Hacker News new | past | comments | ask | show | jobs | submit login
Hierarchy Is Detrimental for Human Cooperation (nature.com)
208 points by rfreytag on Dec 23, 2015 | hide | past | web | favorite | 117 comments



This is an odd paper. It reads like a so-so experimental game theory paper written by an anthropologist. The task was designed to produce Nash equilibria that would fit story they wanted to tell. And they as though they are surprised to see "the behavior of the two partners is only qualitatively similar to" the behavior the game was designed to elicit.

I think that the most interesting finding is buried halfway down.

"Another interesting observation is that for greater rank differences (lower k), the higher ranked individuals are prone to share an amount larger than the Nash prediction." This means that the player in the dominate position behaved more fairly then they needed to.

Given that finding I think this would be a better titile: "Hierarchy is Not as Detrimental for Human Cooperation as You Might Expect"


"This means that the player in the dominate position behaved more fairly then they needed to"

I'd be careful saying more than they "needed" to. If the dominant player had behaved only as fairly as the Nash prediction then human beings would not see that as fair. We all know this intuitively, and we know that other people know it, etc.

As I understand it, modeling human behaviour in regards to reciprocity and judgements of fairness with game theory isn't particularly useful given that this is engaging our evolved behaviour as highly social animals.

It's not my field so I could be completely wrong, but I am under the impression this is very well established.


They rediscovered what Nietzsche wrote about 100+ years ago. High-ranking individuals in any society cooperate with each other more readily than with lower ranked individuals (The Gay Science, Book 1 Aphorism 13 - and many others)


"Rediscovering" 19-20th century philosophy and massaging data to fit it describes much of modern research psychology.


> The task was designed to produce Nash equilibria that would fit story they wanted to tell.

Unfortunately, this kind of malarkey is common in the social sciences today. In fact, more than half of the results are non-reproducible[1][2]. Whether published in Nature or a lesser journal, the current research climate forces me to treat social science results as noise, especially when the results happen to fit one of the major ideologies currently vying for political domination.

Sure, this heuristic is unfair to well-designed, dispassionate studies, but since I'm not qualified to evaluate the merits of each study, I have to make conservative assumptions.

[1] http://www.socialmatter.net/2015/09/18/whats-to-be-done-abou...

[2] http://www.smithsonianmag.com/science-nature/scientists-repl...


>Whether published in Nature or a lesser journal

That's upsetting. Nature is usually quite the stamp of approval. I have some acquaintances who have critical views of Nature, but they dare doubt the rigor of their publications.


The issue I have with these kinds of narratives is they conflate structural and social/power hierarchies. People have spent alot of time rallying against the former thinking it's the latter. The two often correlate in the case of human organizations but are not equivalent. Structural hierarchy also seems to be an intractable problem. It's smarter to construct hierarchies that are more inclusive then to be categorically against the idea.


I believe hierarchy is not detrimental at all (indeed, it is a pre-requisite) for cooperation. Humanity spent the vast majority of its history in loose bands of tribes with no clear inter-tribal hierarchy. Thus we spent thousands of years murdering each other on contact. You may think you are "peers" in the US / West etc, but in fact you are peers in that you've accepted an overarching higher authority in the form of the government - in particular, it's ability to enforce laws by coercion and violence. Thus instead of wondering if the other guy is about to bash your skull, we coexist in relative peace.

International relations, however, lack clear hierarchy - it is a semi-structured anarchy of might-makes-right. Thus we fail to cooperate on many "tragedy of the commons" type problems, such as climate change. Chronic warfare is a defining feature of the 20th century due to this lack of hierarchy.

At the micro level, try running a startup with no hierarchy. You'll find that this will fester conflict, as engineers endlessly argue with no mechanism to bring decisions efficiently and move forward. Indeed, hierarchy is emergent, and tolerated for its virtues by humans because when it does not emerge, the "company" of people dissipate and fail their mission.


I believe this is true too. Hierarchy is a pre-requisite for cooperation.

The most productive OSS projects all seem to have benevolent dictators, don't they?


"Given a longer sequence of repeated interactions, it might be possible that members of a hierarchically organized group would partially overcome mistrust of higher ranked individuals in those that establish generous reputations and these individuals might be able to elicit more contributions from their lower ranked partners. "

To be clear about the implications of this study: Society = longer sequence of repeated interactions. This study is not relevant to thinking about organizing society. It acknowledges it has zero relevance to whether hierarchy is good or bad for human society.


My research group is currently writing up the results for a study where we experimentally studied cooperation that took place over an entire month. We found the very interesting result that cooperation was basically maintained at a high level instead of declining over the course of a 1-hour lab session, as predicted by much previous work.

On a related point, this experiment is a little worrisome because each player only had "9 rounds" to interact with others due to the concern of extended reputation effects. In my experience running this type of experiment, that is not even enough to understand the rules within which people play. (In our case, we had people play 400 ten-round games, 4000 rounds total, and a steady state was reached after about 60-80 games.)

While not completely an "organized society", it does address the generalization limitations of behavioral labs in one important way: the time and span of participation. As you mentioned, this is one fundamental respect where society differs from experiments.


That's why I'm personally fine with doing surveillance without consent for research purposes - observing people behaving naturally seems to be the only way to get results that are not totally messed up by constraints of experiment design. People behave differently when they know they're being studied, doubly so when they know what they're being studied for and triply so if they're being told to "act normal".


You didn't see any horizon effect? People didn't become more likely to "cheat" as the game and the value of their reputation approached an end?


We did see that effect! But, it was pretty minor. In fact, we saw all sorts of effects that would have been impossible to observe in a traditional lab experiment.

One other interesting effect we saw was a day-to-day restart effect: just leaving and coming back the next day was enough to raise the amount of cooperation a significant amount. It's not detectable at the individual level, but in aggregate it's as if going everyone going about their life for a day was enough to instill some new optimism about cooperation at the beginning of a new session than at the end of the day before.

Just that last observation returns to the original point that was made: is it even reasonable to draw conclusions about real-life cooperation from these highly clustered, short-term studies in a lab? We're trying to show that the "virtual lab" on the Internet is a way to get around these constraints.

If you're doing research in this area, check out our OSS projects: https://github.com/VirtualLab


That's a very interesting project. The field desperately needs an alternative to z-tree and psychtoolbox. JS web app seems like a great solution!

I' wrapping up my PhD in neuroEcon at USC. I've started a company to help developers use neuro/econ/psych findings to improve their apps. Let me know when you're done with your post-doc. I might have some applied research questions you'd find interesting.


Interestingly, I think "the field" now refers to an intersection of not just economics and psychology, but also sociology, computer science, and others. More and more behavioral research is going to be done on the Internet and this is going to play a role in designing socio-technical systems.

Although it might sound like an oxymoron, I think computer scientists are going to be both best equipped for and have the most reason to study digital human behavior.


I totally agree. Behavioral science needs an applied branch, and computer scientists are going to have a big role to play there.


Exactly, the paper also expresses hopes that people continue to study of human social interaction:

>> Moving forward, experiments in artificial social contexts like ours appear to be a very powerful tool to examine strategic behavior in socially relevant situations. We hope that our work will stimulate further work along these lines.

It is clear it will be nearly impossible to account for all the factors in one controlled experiment. Note that such "total experiments" in the other scientific fields is challenging as well. Usually one first tries to understand certain fundamental processes, then they move on to the interactions between them.


Undoing hierarchy at every turn is in the zeitgeist. We need to take some time to appreciate when it is useful. My belief is that it is useful for quick action and inevitable when human systems scale.


The problem doesn't seem to be hierarchy, but that it's expected to climb it.

Some people are good in management, some aren't.

But if you're long enough in a job, you have to take on a management position to get paid more or to make decisions in your favor.

Yes the payment thing isn't a big problem in the IT sector, senior engineers often get paid as much or even more than managers. But the decision part is still a problem. Often you end up with stupid ideas you have to execute and the only way to change them is to get rid of the manager, because he won't consider he failed...


Success is as dangerous as failure. Hope is as hollow as fear.

What does it mean that success is as dangerous as failure? Whether you go up the ladder or down it, your position is shaky.

When you stand with your two feet on the ground, you will always keep your balance.

-Lao Tzu


Lovely. Are you a Lao Tzu quote bot ;) ?


Even in a flat organization, some people are good at organizing other folks and getting things done that way, while others are good at sitting down by themselves and cranking stuff out. I'd say the needle is much farther towards "communication and cooperation" (as opposed to "unopinionated worker drone") than in a hierarchy, but the spread of abilities is still there.

The nice thing about being truly flat is that there's no boss to get rid of if they're not working out. You just decide not to work with them.

[Flat isn't totally wonderful and there are still issues with it which I don't care to get into right now.]


> the decision part is still a problem.

"How s plan becomes policy"

http://ogun.stanford.edu/~bnayfeh/plan.html


> Undoing hierarchy at every turn is in the zeitgeist.

Question for folks: Do you really think this is happening or are people just talking about it? I've worked for people who talked a big collaboration/flat structure game, but when it comes down to important projects, all of the sudden, one guy is the decider, and he is only giving responsibility to his lieutenants, who have direct command over the people below. He can count on his hand the number of people who have true control over real resources, which makes it easy for him to reason about where costs are going.

Everyone tries to be nice about it, but my sense is that almost no one actually believes in giving serious responsibilities to a wide range of people. They have their go-to people, and everyone else has to prove themselves by following directions well and maybe they can get a position in that structure at some theoretical later date.

I haven't worked in very many different companies though, so I would love to hear if that's actually becoming a bigger thing in other places. I know Valve is different, but they seem more and more like an outlier to me.


It's useful to make the distinction between structural hierarchies and social ones (and useful to be really skeptical of the zeitgeisty posts that don't differentiate between the two).

So flatten your company structure however you like, there will be a social hierarchy in place as well because these are Humans we are talking about. The people who have interesting things to say about this will talk about how this is the real power of a flat structure, that the social hierarchy is much more dynamic and re-arranges itself on demand and multiple context dependent hierarchies will exist all inter-relating. Your static and slow to adapt org-chart hierarchy just gets in the way.

Of course there are a lot of cons as well as the pros, you will get cliques, introverts, regardless of ability, will be disadvantaged, these subconscious social hierarchy building behaviours incorporate all those ugly human prejudices we know are maladaptive, etc.

But overall I find it a really interesting argument.


This is what really happens, and generally what should be happening. But it's against the zeitgeist - it's undemocratic, it's not inclusive, etc. - that's why you mostly shut up about it and rely on leaders self-selecting themselves when a challenge arises. Humans understand this implicitly, therefore only complain when either a) the matter is insignificant to anything (e.g. things labeled as "just politics"), or b) they have a personal stake in breaking the existing hierarchy (e.g. they want to be leaders themselves).


I read something very interesting a while ago, which I'm not completely unable to find, about feminist movements rejecting, and then embracing, explicit hierachies.

The thrust of it was that without an explicit hierachy, informal hierachies in social structures always emerge. Those informal hierachies are much harder to navigate and much more likely to opress than explicit ones.


https://en.wikipedia.org/wiki/The_Tyranny_of_Structurelessne...

From a feminist perspective some complaints are that lack of structure creates a vacuum in which sociopaths and bullies can run free. Also it doesn't have the capacity to enact top-down diagnostic and corrective action to combat institutional or otherwise widespread behavior deficits (for example discrminatory or sexist attitudes.)


The article you're thinking of is called "The Tyranny of Structurelessness".

https://en.wikipedia.org/wiki/The_Tyranny_of_Structurelessne...


Hierarchical systems do NOT scale well. If you want to scale something it needs to be distributed and in human terms that means equality, action by consensus, local decision making. The properties of the system must emerge from the global rules.

It's in the Zeitgeist because people are coming to understand how monumentally stupid it is to continue operating in this horrific way - both in terms of overall efficiency and in terms of human suffering.


Hierarchical systems absolutely DO scale well. They are just not efficient in the task of adapting to a changing environment.

The proof that hierarchical systems do scale is in the military. They operate in absolute hierarchical fashion. In the past they did so with very limited communication, with headcounts unimaginable in today's organizations.


No they don't.

Pointing in the vague direction of an exemplar system and describing it a positive argument does not constitute an argument.


>Pointing in the vague direction of an exemplar system and describing it a positive argument does not constitute an argument.

Neither does saying "hierarchies do not scale well", or "no they don't", without providing any evidence. So at least you are both "arguing" on the same level: unsubstantiated opinion.


Pretty much every large business and government is also explicitly hierarchical. For a particular firm it's such a common choice one has to imagine it's a good one. For organizations or broader scope than a single firm (industries and economies, cities and nations) independent competition and cooperation seems to work quite well.


Er... what? Hierarchical is the very definition of "scales well". That's what you want to do when you have many instances that all need to talk to each other. Distributed means combinatorial explosion. It's... like secondary-school-level graph theory/probability.

Distributed is what you want to have for a) robustness (because redundancy means the work can be redirected to agents that are still alive) and b) adaptability (because system can restructure itself locally without the need to coordinate with the entirety of it).

Of course you want to apply those two recursively, e.g. hierarchies that encompass many peers at each level, etc., but that doesn't change the point. Frankly, the whole field of distributed computing is one big proof that hierarchies are good and desirable.


> Hierarchical systems do NOT scale well.

Ahem. Ever hear of DNS?


The experiment seems poorly designed if the purpose is to simulate the real world:

"In the cooperation phase, both players ... contribute simultaneously ... to a common pot, unaware of the partner’s contribution." The higher-ranked player then gets priority in determining how the payoff is split.

Who goes to work without agreeing on a salary?? Who co-founds a company without agreeing to an equity split??

The connotation for the so-called "lower ranked player" is also misleading, as in the real world the employer is more similar to their "lower ranked player": the employer usually commits to paying the employee around three months' worth of salary first. The employee then gets to choose to slack off or work hard. So, in a sense, the employee gets to choose how to divide the payoff: the employee always gets the salary, while the employer gets (output - salary). Of course, in the real world the game is then iterated, as the employer gets to choose to fire the employee or continue the relationship.

Any experiment in game theory that doesn't involve iteration is highly unrealistic -- the fact that we have a reputation to keep and have to deal with each other over and over again is pretty darn important! Frankly I'm a little bit disappointed that Nature has chosen to publish this paper, as I don't see what insight it offers.


"The experiment seems poorly designed if the purpose is to simulate the real world"

The stated purpose is to try to reproduce the same experiments and behaviors that observed in other primates.

Some insights I get:

1. In those conditions, our behavior is very similar to other primates.

2. The splitting phase is not influenced by the collaboration phase (I can't avoid to note that you don't choose to apply this insight to start ups).

3. There is some component that we don't understand that account for the difference with the Nash equilibrium.

4. When you have a hammer everything looks like a nail


>> 2. The splitting phase is not influenced by the collaboration phase (I can't avoid to note that you don't choose to apply this insight to start ups).

Vesting is supposed to serve that purpose by allowing a co-founder who doesn't contribute to be fired. Certainly it can be harder to fire a non-performing co-founder though.


> Who co-founds a company without agreeing to an equity split??

Lots of people! This is the normal state of affair in many spin-offs where the product happens before the company and the "ownership" must be divided after most of the work is done.


I remember reading a study recently about the health aspects of power disparity. It concluded hierarchical social structures impose significant physical and mental health risks on those at the bottom. Unfortunately I can't find the source.


An interesting observation related to this is posture.

The posture that show higher status, straight back, looking to the front and not down, relaxed muscles, etc.. need less energy and is better for your health.

Another interesting though is how good food makes you bigger an taller.

So, however is low in the hierarchy, is always less strong and in worst shape, what makes them to get less food, and so on.. It's easy to see how a feedback of this kind begins.

Of course, all this is carry on to the next generation by several dynamics.

I try to remember all this when I hear how some communities are hopeless or just 'don't want to work'.


On the other hand, in my experience absolute equality can also result in stressful situations because of things like diffusion of responsibility and communication difficulties. Sometimes it's just more efficient to outsource decisions to decision makers rather than worryig about them yourself. I think the Goldilocks principle applies here.


You're probably thinking of the Whitehall Study[0]. There's a more recent study about how being the person at the top can be really stressful too[1].

[0]https://en.wikipedia.org/wiki/Whitehall_Study

[1] http://www.bmj.com/content/351/bmj.h6424



Hierarchy may be bad for cooperation, but it is great for quick decision making. Further, I would argue that the decrease in "investment" shown in lower-level members of a team is not a symptom of hierarchy, but a symptom of bad leadership--the appoonted leader doesn't know how to lead, and the decision to make him leader seems unfair.


Never mind how good is the leader.

I think you miss what, in my opinion, is the most relevant point: The collaboration phase doesn't change the splitting phase, they are independent.

That means that never mind what you did in the collaboration phase, what decide the result in the splitting phase, is your place in the hierarchy.

Of course, that leads to those low in the hierarchy investing less, as they know that, never mind what they invest, what decide what they will get is the hierarchy that is already decided.


It really isn't. It's great for quick bad decision making, but so is rolling a die. It takes so much time and effort for information to flow up and down the hierarchy that it takes a long time for informed decisions to get made, let alone good ones.


That's where chain-of-command and executive decisions come into play. In an army field manual for tactics, it states that each level in the hierarchy of a chain of command should be responsible for a certain level of the battlefield; the Commander in Chief decides who to go to war against, the general decides which city is a priority, his liutenants decide how to assault the city, the sergeants decide which building to run to, and the private decides how to get to that building in a timely manner.

Most businesses do not do this. They rely entirely on decisions from higher up, which are often uninformed about small-picture things, even if they understand the big picture.


This doesn't alleviate the big problem with command hierarchies. If the top decision is wrong, and the commander in chief / CEO is ignorant to the advice of their subordinates, there usually is no other recourse than mutiny (war time) or resignation (businesses). This is played out at every level of the hierarchy.


Which command structure is most appropriate will depend on the challenges the business is facing.

If the business has solid fundamentals and basically needs to keep making incremental improvements and watch the money roll in, almost any management structure will do it.

But if a business has needs exit a market with poor fundamentals, and that requires them to close factories, end product lines and fire people? The people getting fired won't be happy about that. So if your management structure requires their consent, you're going to have a bad time.


Whether this is a problem or not depends on the domain. In case of military, it's usually much more important that everyone knows and follows the same plan.


That's all fine and good with the military, which has had 1000s of years to hone their model of their domain. Land/sea/air. Geographical boundaries. But in the typical company, there is nowhere near as clear a picture of what exactly the different levels are. Are Design and Engineering on the same level? Are they underneath Product or is product just part of design? Is Marketing? What about the Data team? Are the server and client two teams at the same level, or are they just different regions of the "world"?

> They rely entirely on decisions from higher up, which are often uninformed about small-picture things, even if they understand the big picture.

That's the thing.... you can't really understand the big picture without understanding the small one. We lionize Bill Gates because he often knew more about about the details of software decisions than the low-level employees who presented plans to him.


Even in the military it's a very flawed process, and, in the US anyway, they go out of their way to enable independent decision making at the unit level.


You know what's worse than rolling a die? Distributed decision making. Yes, moving information up and down the hierarchy takes time, but coordinating a distributed structure takes exponentially more time (and I mean it literally, it's graph theory 101), and with decision making it's usually more important that everyone is on the same page than that the decision is optimal. Army, which pops as an example in this thread, understand that perfectly well, hence the chain of command.


I agree with you. I just cannot understand where people are coming from when they rush to defend the hierarchical organisation of society. It's just like you've attacked their religion.


They're coming from observation. IMO it's the obsession with destroying hierarchies and "democratization" that seems ideology-driven. Centralized solution are good at coordinating and bad at responding to local challenges. Distributed solutions are good at adapting to the local environment, but absolutely suck at coordinating. We have evidence for this evaluation all around us, in every second of our lives, starting with how our bodies are structured (hierarchy which is made of distributed elements at each level), going through basic social interactions (ever been to a YouTube party, or tried to organize a family trip?), and ending with politics.


Ah, the ol' "discredit my opponents by painting them as irrational extremists" tactic. Good show.

Maybe people are defending hierarchy because it's worked reasonably well in many different parts of their lives? Are there any large, successful businesses that aren't hierarchical? What about large governments?


> Hierarchy may be bad for cooperation, but it is great for quick decision making.

Not really. Its good for applying established processes and resisting change, which includes being good at making quick decisions on situations that have been anticipated (but, also, often very bad at making quick decisions outside of the realm of pre-considered possibilities.)


I would think it'd be even more efficient the more unknowns there are. If no one has the ability to listen to the differing opinions of the group and then initiate a course of action by deciding on one or more of the available paths, there are plenty of cases where no consensus is reached and everything goes in a bunch of different directions at once, or nowhere at all (as anyone who has witnessed the parade of "YYYY will be the year of Linux on the desktop!" jokes every year for the past 15+ years can attest to).

Now, hierarchy may not produce the most correct decision making, but it does produce decision making (not even "quick decision making", just "decision making"). Nonhierarchical organizations can't guarantee that. I also think there are large categories of business, design, marketing, or engineering decisions about which groups might never reach a consensus because each position is true and correct and attainable in isolation, but are not attainable when combined together.

I strongly believe that good engineering is all about optimizing compromises, and that it's irrational human nature for us to even think that "no compromises" solutions to any nontrivial problems even exist. We might convince ourselves that the compromises we're making don't matter, but all that does is help hide a potential source of future problems. Compromise isn't a dirty word in engineering, or at least it shouldn't be.

That's what having a hierarchy gives you- a way for people to argue their position as strongly as needed, an understanding that it's natural for people to focus on different ways of looking at the same problem and coming to conclusions that are simultaneously correct and incompatible, and a mechanism (the hierarchy) that ensures action by either enforcing compromise or just picking a winner in situations in situations where a decision must be made but more than one correct but incompatible solution is being fought for within the group.


> If no one has the ability to listen to the differing opinions of the group and then initiate a course of action by deciding on one or more of the available paths, there are plenty of cases where no consensus is reached and everything goes in a bunch of different directions at once

Organizational structures without hierarchical authority structures, e.g. network models, don't necessarily rely on consensus; they may well have distributed authority where decisions can be made at the point in the organization where the issue is encountered, rather than being kicked up a hierarchy until it reaches someone authorized to make it.

> Now, hierarchy may not produce the most correct decision making, but it does produce decision making (not even "quick decision making", just "decision making"). Nonhierarchical organizations can't guarantee that.

I don't see any evidence that that is particular true of hierarchy as opposed to non-hierarchical organizations. Hierarchical organizations are certainly not incapable of analysis paralysis; having a hierarchy doesn't guarantee a decision.

Hierarchies might promote consistency in the decisions that do get made, but that's a different issue.


Hierarchy is defined as priority of access to resources and probability of winning competitive encounters

...and cooperation is some game that sounds like the lower-ranked participant will have a negative expected return (double a common pot -- as long as it's above threshold -- and let the higher-ranked person take more than half).

.

I'm thinking this won't apply very well to situations where everyone has a positive expected return. Like most real-world organizations, companies, etc.


There are number of primate and human studies that suggest that we value perceived justice and fairness and just offering positive expected return is not enough.

A Theory of Fairness, Competition, and Cooperation http://qje.oxfordjournals.org/content/114/3/817.short

https://epub.ub.uni-muenchen.de/726/1/Fehr-Schmidt_Handbook_...

Justice- and fairness-related behaviors in nonhuman primates http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3690609/

Chimpanzees play the ultimatum game http://www.pnas.org/content/110/6/2070


"I'm thinking this won't apply very well to situations where everyone has a positive expected return. Like most real-world organizations, companies, etc."

Don't understand your point. In the experiment, the low ranked persons get a positive return.

It's just that, the return, it's determined, not by the investment in the cooperation phase, but by the position in the hierarchy.

How that doesn't apply?


Because they get less than half of the doubled pot?

...re-reading, it sounds like the pot is actually set to double the threshold rather than doubling whatever's in it. Which means the way I was trying to calculate things out won't work, but also explains why the splitting phase behaved like it did.


It does sound like that's about zero-sum games, but there can also be competition for the rewards of positive-sum games. If we collaborate to produce 100 pounds of bananas and you take 70 pounds and I take 30, that's evidence of a hierarchy even if I also come out ahead.


Ultimately, what this paper means to me is that there is evidence (based on simplified conditions, granted) that people collaborate better when they believe they will receive a fair share of the outcome. This makes a lot of sense to me, especially when comparing to real life.

For example, why is it that in many cases it is the small startup that creates innovative new technologies compared to large and well funded established players? I guess there are many factors at play, but I would argue that at least one factor relates to a more successful collaboration within the smaller environment, which in part is sourced upon the fact that there is promise of a higher payout in the case of success (equity as part of the salary).


> that people collaborate better when they believe they will receive a fair share of the outcome.

There's research demonstrating that group effort and outcomes are detrimentally impacted by increased pay dispersion among the group members. Pay dispersion could form an "informal hierarchy", I suppose.


When I think of hierarchy - in the context of a shared task - I think of who gets to make decisions about how a task is done, but here it seems to be about who gets to decide how the bounty is split when the split happens after the task is complete and bounty already earned. So, testing hierarchy of post result bounty splitting rather than hierarchy of task co-ordination.

What does this relate to in the world? How is it a test of a relevant type of hierarchy? Am I missing something (I would not be surprised, we all have mental blindspots)?


Hierarchy in the natural world is decided by power. And power means that you can get the most, never mind what the others think.

If a chimpanzee leave other to decide because the other is better in the task at hand and, then, they split the bounty, they have both the same hierarchy.

If a chimpanzee leave other in the group to decide how to do the task and then keep everything for himself, he is in the top of the hierarchy and the other is just working for him.

In the not so natural world too, by the way.


> power means that you can get the most, never mind what the others think.

No, sometimes the bounty is having influence. You're not looking far enough up the hierarchy of needs:

https://figures.boundless.com/29841/large/Maslow's_hierarchy...


I wonder if they controlled for specific relationships like the participants never have met prior to the trials or anything of that sort. I can't find anything the article to indicate that.


Did Tony Hsieh commission this study? I'm just wondering ;-) jk


If you exploit people without giving them a fair share then they will eventually revolt.


Hierarchy, not money, makes the world go round.


Neither of those categorical statements are true. The world is a heterogenous overlaying of all possible political systems working simultaneously. There is capitalism (money), there is delegation (hierarchy), there is anarchism... often in the same room.


Linear hierarchy may have been effective for quick decision making in the age of paper and print but as Marshall McLuhan and Barrington Nevitt point out in their book Take Today: The Executive as Dropout (1972)[1], in the age of electric media distributed networking and flat hierarchies become the most effective organizational form and "Johnny on the spot" is the one who makes the quickest and most accurate decision as opposed to the executive:

McLuhan: Take today means that at the speed of light, today includes all the past that ever was, and all the future, it's here now.

Nevitt: That is we can retrieve the past, we can be in touch with the present, and the future of the future is the present, so we are in touch with all times now, today.

McLuhan: But let us notice that in a world of simultaneous information, you have basically an acoustic world, not a visual world with a point of view, not a positional world from which to look at the future or the past, but an acoustic world in which you are bombarded simultaneously by every kind of data from every direction ... the simultaneity of information means that you live in a world which is simultaneous in terms of its information structure. From every direction you have information electrically. Now this creates a new kind of space.

Nevitt: Yes, it is a new kind of space for us and yet an old kind of space for humanity.

McLuhan: Well, for pre-literate man.

Nevitt: Yes.

McLuhan: The space in which we live is identical to the space of pre-literate man, pre-visual man, it is acoustic space.

Nevitt: Post-literate space and pre-literate space are similar.

McLuhan: Yes but the question that seems not to have entered most minds is, what is the structure of acoustic space? In fact, most psychologists have never heard of acoustic space and they don't understand its properties. Its properties are those of a sphere whose center is everywhere and whose margin is nowhere.

Nevitt: That is, no boundaries, and no single center, but centers everywhere.

McLuhan: centers everywhere, margins nowhere.

Nevitt: No points of view. ...

McLuhan: The old structure was made up of fragmented specialist jobs.

Nevitt: Exactly, and hierarchies of responsibility which were delegated from the top down and channels of communication which were supposed by the organization chart upon which they flow from top to bottom, and bottom to top.

McLuhan: So the theme of our book is that you cannot use an organization chart with electric services added.

Nevitt: So that "the executive is drop out" means three things, it means first of all: the man who finds himself as top executive in a large organization, gets out of touch with the action, that is, he gets so far out of touch with what's happening, that as a human being, he has no more satisfactions in the process. That’s the first thing it means. The second thing is that specialism as a means of coping with this acoustic space, doesn't work, that is it has dropped out, it is no longer effective. And the third thing is, that in a knowledge environment, that is where the information is everywhere, like acoustic space with centers everywhere and margins or boundaries are nowhere, then every man knows how to be an executive, and "johnny on the spot" is the man who really makes the decision.

McLuhan: Oh, the hijacker is a nice case of a man who wants to be an executive and wants to take over an operation but he also wants coverage, he wants part of that global theatre for himself, he wants to be in the center of the show.

Nevitt: And the global theater itself becomes possible, and the hijacker becomes possible only in a world which is such that you are on camera.

[1] https://www.youtube.com/watch?v=HTyzjC_s-kk


>flat hierarchies become the most effective organizational form

Luckily, as folks with an understanding of data structures we can see this is a very naive approach. 'Electric services' must have seemed very new and frightening to inspire this scattered response.


Care to elaborate?


Caching, caching, caching. And load balancing.


McLuhan and Nevitt are talking about human organizational structure, not technical design.


Information flows are information flows whatever the medium.

Hierarchies are efficient, that is why they exist.

If we want to know what branch of a tree structure to find information we examine a node after gathering info on how that data structure is put together and work our way down. If we want to know who is responsible for some aspect of an organization we ask the boss after gathering some basic info on how that organization is structured and work our way down.

As much as electronic signals have reduced friction, having a complete lack of hierarchy will quickly eliminate any efficiency advantage that provides as long as there is any communication friction whatsoever.


Just to reiterate, the study we are discussing and my original post are about inefficiencies in the design of human social hierarchies and not inneficiencies in data structures.

>Hierarchies are efficient, that is why they exist.

They exist because they are legacy systems that were once efficient or at least practical.

>If we want to know who is responsible for some aspect of an organization we ask the boss after gathering some basic info on how that organization is structured and work our way down.

This is exactly what is now inneficient. When information was centralized due to the constraints of paper and print media this may have been efficient. But now that information is decentralized and distributed it becomes much quicker to simply query the information instantaneously yourself rather than having to go through a third party.

Centralized information leads to pyramid shaped social hierarchies.

Distributed information leads to flat or spherical internetworked social organization.


> it becomes much quicker to simply query the information instantaneously yourself rather than having to go through a third party.

But we are still going through a third party to make that query. We are not doing it ourselves. Information is more centralized than ever. Google itself organizes it into a hierarchy. It may be more fluid than a bunch of people shuffling papers in an office or government building but the basic structure is still there. Besides, raw information is only one small part of the story. The other half of information is generation which is still boxed by responsibility (a long term information store in human form) that is most often, even today, most easily navigated to through a hierarchical organization. It might not look like the old hierarchies but that doesn't mean they are not there.


You seem to be equating the absence of hierarchy (which is critically about authority) with the absence of structure for communication, but the common alternative to hierarchical organization -- network organization -- still has that, it just detaches it from authority structures (and communication lines tend to be shorter, because paths which would be lateral on a hierarchical org chart and require steps up and down the chain of command are, instead, direct connections.)


Sure, but if you have a general graph (network) of more than a few nodes all trying to communicate meaningfully with each other you quickly run out of bandwidth.


If they are communicating through a heirarchy, you've got the same total communication volume over a smaller number of channels; either you don't get as much information integrated where it needs to be, or you do so on a slower best-case path; in addition, each non-leaf node in a heirarchy is a single-point-of-failure for all communication (and decision-making) involving the leaves underneath it.

Heirarchy is both more failure prone and has worse best-case communication time between nodes than network, which is why many nominal hierarchies end up routing around those limitations and acting like networks, where formal authority follows the hierarchy, but much practical communication and decision-making is done on a network model, because real humans are pragmatic and route around the failures that inevitably arise in a hierarchical organization. This ends up compromising the ability of the hierarchical command structure to impose uniformity.


Can explain why Republicans are less empathetic towards their fellow citizens, as they tend to be more about libertarian self empowerment, where they are kings of their own world.

Example: If you have a gun, you are placing yourself above others in order to gain power over them, and are therefore less likely to cooperate with your fellow citizens by offering to pay for their health care through taxes.


"Republicans are less empathetic towards their fellow citizens"

Seriously?

It's well established that conservatives/republicans donate more to charity.

Here's Nicholas Kristoff's column on it:

http://www.nytimes.com/2008/12/21/opinion/21kristof.html?_r=...

Not only do Republicans donate more in absolute terms to charity, they also have lower average incomes, which means that the average giving as a portion of income is much higher. And it's not just donations to their church, Republicans are even more likely to donate blood than Democrats!

I'm not writing this because I want to claim the moral high ground for the Republicans, I see Republicans and Democrats as each having different takes on how best to help the poor. Neither side has a monopoly on empathy.


"they also have lower average incomes, which means that the average giving as a portion of income is much higher"

I'd be cautious about comparing the average of one distribution with the average of another. Unless you know how the quantities are distributed (and income in particular is unlikely to be normally distributed), you risk making some misleading conclusions.


> It's well established that conservatives/republicans donate more to charity.

Liberals solve social problems through taxes and big government, instead of donations.

Neither side may have a monopoly, but one side is far more effective than the other at solving social problems.


I honestly don't even know which side it is that you think is "far more effective" at solving social problems. Do you think it's the Democrats? Because "big government and taxes" are also far more effective at creating social problems that local charities. Have you read about the corrosive effects of welfare on black communities? Have you ever joined in complaining about the inability of public school systems to successfully adapt and harness changing technologies? Have you noticed that our biggest social problems just "happen" to also be located right where Democratic policies were enacted (i.e., the policies preceded the social degradation)? All of these things are certainly at least debatable, but they are also certainly not so obviously false that they can simply be discarded without consideration.

"Big government and taxes" may theoretically be more capable of solving problems, but there's a big gulf between theory and reality.

And back on point so this is at least less an off-topic diversion than it otherwise would be, are you A: inclined to agree that hierarchies are bad for cooperation even as B: you declare that the largest hierarchy in the world, "government", is better at solving social problems than distributed non-hierarchical networks? I don't know, because you, mozumder, personally didn't say, but I rather suspect there's a lot of people here holding both those beliefs simultaneously without noticing the profound philosophical conflicts they have.


Have you read about the corrosive effects of welfare on black communities?

No. I have read about welfare reducing poverty. I'm sure you have, too.

Have you ever joined in complaining about the inability of public school systems to successfully adapt and harness changing technologies?

No. I have read about public schools outperforming private charter schools.

Have you noticed that our biggest social problems just "happen" to also be located right where Democratic policies were enacted (i.e., the policies preceded the social degradation)?

No. I have read about liberal policies solving social problems, as usual.

All of these things are certainly at least debatable, but they are also certainly not so obviously false that they can simply be discarded without consideration.

Why not? I just discarded them without consideration, and you can, too.

The question isn't whether or not liberal policies solve problems far more effectively than Republican policies. The question is how quickly you will accept that truth.


"No. I have read about welfare reducing poverty. I'm sure you have, too."

Only in the short-term. In the long-term, it creates generations of people completely dependent on the government, never really getting the chance to get themselves out of poverty.

All of my extended family are like this. They don't work because the government gives them enough to live.

"No. I have read about public schools outperforming private charter schools."

Where is the proof of this? In my area, the public schools are terrible. It sickens me that I have to pay taxes to support them when they just continue to spiral downward. It's also impossible to get rid of bad teachers, because whenever some sort of solution is suggested (for measuring effectiveness), the unions come back and say it isn't possible.

"The question isn't whether or not liberal policies solve problems far more effectively than Republican policies. The question is how quickly you will accept that truth"

Most liberal solutions that I've seen leave out human nature and a history of complete failure and re-dress it as something 'new'.


it creates generations of people completely dependent on the government

Where are these magical people that aren't dependent on government?

Are you one of those? How do you travel? Do you build your own roads? What do you do for mail? Do you use UPS for all mail, and they use their own roads?

Did your employees learn literacy from private schools? Or did government teach them? How did they learn to follow work instructions?

I would like to find these magical people that aren't dependent on government. They sound like they are awesome people. Maybe they also have guns to defend themselves from invading armies, and have fire-proof houses?

You are dependent on government. You can ignore your ego that actively harms you by telling you that you aren't dependent on government. The quicker you learn this, the better off you will be.


I'm receiving services in return for the incredible amount of money, in taxes, I pay into the system.

In reality, the government is dependent on me.

Our goal should be to encourage more people to be like me (pay more into the system than I will ever use) so people that truly need it can use it.

Welfare and an abundance of social programs encourages just the opposite: live on welfare and never make enough to give back into the system.


In defense of grandparent, let me point out that you ignored the adverb "completely". There is "dependent on government" and then there is "completely dependent on government".


> It sickens me that I have to pay taxes to support them when they just continue to spiral downward.

This is a feedback loop. I'm not saying that throwing more money at public schools will help them get better, but taking money away will certainly make them worse.

> Most liberal solutions that I've seen leave out human nature and a history of complete failure and re-dress it as something 'new'.

Actually agree. Most liberal policies are somewhat naive. But on the other end you have the free market people crowing about benefits of deregulation and wanting to disassemble public services because "I don't want my tax dollars going to [parks, buses, schools, etc etc] I don't even use!!"

There's a medium ground here. The point of society isn't to preserve your tax dollars or to give everyone welfare. The point is that when we all live together as a group, we all do a lot better than we did when it was six of us in a group starving out on the tundra.

There are some public services that we just shouldn't live without, especially if we want to stay competitive as a country. However, some of them need complete overhauls or just shouldn't exist in the first place.


I agree with you here. We shouldn't be taking away money from our public school systems. We should be fixing the current situation and making people accountable for the wasted money.

My parents grew up near detroit and they told me about how the state poured money into the school system for years.

They are now some of the worst in the nation, with abysmal graduation rates.


How much did they pay into the public schools? Did they spend as much as public schools around the Washington DC area (best in the country?) And, how much was it compared to what they would pay to private schools?

Yah they didn't pay enough into the system.

A real system is a lot more expensive than what they think.

The solution to the education problem is: spend more money.


See: detroit. 5x was spent for 30 years until the entire system collapsed when enough tax payers fled the city.

This is the result of your plan.


> I honestly don't even know which side it is that you think is "far more effective" at solving social problems.

...and I really hope that it was intentional. :)


Taxes only go so far. Higher wealth elasticities of demand among those of higher income means less sensitivity to it, and the greater purchasing power means a lower opportunity cost to avoiding those taxes. Then, once the taxed income is retrieved, you must rely that the state apparatus will properly allocate it without pork-barreling, corrupting or shoving some omnibus legislation on transit until it finally becomes executive procedure.

This means that, at some point, all forms of taxation eventually boomerang and become regressive. Even when you introduce tax credits for lower income citizens, the bureaucratic and personal costs to managing them also tend to become regressive since third-parties will usually be paid for consulting and navigating across the labyrinthian tax code. This creates a situation where you open up for rent-seeking players whose business models depend on the aforementioned labyrinth, and who as a result get to redistribute income via wasteful inefficiency, lobbying and then offset the gains from tax credits by the losses of navigating them.

"Big government" inevitably leads either to totalitarianism, or to "kludgeocracy" as in the United States [1].

[1] http://www.nationalaffairs.com/publications/detail/kludgeocr...


That's your assertion. It's always fun and easy to spend other people's money which is why the liberal solution is popular. Building up a debt of, for instance, nearly $18.7 trillion and rising has no eventual knock on effect and is quite in order? Probably misinformation but I saw an estimate that as much as some 70% of the American public depend in significant ways on their government for financial survival. You call that 'solving social problems'? Sounds ultimately terminal to me. But that's just me sounding off. I do gather that some really well-informed folk do take that view and can back it up.


Make that 100% of the American (or really any state's) public. Do you think your life would just go on without any troubles at all if government suddenly disappeared?

I guess what I'm trying to say is that everybody[0] depends on the social structures around them (which includes the government). It's just that for whatever reasons, often pretty arbitrary ones, there is a consensus narrative which labels some of those dependencies as "bad" and others as "good" (or "normal"). Then people invent lots of after-the-fact rationalizations and categories to try to ignore the "good" dependencies and focus only on the "bad" ones.

[0] There may be some persons out in the woods who live in a cabin they built using their bare hands, with only the tools that they built themselves, without using oil-based fuels or something like the weather forecast. I'd say it's safe to ignore those persons for the purpose of this discussion, if they even exist.


>Do you think your life would just go on without any troubles at all if government suddenly disappeared?

The classic "I'm going to pretend your criticism of one part of government means you want to completely abolish all government" strawman.


I'd agree with you, but vixen99's post does not actually articulate an awareness of criticizing a particular part of the government. There's talk about "dependence on the government" by 70% of the public (probably the incorrect number according to what they really mean, but let's go with it). My point is really that this statement is meaningless because it's really 100% of the public who depends on the government when you think of it.

For example: people who work for recording companies undeniably depend on government handouts for their financial survival (think copyright law). Do you think vixen99 intended to include this group of people or not?

So my criticism, which you unfortunately missed (because I admittedly could have been clearer about it) is that the language used to attempt to discuss an argument is already wrong.

The intended argument itself may have merit, but as long as you talk about it in a misleading way you're not helping the debate.


Much of conservative charity is to churches and religious causes. It's the difference between praying cancer away and investing in cancer research. I can go on for a while about the difference between state offered social services in Massachusetts being far, far more effective than church offered social services in Texas. This difference in approach also has to do with 45.83% of people in Massachusetts are Catholic and with political beliefs in line with the Kennedys and only 18.72% of people in Texas are Catholic with many people from other forms of religion at war with government that wants to teach children science and evolution in schools and treat homeless people with medication, stable housing, and education rather than treating homeless people with prayer and preaching about Jesus.


This is actually not true, here's a quote from the link that I posted.

"Conservatives also appear to be more generous than liberals in nonfinancial ways. People in red states are considerably more likely to volunteer for good causes, and conservatives give blood more often. If liberals and moderates gave blood as often as conservatives, Mr. Brooks said, the American blood supply would increase by 45 percent."

It's true that a lot of the money is donated to churches. But even when you ignore the money that is donated to churches, Republicans still donate a higher percentage of their income to secular charities than Democrats do, donate more blood, and volunteer more.

"many people from other forms of religion at war with government that wants to teach children science and evolution in schools and treat homeless people with medication, stable housing, and education rather than treating homeless people with prayer and preaching about Jesus."

I live in Texas, and I can tell you from first hand experience that this is a ridiculous and false stereotype.


The language used in the NYT article you referenced is fairly loose for something that is trying to draw conclusions from statistics. And it is not hard to find people questioning the reliability of the source work, Who Really Cares by Arthur Brooks.

For example, you and the article's author state that conservatives donate more blood (which may well be factually correct) and seem to assume that there is a direct causal relationship from political belief. It may instead be that non-conservatives are disproportionately likely to belong to communities that have negative attitude to blood donation. See [1] for a possible example. There may be many other confounding factors.

"People in red states are considerably more likely to volunteer for good causes"

"People in red states" are not all conservatives. If those people in aggregate are more likely to volunteer then the causal factor(s) may have more to do with geography or regional history than political outlook.

[1] http://www.ncbi.nlm.nih.gov/pubmed/20840534


Republicans are also much older than Democrats. A very large portion of Republicans are retired and collect social security which is paid for by younger more likely to be liberal Democrats who don't have to time to volunteer because they are working to support their families and pay into social security to support retired people who just happen to be more likely Republican. The popular conservative argument against a basic income is that people will just suck the tit of big government and be lazy. That retired people on social security volunteer with their guaranteed minimum income is proof that a basic income does work. People instinctively care about other people and they do not want to be bored. Thank you for pointing that out.


That's not what is says. It says that if you place yourself above the others, you will need to contribute more. Probably because others will be reluctant to do so out fear of you grabbing most of the earnings.

It might mean that you need unions for people to feel that they are getting their fair share, making them more invested.


I'd be afraid to get a union involved, because in my experience they make terrible decisions (reducing "full time" from 40 hours to 32, making it near-impossible to fire troublemaker workers, allowing each worker to be used for work they're not actually trained for with no prior warning).

Then again, the only unionized place I've worked in is a Kroger.


I'd love to see a study of the relative effects of hierarchy and cooperation in problem-solving effectiveness and inventiveness, with some interest in what positive non-toxic hierarchy looks like.

The right kind of leadership really does seem to make everyone more focused and effective. But there are so many wrong kinds of leadership that the best outcome happen very rarely; bad leadership seems to be immensely destructive of group ability, and is much worse than ad hoc cooperation with no formal leadership.


It seems to me that FOSS projects like Linux have a "positive non-toxic" hierarchy. The hierarchy there is only based on merit and imposes no power constraints over the members, who all have the same access to the source. If you disagree with the elders, you can fork.

I think if your project or team needs hierarchy for the decision-making, then the hierarchical order should have no power (for example, the leaders could be voted in democratically), because that way you get the positive effects and avoid the negative effects.


No, I don't think that hypothesis has anything to do with the study at all. Just for starters, cooperative behavior declined in the lower ranked individuals in the study. From the abstract: "cooperation declined in the presence of a hierarchy due to a decrease in investment by lower ranked individuals."

I admit I have only read the abstract so far. Did you even read that?




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: