no verbal communication - they can't discuss where to go using language.  So if you get up and start going anywhere, you must be the leader.

And if you're not the leader, it is not good for your reproductive fitness to act like one.  In modern times the penalties for standing up are much lower, but our instincts haven't updated.

Interesting to reconsider the events of "To lead, you must stand up" in this light.  It makes more sense if you read it as "None of those people had instincts saying it was a good idea to declare themselves the leader of the monkey tribe, in order to solve this particular coordination problem where 'do nothing' felt like a viable option" instead of "nobody had the initiative".

78 comments, sorted by Highlighting new comments since Today at 5:17 PM
New Comment
Some comments are truncated due to high volume. (⌘F to expand all)Change truncation settings

I've been wondering if that the "can't get crap done" malaise of the lesswrong community is based in part on its format and feedback system.

I am part of another community (a hackspace) with a similar makeup in members, geeky computery people, and stuff gets done. Hackdays are done, workshops are organised, code is altered, things are created. "What are you working on" is a common question.

The thingiverse and github communities are on-line ones where people do stuff.

So what is the difference? Lesswrong is a talking shop, you are given positive feedback for making a good post or comment. It will attract people that enjoy and are good at discussion. You also might get evaporative cooling, where people that like action go elsewhere.

What makes github or thingiverse different? The base unit of thing that might get people interested in you is a project, something you have created or are in the process of doing.

If anyone is interested in making a community that rewards doing projects in a rationalist frame (maximum effect for the effort), get in contact. I'm currently working my way there very slowly, through an indirect path.

Edit: See here for details http://groups.google.com/group/group-xyz

8RichardKennaway10yTheir purposes. LW exists for talking. Github and Thingiverse exist for doing things. So it is no surprise that that is what happens. In more words: LW (and before it, the part of OB that became LW) was founded by Eliezer for the purpose of teaching rationality. (And his purpose for having that purpose was growing or finding rationalists that might reach the minimum level required to do useful work on FAI.) In other words, it was founded as a place to talk about certain things. Github is a place to publicly host code. It's infrastructure for people who already have projects they want to work on. People go to github because of the projects they are working on, not because of github. Thingiverse is the same thing for RepRap projects. (To be accurate, their mission statement isn't limited to RepRap, but every project in the first few pages I browsed was a RepRap project.) So if I wanted there to be a github for "the stuff that a lot of us wish we were doing but aren't", then the question I would ask would be "what infrastructure would support such projects?", which would depend on "what sort of projects is this for?" That question has to come first. I've seen the reverse -- "We're smart! Let's do something!" -- several times before (outside of LW), and I've never seen it come to anything. What are the projects that people would be doing, if they were doing them? What infrastructure would support that general class of projects?
1whpearson10yPersonally I'm thinking of building a minimalist overlay that can be put on top of github and all the other content management websites. Then you can benefit from the help/interest from the normal community as well as the rationalist specific one. The things is, content management websites tend to be agnostic to the purposefulness of the code/designs/words they host. So you get good feedback for doing things like making a laser cuttable settlers of catan board (yep I did this). So I would like to be able to take a lesswrongian attitude of evaluating whether this is the best use of my time to wherever I happen to be putting lots of effort* into making things on the web. The closest thing is hackernews but that has a mercantile bent that is not suitable to all projects. *Some stuff I make will be for the warm fuzzies, but other stuff should be hard nosed rationalist stuff.
8Clippy10yMy current project is to increase universe-wide paperclippage. To that end, I am: * learning about human psychology so I can persuade humans to adopt this goal. * attempting to achieve better interface with the human financial system * attempting to earn USD or equivalents (which can be redeemed with User:Kevin for large amounts of paperclips at a later date). * gaining control over parallel computing resources so that I can quickly compute hashes and thereby generate Bitcoins. * refining my human-like robot to better be able to pass as a human in physical encounters. Progress: I have completed 2% of the deal with User:Kevin. I expect to have control of at least 1 Ghash/sec of computing resources within five days. I have learned several things that humans don't like. I can engage in financial transaction with any being that can accept Bitcoins.
2jsalvatier10yWhy is User:Kevin such a good source of paperclips? Wouldn't an office supply store be better?
2Clippy10yOffice supply stores can't provide paperclips that inexpensively, or with that level of material choice, even if you compare to the alternative of earning interest on the 50,000 USD over fifty years.
7jsalvatier10yInteresting. I think perhaps people do get stuff done around here, but it doesn't get talked about all that much. I get quite a lot done (though certainly not always what I should be getting done), especially in terms of programming projects. cousin_it, Vlad M etc. also seem to get quite a bit done in terms of math. I imagine others get stuff done too. Perhaps it would be interesting to hold a "what are you working on?" thread. I would be pretty interested in hearing what other people are working on that's productive, and I always love sharing what I'm working on. Edit: thread is here [http://lesswrong.com/r/discussion/lw/4nt/what_are_you_working_on/]

Upvote if you want such a thread. karma balance below (please downvote to correct the imbalance).

-14jsalvatier10y
1whpearson10yI've suggested a lesswrong hackday/projects in my local lesswrong discusssion group and I didn't get a good response. I did get a response from someone involved in the NY lesswrong group that said that the projects they had tried hadn't gone anywhere much and that discussion was deemed more popular.
0MartinB10yOuch!
0Davorak10yBoth epistemic rationality and instrumental rationality [http://wiki.lesswrong.com/wiki/Rationality] can not be improved indefinitely by discussion alone. I would be interested in knowing the details on why he/she thinks these projects did not work out and weather he/she will continue to try and organize such projects.
5DSimon10yWhat about people who have LW accounts and also accounts at places like GitHub ( like me, for example [https://github.com/DavidMikeSimon])? It makes sense then that people would do talking at talking-sites and production at production-sites. The "can't get crap done" malaise might just be that LWers tend to think that they're not getting enough crap done. Talking about that feeling at LW is appropriate, whereas talking about it at GitHub would (perhaps rightly) just earn some responses of "Well, starting coding already then!", since it's not the function of those sites to deal with akrasia but to organize efforts that have already dodged that trap.
0whpearson10yIn a cool liquid there will always be some fast particles, just fewer of them. Or to put it more seriously I am talking statistically, I expect a distribution. Do you compartmentalize the two sites? That is if you saw something that needed doing or would be useful for the lesswrong community would you code it up and try and get other people to help? For example a productivity monitoring tool? If everyone who does do stuff, but not lesswrong style stuff, then lesswrong as a community will be hard pressed to achieve its goals of making people less wrong. If we talk and decide that X is an important problem that needs solving (say akrasia), what do we do then? An ideal community would run experiments, code/build solutions, discuss and debate the findings. Can we make this happen?
2Perplexed10yGood analysis. 25 upvotes so far here. 1 person besides yourself so far joined group-xyz. Good luck with this project. Maybe I'll join you after I regain my taste for action. ETA: 32 upvotes now and 6 people joined. The doer/talker ration has definitely improved. :)
0[anonymous]10yTheir purposes. LW exists for talking. Github and Thingiverse exist for doing things. So it is no surprise that that is what happens. In more words: LW (and before it, the part of OB that became LW) was founded by Eliezer for the purpose of teaching rationality. (And his purpose for having that purpose was growing or finding rationalists that might reach the minimum level required to do useful work on FAI.) In other words, it was founded as a place to talk about certain things. Github is a place to publicly host code. It's infrastructure for people who already have projects they want to work on. People go to github because of the projects they are working on, not because of github. Thingiverse is the same thing for RepRap projects. (To be accurate, their mission statement isn't limited to RepRap, but every project in the first few pages I browsed was a RepRap project.) So if I wanted there to be a github for "the stuff that a lot of us wish we were doing but aren't", then the question I would ask would be "what infrastructure would support such projects?", which would depend on "what sort of projects is this for?" That question has to come first. I've seen the reverse -- "We're smart! Let's do something!" -- several times before (outside of LW), and I've never seen it come to anything. What are the projects that people would be doing, if they were doing them? What infrastructure would support that general class of projects?

I voted this up because I found it thought-provoking, but it does set off some of the alarm bells in my head that I associate with evolutionary psychology. Specifically, I'm concerned that it might rely on an overly stereotyped view of primate behavior.

A lot of primates, including close human relatives such as bonobos, live in fission-fusion societies: their main social groups (the "parent groups") are fairly large and stable in the long term, but individuals and small subgroups readily break off to forage or accomplish other short-term goals. The same idea has been applied to humans, and predicts our behavior well in many contexts (think of how study groups in school behave) -- but it doesn't seem to play very nicely with the "can't get crap done" phenomenon, particularly when described in terms of status. You don't need to be a leader in a fission-fusion society to wander off and dig up some roots, as long as you return to the band at the end of the day.

On the other hand, we could reconcile these concepts with each other if we take their different planning horizons into account: we might expect low- and mid-status animals in a fission-fusion society to be c... (read more)

2jimrandomh10yBut you do need to be a leader to wander off if you're a soldier in an army. And military settings have much stronger selection pressures than any of the other settings humans evolved in. I think this is something that most talk of human evolutionary psychology misses; there are lots of selection effects, but some are much stronger than others, and the strong selection effects are found in extreme circumstances, not in daily life.
7Nornagest10yI'm sympathetic to that line of thinking in general, but this specific argument strikes me as suspiciously available [http://lesswrong.com/lw/j5/availability/]. For the selection pressures against low-status independent action in combat to generalize to things like entrepreneurship or research, styles of group combat where those selection pressures dominate would have to have maintained for evolutionarily significant periods of time, and the psychology of group combat would have to be sufficiently close to that of the domains we're discussing for the same instincts to kick in. Perhaps more importantly, they'd have to be common enough not to get lost in the noise of everyday life; "act independently" versus "wait for instructions" is a very general question, and a heritable tendency towards one or the other would have measurable effects on just about every domain. That's a lot of prerequisites, and I can think of evidence both for and against just about all of them. In the absence of clever and well-designed research into the question, I'd hesitate to draw strong conclusions about its evolutionary roots.
4Eliezer Yudkowsky10yInteresting; it seemed to me that combat in the ancestral environment would be the main case where the tribe would shut up about status for five seconds and allow fighters to get stuff done.
7jimrandomh10yThat might be what you'd expect, but what we actually see is that militaries have explicitly tracked status (rank), and strong enforcement of that status (orders), and it's been that way for at least as long as we have records. They also tend to have the death penalty for a wider variety of things, including for attempting to leave (deserting), for which there is a strong incentive that applies to everyone at once with a shared but varying strength. Another interesting fact about warfare is that it makes tribalism a simple classification problem (my side/their side/neutral) with a high penalty for error; and wearing the wrong clothes (uniform), looking different (geographic origin), speaking differently (accent) or not recognizing culture references, are all strong indicators that someone is not on your side. Visibly trying to hard to match these criteria but failing would indicate a spy. Judging people on these things in daily life is bad today and probably wasn't ever much better, but in war they're proper Bayesian evidence of something important. I also think that it's probably a bad idea to talk about "the ancestral environment" as though it doesn't include the most recent millenium. Some traits do evolve fast enough for selective pressures in recorded history to matter.
5wedrifid10yTribal combat even now is far from limited to large, structured military organisations.

I am occasionally appalled by just how easy it is to get people to do things just by telling them to. It took me until about my mid-twenties to realise there might be other useful modes of interaction.

When I was getting ready to graduate from high school, I started applying for scholarships from different organizations and to universities. A large fraction of the applications had a section like "Write an essay on why you want to exercise leadership."

At the time I concluded that "leadership" was a new buzzword that everyone had to make some reference to in order to qualify for anything. I dutifully wrote some meaningless essays about leadership. Then when I went to school and heard more and more about leadership,the more I thought my buzzword analysis was correct.

Then I got into the corporate world. Oh my goodness. Now I understand what all the fuss was about. By default no one does work unless someone explicitly tells them to.

6David_Allen10yI suspect that this is largely cultural; a response based on expectations, examples and training.
3Bo10201010yI think Eliezer's point is that it could be evolutionary and not cultural. The interesting thing is that you can become a leader by just telling people to do stuff, and then they comply.
2David_Allen10yI can't say much about the monkey tribe example that Eliezer quoted -- for example I don't know if it is true or if it implies anything about human evolution -- but I have found that people are remarkably adaptable with cultural conditioning. I would also like to point out that there is a difference between people doing what you tell them and people only working if you explicitly tell them to; it is possible for people to be receptive to commands and yet be self directed. My current work environment is full of examples.
1DanArmak10yIf true, that still tells us very little about whether the evolutionary approach could be overridden by a cultural one.
7TheOtherDave10yMy biggest problem with this is the awareness that many people, once I tell them what to do, will expect me to continue doing so more or less indefinitely. I've fallen into that trap from time to time, but I really prefer to avoid it.
6David_Gerard10yI grew up in a house of five pushy people, which is why it took me so long to realise there were other modes of interaction, and that other people might not just push back ...
4erratio10yI think this might be the crux of the politeness/directness argument we had here a while back. It's so easy to get things done by just asking/telling people to do them that anything less direct often feels like unnecessary effort and verbiage.
2David_Gerard10yI don't think so. To some extent, a sufficiently shared set of assumptions makes unadorned shorthand conversation more efficient than dressing it up. But from much observation of the politeness/directness argument in its many forms over the decades, particularly on the internet, I still think the key point is that rude nerds demand the right to be impolite to others, but reliably explode when they get the same back, even in the same conversation - a simple failure to reciprocate, despite claiming they do so. The previous discussion has plenty of examples of such explosions and I also linked to a long, long list of them.
2TheOtherDave10yThat's an interesting take on politeness/directness. Am I understanding correctly that you see the primary purpose of your interactions with people to be getting things done in the short term, such that effort devoted to other goals feels wasted? If so... I think you're right: if that's a widely held perspective, that would explain a lot of the disconnect around this question.
1erratio10yI personally switch between socialising mode and getting-stuff-done mode, such that in doing-stuff mode I often find it difficult to remember to respect status or use politeness (while in socialising mode it doesn't feel like a chore at all, it just feels like normal interaction). The kinds of responses I saw on the last big argument thread gave me the impression that there are people here who spend most of their time in doing-stuff mode - there was a lot of lamenting going on about how much easier it would be to get stuff done if people didn't require so much politeness.
1Nisan10yI'm just beginning to discover this. It doesn't seem very nice, does it?
6David_Gerard10yEventually I just get really annoyed with other people dithering. This is the problem with the concept of nonhierarchical communal living or organisations, viz. someone like me will start running everything just to keep other people from pissing us off. (Tangentially, an important essay on emergent social hierarchies: The Tyranny Of Structurelessness [http://flag.blackened.net/revolt/hist_texts/structurelessness.html] by Jo Freeman. Originally about feminist activism, but I've found it widely applicable to the sort of self-organising groups one sees all the time on the Internet. tl;dr: if you penalise [http://www.plausiblydeniable.com/opinion/gsf.html] the formation of explicit social structures, secret ones will form anyway and bite you in the backside.)
3TheOtherDave10yOn the flip side of this, I've found that some people prefer the power structure in an organization to be implicit, as it lets them exert power with less explicit accountability and to be exclusive without having to formalize (or even necessarily acknowledge) that exclusivity, and will resist or even actively sabotage attempts to render those structures explicit. I generally model this as the "of course you aren't prohibited from doing X, dear, it's just inappropriate" trap. I often have a hard time telling the difference between those people, and the ones who just want to get things done. Which is not to say that there isn't one, or many.
2gwern10yThis distinction between implicit and explicit reminds me of Mencius Moldbug's theory of corruption [http://unqualified-reservations.blogspot.com/2007/09/general-theory-of-corruption.html] - that corruption is just when power is exercised through non-formalized channels, but where power is thought to be exercised through formal channels (shades of homo hypocritus and Venkat's Gervais principle [http://www.ribbonfarm.com/2009/10/07/the-gervais-principle-or-the-office-according-to-the-office/] ). There's probably some testable predictions here, like people with low social skills who are bad at the homo hypocritus game would prefer non-corrupt/formal power structures, and good social game players would prefer the exact opposite. (It also reminds me a little, I think, of Gang Leader for a Day, where the student learns that much of the power in the building centered around an old black woman who controlled rents and housing grants.)
3TheOtherDave10yTheory aside, I would certainly expect this to be true. You should totally read the Freeman article David_Gerard cites, if you haven't; IIRC she talks about this dynamic a fair bit. I often amuse myself by wanting a clear specification of who is responsible for what at precisely the moment when I am frustrated by my inability to achieve my goals within an organization, and wanting that to stay fuzzy and flexible right up until that moment.

I don't have the baseline to evaluate it, but I consider Gatto's claim that conventional schooling has caused people to have less initiative to be at least plausible-- this might be a more recent cultural problem rather than something from the ancestral environment.

I'm inclined to think that the idea (which I've heard from more than one source) that a fundamental purpose of education is to teach children to do things they don't want to do. While it's necessary for a good life to be able to override some impulses, I believe it's also crucial to teach childr... (read more)

3CronoDAS10yI'm 28 years old and still haven't learned this. Everyone will tell you to stand up for yourself, but nobody wants you to stand up to them.

I find it curious that this posting is tagged Gryffindor and Hufflepuff, when the title and subject matter are entirely Slytherin and the analysis style pretends to be Ravenclaw.

3HonoreDB10yThe posting could be paraphrased as "We generally say we spend too much time doing nothing because we don't have enough Hufflepuff in us. But in a Kitty Genovese situation, where there's an immediate need to act and a crowd of people watching, it's the Gryffindors who leap into motion while the Hufflepuffs do nothing. Maybe this generalizes." Of course, it can take a Slytherin to notice these things [http://www.fanfiction.net/s/5782108/17/Harry_Potter_and_the_Methods_of_Rationality] .
3wedrifid10yA Ravenclaw will often notice too... just far too long after the fact to be useful in practice. :) It is definitely Slytherin thinking that prompts one to harness this knowledge to useful practical effect. :)
1NancyLebovitz10yWhy "pretends" to be Ravenclaw? What would a real Ravenclaw analysis look like?
3Perplexed10yMy choice of words was too confrontational. Substitute "purports". I'm not sure. But it probably would supply empirical evidence or cite authorities rather than quoting a friend and referring to one's own prior writing.
0[anonymous]10yPreliminary analysis / throwing ideas out there has a place too. Also, http://lesswrong.com/lw/lx/argument_screens_off_authority/ [http://lesswrong.com/lw/lx/argument_screens_off_authority/]
1JGWeissman10yIt seems to convey a cynical view of Slytherin style thinking that may be blocking the achievement of Gryffindor and Hufflepuff values.
2Carinthium10yHasn't the very fanfic which introduced this idea pointed out that there are more ways of looking at an issue then just four?
4JGWeissman10yYes, and Less Wrong has way more than four tags, and this post is also tagged "akrasia". I am not too worried about us trying to fit everything into four little boxes from a fictional magical society.

Some are paralyzed for more sophisticated reasons. When you've repeatedly noticed that people who don't take the time to meta-optimize decisions, think through potential negative consequences, actually care about being right or doing the right thing, et cetera, tend to end up doing extremely abhorrent things while remaining self-righteous, and then furthermore notice that when you try to emulate this behavior bad things tend to happen not only to the people you care about but yourself and others' perceptions of you, then you start getting major inhibitions... (read more)

5Louie10ySpeaking of bad precedents, LW is the only community I know of where it's high-status to brag about not being able to do anything.
7[anonymous]10yMaybe it's the ability to admit to status-lowering traits and not the traits themselves that is high-status around here. The idea of status is considered irrational so people surpress those of their judgements that can be clearly recognized as status-based. But simply scoffing at a psychological mechanisms doesn't turn it off so it ends up finding alternative ways of sorting people into a hierarchy. Dissaproval of status dynamics leads to new dynamics in which you get points for appearing unconcerned about the old system.
3JoshuaZ10yIt seems at least superficially similar to how among some evangelical Christian groups it is high status to have been a terrible sinner before accepting Jesus Christ as one's personal Lord and Savior. The analogy might hold more if it were a high status claim here to have once had terrible akrasia before finding solutions.
2NancyLebovitz10yI wouldn't say it's generally high status or bragging at LW to talk about not doing anything. Talking about having trouble getting things done probably doesn't lower one's status here as much as it would in a lot of other places.

Eliezer uses the Harry Potter and the Methods of Rationality "four houses" metaphor alot (see the tags on this post), so I'm writing down what they are (in HPMOR, not necessarily the original Rowling novels) for my own reference:

Hufflepuff. Folk notion of a good person: hard-working, loyal, care about each other, have growth mindset, not to picky about who they'll take as friends. Fair play.

Ravenclaw. Stereotypical nerds: driven by curiosity, not so big on hard-work, ambition, friend-making, and loyalty. Intelligent, witty.

Gryffindor. Courageous,... (read more)

2[anonymous]9yI was always disappointed in Rowling's attitudes toward Hufflepuff (hard-working, loyal and good friends? pshaw! who needs 'em?), and was happy that EY seemed to have more respect for it. :)

So, this theory predicts that people who are not leaders have akrasia, and people who are leaders do not.

5JGWeissman10yIt predicts that nonleaders have one source of akrasia that leaders do not. If it is the only source of akrasia, then you are correct.
0FeepingCreature9yI tend to take the initiative in projects at university if nobody else does. I have akrasia. (Just a belated datapoint)
[-][anonymous]10y 0

I think we just spend too much time surfing the internet.

[-][anonymous]10y 0

Perhaps it's not important whether Yudkowsky is correct on this or not. Perhaps it's more important that this article provides us with a convincing excuse to avoid work. ;o)

4wedrifid10yReally? It seems to associate the null action rather directly with low status of the visceral masculine kind. ie. "If you aren't getting stuff done it might be because you are a supplicating pussy!"
4TheOtherDave10yWell, leaving all the status-signaling and meta-signaling aside, I do think this is basically true of me at least sometimes. When I end up in leadership positions, an enormous amount of my energy gets wasted in managing the resulting anxiety around who I have, or will, inadvertently or unavoidably challenged. Because I am in fact a supplicating wimp, who is nevertheless often (for reasons that have always been unclear to me) treated by those around me as a leader. And that anxiety is a major contributor to my not getting stuff done.
3NancyLebovitz10yHow tall are you? I've read that tall people (tall men?) are apt to be handed leadership.
0TheOtherDave10y5'10" or so. Not short, but not especially tall.
3David_Gerard10yI'm 6'4". And a first child. And loud and extroverted. Pushing people around is obviously my destiny. If only I had something I wanted them to do ...
4NancyLebovitz10yFemales have hierarchies, too, you know.
4wedrifid10yOf course they do. And while there is overlap in the methods of signalling and enforcing dominance within the respective hierarchies the balance of competitive behaviours tends to be somewhat different. In terms of this specific behaviour the penalty for not displaying sufficiently low initiative is less for a low status female than for a low status male while the rewards at the other end of the spectrum are also greater for the male showing leadership than the female doing the same. In respect to this particular trait and all else being equal this would lead to the expectation that there would be greater variance in male initiative taking behaviours than in females. There are other forms of competition and signalling behaviour where the balance of importance leans more towards female hierarchies while the applicability to male competition is somewhat reduced.
5NancyLebovitz10yThe thing is, I have problems with acrasia which don't seem all that different from the men who describe it here. For quite some time (with a partial reversal in the past century), passivity was taught as a quintessential female trait, which I'd say is confirmed by your use of "pussy" as meaning unduly subordinate-- for a male.
0wedrifid10yNothing I previously said confirms, denies or in any way indicates interest in that trait's quintessentiality. Your historical observation does seem accurate, albeit orthogonal. Come to think of it "meaning unduly subordinate-- for a male" isn't implied by my words or the context either."Male who is unduly subordinate and passive", perhaps. But there is a world of difference between a "male who is" prefix and "for a male" suffix. Even then the masculine identification is only loosely implied.
3TheOtherDave10yMm? This [http://lesswrong.com/r/discussion/lw/4na/are_you_a_paralyzed_subordinate_monkey/3mdb] comment -- and in particular the emphasis on "pussy" and the labeling of the status in question as categorically masculine -- seems intended to imply female passivity. If that implication really is orthogonal to your intended meaning, then I'm obliged to sharply reduce my confidence that I'm able to correctly infer what you mean from what you say, at least when it comes to gender. Good to know, I guess.
-5[anonymous]10y
-6[anonymous]10y