I* watched a video by Professor Dunbar on Dunbar’s Number, which set me to thinking again about the design constraints and optimal group size for a group in a modern society. A different warning**. A note on methodology.***
Our common goal is increasing the level of civilization. That is a human activity, one we all engage in daily, one which affects us all, daily. The fact that our level of civilization falls every now and then means we don’t know enough about the process, falls are bad for people. Increasing levels of civilization are good for everyone, we need to think about how to improve our social organization so as to increase the competence of the total civilization while avoiding cliffs.
A civilization’s future is not predictable, as it is a natural system, an evolutionary system. We know that evolution is a mechanism that finds and accumulates optimal elements of physiological and other systems, and that evolutionary algorithms are highly efficient in searching huge, high-dimensioned ‘spaces’ for optimal solutions. Thus we wish to organize for efficient evolution.
Civilization is an evolution of human societies, made up of individuals in groups. Characteristics of individuals and groups and changes in those characteristics thus control the rate of evolution toward higher levels of civilization, as well as the dramatic falls. We wish to maximize the rate of rises and minimize the catastrophes that loss of civilization always are.
That is a problem of networks, individuals and groups and institutions are modeled as networks, with nodes having characteristics which are weights in equations governing interactions, messages, from node to node. The model is not the reality, of course, but it makes clear that moving messages positive to civilization’s development, no human left behind, through networks of minds and doing so without coercion or dishonesty must be the goal. That is necessary to make them maximally active in a world-community, and thus most active in working on our mutual goals for a healthy and wealthy future for all.
That must be the goal if we are to use all of humanity to improve our civilization, a cooperative bootstrap, generation after generation. Not idealism, optimal operation of the obvious engineering specification. Why settle for provably-less?
The primary constraint on human society is the information learning, holding and processing capability of individual human minds and of the largest effective group of aggregated human minds that combine on individual issues. Within those capacities must lie all possible sustainable human civilizations. Within that set there is an optimum for all, short-term and as long-term as nature allows.(1)
Within those individual and group capacities, constraints, our problem in designing and operating civilizations is to make individuals as effective in using their minds as possible and the effective group size as large as possible and thereby optimize operation of the societies within it. This is true however you define ‘optimize operations’, but as the effective size of the collaboration increases, that definition will generally include more of the values of the minds in the collaboration, and is thus subject to the same optimizations.
Constraints to largest-effective aggregations are :
- Limited human brain power, and so limited amount of an individual’s information and thought and action.
- Limited human life, reducing the amount of information can be integrated in one mind.
- Our current methods of aggregation, which includes the training in methods of combining our individual capabilities.
- The effectiveness of the methods we use to combine our individual capabilities.
- The number of individuals in the network who can combine their mental and physical capabilities, which is in turn limited by all of the above and the technology employed.
There is not a lot we can do in the short-run wrt individual raw human brain power and work output.(2)
Likewise, there is not a lot we can do short-run wrt the length of human life.(3)
So, the effectiveness of aggregating minds is the target offering the largest opportunity for improving our evolution of civilization.
Methods of aggregation are astonishingly variable in every dimension, e.g. US Marines vs Wikipedia, and the outputs are more so. Although much study has been done, 10s of 1000s of papers published and seminars and training sessions presented, best practices for extracting maximal efficiency and effectiveness from groups in any given setting are far from certain and rarely practiced if known. (4) One thing that is clear in all studies of people working in groups, trust is a necessary component of group effectiveness.
A functioning society demands wide-spread trust in order to make the social units with the most-effective aggregation methods and processes as large as possible. Losses of widespread trust inevitably result in a fall in the system complexity. However produced, those are costly in every dimension. Trust takes a long time to recover, perhaps 100s of years. Maximum trust takes social contact, personal relationships and personal understanding. If we want a society of trust, it must be a personal society. The most trusting relationships in society are within families and clans.
Thus, this discussion focuses on families. First, people learn to trust in their families, and learn the basics of cooperation and teamwork. Second, families are still a major component of societies, tho much diminished in importance compared to historical times. This considers changes to family structure and how they might improve the functioning of all other institutions.
Families, extended families and clans are termed ‘family’ hereafter, nothing is implied about its structure in a social arena. This only considers the issues of scaling trust from a family base, although the aggregation of knowledge about optimizing individual lives and life’s human satisfactions for individuals that a family represents is an equally important factor.
Individuals can directly trust people they personally know well enough to trust, a zero-th level relationship. That means 0th level trust relationships, with their requirement of personal relationships, are limited to the maximum number individuals can know personally. That is Dunbar’s Number, generally thought to be a maximum of 150 people, the number at which migratory tribes generally split due to limits on the carrying capacity of those landscapes using primitive tool sets. Although individuals can know and know of perhaps 1000 in business and political arenas, most don’t due to limits on time, interest, attention. (5)
Social networks, however, can trust proportionately to the size of 1st-order networks, the number of people who you can trust as a result of someone you know trusting them. That is approximately the number of people in your 0th-order network * the average number of people in each of theirs, less the overlap. Generally, a good fraction of N^2, as social animals we build and maintain connections up to our mental and time limits. Modern communications have increased the number of people we communicate with, acquaintances, but not the size of our most intimate group. This has not produced corresponding levels of trust.
Humans strive to create 0th-level networks, it is a major part of growing up to socially bond with peers and create life-long friendships. In the modern world, those relationships may be long-distance, but nevertheless people depend on long-term friends’ opinions and advice. The fact that we all go into different professions and business is an advantage, groups of friends have a wider set of understandings as a result, freely shared. One of the ways we judge people’s trustworthiness is the kinds of friends and lengths of friendships.
Nevertheless, bonds within families are generally stronger than those with non-family, even best-friends. Also, family is better connected, and there is more sharing of resources and information flow within most families than with even the best of friends. Likewise, 1st-level links based in family are likely to be stronger, everyone does more for a friend’s relative than a friend’s friend. Another of the ways we judge people’s trustworthiness is their family and how close they are to their family. Thus, in order to maximize a trusted network, we want to maximize the size-strength of our social networks, not just their size.
That means big families because a limit to the size of a maximum-trust 1st-level social network is the average size of a family. The size of the family and and the number of individuals each may personally know constrains the total number of people a family can 1st-level most-trust.
In previous times, families were generally much larger, meaning more people to have more relationships outside of the family. In fact, with clans as the social unit, those networks could be very much larger, and that must have always had some advantages, as it was the common social organization well into the modern era, and still is for large parts of the world.***
Further, families and clans may have a reputation as a unit, e.g. being carefully honest in business and personal dealings. That reputation is something that individual family members are taught to live up to, an advantage in their dealings, and an advantage to people who deal with them, because more trust means less checking of everything, so less cost to any transaction.
How then, to organize a modern family that is a stable base for humans to grow and live within and allows a more optimum, trusting combination of human minds on larger tasks? This essay does not consider the organization, merely the constraints of size and human minds on the organization of a family and the resulting possible societies.
Assume 100 as the largest family, with 150 possible if human longevity continues to increase and we have 5 generations living at the same time. With any set of mixed-age people, there will be the usual closer relationships within the larger group, child to mother, child to father, parents, current lovers, specialized work groups, etc. Nevertheless, a family of 25-50 adults with their children, children’s spouses and grandchildren is a normal human group, a large band in hunter-gatherer times.
That unit scales well in our age of modern communications. Everyone in a large-scale society of such units will know many people outside of their primary social group, local community, business, professional. Some individuals can probably keep context of up to 1000 people in total, but probably 250 will be an average in era of efficient and effective communications, tho that set will change throughout most people’s lives. Assuming 100 family, 150*50 is a lot of social and social-business connections.
There are many reasons to organize a society out of such large-family units. First, it is optimal for producing a society of cooperative individuals, as parts of extended families. If individuals do not learn trust in their family, it is less likely that they can ever learn to trust. Living and working in a large-family can be managed to much improve cooperation and trust and family loyalty. Clan leadership and management isn’t something moderns have focused on, but there must be much old wisdom that can be resurrected.
In a society intending to trust, the goal of every family would be to extend their trust network by choosing to know and do business with the most trusted families. They would best do that by becoming the most trusted themselves, trusted in opinion and in business because transparent and trusting transparency, doing their part to check on the accuracy of others, the correctness of their transparency. In this world, every family-business knows they are in the trust business, first and foremost. By emphasizing that, lower costs and more rapid evolution of business and society can be achieved for everyone.
This is not idealism, this is hard reality, analysis of the requirements of society that function to a specification. Trusting societies work better and can attain higher levels of civilization. This benefits the trusting very directly, as well as increasing life’s satisfactions and lowering costs of everything. The problem is making the transition, not with the goal, which is entirely feasible, social engineering ourselves for a better future, as civilized beings do.
How to check this line of thought? First, is it plausible from what we know of human behavior? It seems to fit with a need people have to join small, intimate groups and be intensely social, families in less-modern times were very important for an individual’s social standing and family loyalty was an important part of enhancing that social standing. Throughout history, people have joined into sort-of-family communal living arrangements of many types and adhering to many different ideologies and religions, and most business in pre-modern eras was based on families and extended families. Most of the communal living arrangements were also businesses, e.g. every monastery was more than one business, and many provided food and lodging for wayfarers. Many communes were also businesses in all ages. This is an extension of hunter-gatherer standard behavior and so makes use of the human characteristics we evolved over several million years in such groupings. Also over the last 10K years in intense wars.
At a somewhat higher level of abstraction, my previous Night Thought suggests that we measure this social unit by its information production and entropy production rates as compared to our current system. If it is more stable, sustainable and produces more entropy per person per year, it is a winning design.
Information is anti-entropy. More intense social interaction of larger social groups produces and consumes more information, Facebook is proof of that.
More sustainable and stable is not provable for any in-reality system, designed or natural, but larger families will be more local, and also more in touch with more aspects of everything. The ‘understand global, act local’ and longer-term view that such a family would have favor stability, evolution not revolution.
*Generalissimo Grand Strategy, Intelligence Analysis and Psyops, First Volunteer Panzer Psyops Corp. Cleverly Gently Martial In Spirit
(1) That is, so long as we don’t do our own civilization to death by allowing short-term wins for powerful minorities.
(2) Although we could stop using pedagogical methods that cripple human minds and stunt bodies, rather emphasize learning points of view instead of sterile facts in a misleading context. In work output, companies that put more responsibility on people ‘doing the work’ for organizing their work efficiently consistently do better than their competitors. Cooperatives and smaller companies without any management at all seem to do OK, by anecdote.
One point of view that any education should include is that part of every life is continuous work at extending our grasp of reality and using that to increase our effectiveness in attaining our goals, and that this extends into daily work, daily interest of all kinds.
(3) Tho we could stop the many factors which so obviously shorten lives, disarmed citizens, armed and centralized governments, pollution, choices of drugs, foods and lifestyles.
(4) Freedom seems to work best, most of the time, but it is rare, and many counter-examples encourage the powerful. For example, Stalin’s engineers in prison camps produced some great designs for military weapons, tho that model has roved unsustainable. Things like that are one aspect of ‘unsustainable’ and happens with too-strong government of every persuasion.
Along with ‘avoid too-powerful government’, among the few things most social scientists looking at, the efficiency and effectiveness of organizations is inversely correlated with size and age. Organizations uniformly devolve unless both small and lead with un-remitting attention to dealing with both the group’s cooperation and the aspects of external reality that are important to their success. Nevertheless, the goal of every organization is to grow, therefore implicitly to devolve, sooner or later. Organizations of genuinely superior performance are rare and rarely copied, sooner or later to grow and devolve.
Our current methods of aggregation cripple our ability to cooperate as effectively as we could, e.g. the wide-spread adoption of ‘Political Correctness’ and other power relationships which prevent trust in discussing and dealing with common problems. This is more common in larger organizations and organizations less connected to immediate feedback, e.g. market discipline, one aspect of which is individual support.
Most functional organizations try hard to deal in personal terms as much as they can. Using the same close to home supplier, knowing people in common, having reputations locally, makes for trust, which makes for lowest cost because your losses are lower, nobody sues you, and you don’t need to sue anyone.
Having a family-clan-business as the fundamental unit of society makes people responsible for their joint reputation, and honesty, integrity and charity in their community will come naturally as a result, or people won’t deal with you.
I do not claim this will be an honest society, unless it is an entirely voluntary society. Any coercion generates reasons to hide one’s actions that evade the effects.
(5) Those are the reason that platoons and companies are the same size in all military units of the armies throughout history : A platoon is the biggest group one mind can control in detail and a company is the largest group in which every person can know everyone else and keep their capabilities in mind. Other organizations are organized in the same hierarchy, the optimal organization for different types of missions has received endless attention, and our largest organizations continue to violate everything known.
**Another warning, this one of my complete lack of intellectual foundation in this bit of thinking. Most people are like me, and it doesn’t generally matter how un-founded our opinions are. Unless you do it in public, standard leader behavior. Unless you are a writer. Or some variety of mental case.
Or open system designer of social systems, as I am. Designer of the next standard unit of a global civilization, in fact. We open system designers who expect to make a difference have to A) Do this in public. B) Be right. C) Get everyone to sort-of-agree, win social acceptance, tho most won’t change their behavior. D) As part of that, generate examples of wins with our designs.
Public is easy, we already had blogs. Publicity was a bit harder, but intelligence operatives have a network, so it is pretty much standard social networking, scratch my back, etc. Intelligence operatives and PR agencies and political campaign clubs and …
Once you grasp social media, and are willing to work your ass off doing favors for people, you can make that publicity happen, whatever the social group. Public Relations knows how to sell people, to get social acceptance for people and ideas. For people, money of any origin, fame of dozens of kinds, infame of most kinds, including pedophilia, war crimes, large-scale theft of the wealth of ordinary citizens, leading mass rapes, … Sorry, I am not joking at all, of course infamy is as good as fame, any day of the week, in the highest corridors of this society and its ruling elites. Any PR agent can handle the situation, they rehabilitate preachers, politicians, war criminals and actors all the time. Not quite the power to rehabilitate pedophiles yet, but the power to have the media ignore their crimes is almost as good.
Being right is the hard part. You have to think about things to find all of the bases your design has to cover, this is one of those efforts.
***Methodology and assumptions. This is a design for a civilization made up of links between individuals and families, very similar to the society of an old world novel such as written by Tolstoy.
For purposes of calculations, we consider a family to be 50 adults, and the entire adult society to be constituted from like families. Families may grow larger than that, children and adults they marry when grown, but 50 is the core of the stable, mature family, and young adults may choose to form their own family-businesses or stay at home and bring in a new generation of mates.
That family is a 0th level of trust : they all know each other personally, adults and children.
50 is the maximum number of adult-level links that the family can maintain without more than one person being members of more than organization. If organizations never grow beyond 100 before fissioning, a world of 10B people would require one such chapter for every 100 families. that is 10,000,000,000 / 100 = 100,000,000 local chapters. Those local chapters would require 1M regional aggregations, etc., for a maximum of 5 organizational levels to encompass 10B people into one organization. Few organizations need to be so large. Nevertheless, it will not take many international organizations and families belong to them for reasons of social, business or professional benefit to connect statistically-all humans to all others by 7 links of individuals who personally know other people who know … That is approximately so even now, but the relationships are ‘know’, not ‘know personally’, and convey little trust.
‘Government’ and private companies would conform to the rule to keep all organizations below 100 in number of members. Many companies now have more than 10 levels in their organizations (the local telco had 17 when I did a contract for them), so a restriction to no group larger than 100 is no problem.
So the limit of 100 in a group is not a restriction on any one organization, nor does it restrict scaling in other ways, e.g. connecting to the many families that each family will be related to as well as the local, regional and global organizations that will evolve. Assume a straight line maximum of the most social of the adults in the 50 being able to deal with a full 30 people (outside the family, to make this easy) and 30 contexts, companies, business deals, social organizations, and other families. At an average of 25 people per organization (80/20 rule says skewed small), that is 30*25 = social entities, individuals who represent groups, who can be within that most-social person’s social network. So that most-social individual can represent the interests of 100 people to an aggregate social network of 750 people. This is a social group who can trust each other because someone they trust trusts someone who vouches for them, a 2nd level trust social network.
Perhaps not so large a number you think, how can we scale to a total-trust society, the goal of our exercise? There are two ways of scaling that, the total number of adults in the family, which we have set at 50 and so much larger than any family today, and by 3rd and nth order social connections.
First, the family. 750 the maximum people for just one of 50 adults in our family, the most social. Assume that the least-social person can only deal with 1 outside group, thus 25 people, but represents the group and is trusted by them to convey trust about others. For each adult in the family, that is a The average is 30*1/2 = 15 groups handled by the average person in the 50 adults. In reality, the curve would be a normal distribution, most people handling a handful of outside organizations, so perhaps 250 for any given family. However it works, that is a much larger and more trusting network of people than exists anywhere in our modern society, even the Marine Corps or other military services.
That seems more than enough for me, but perhaps not, the world is a complex place and a big families will have many interess, so who knows. However, we can allow people in related families, with whom a family necessarily has personal and often business connections, to represent us. That is one ‘Nth order’ social connection, adding function to already-strong social connections.
Obviously, already-strong trust in a social network can be extended to another level, trusting the people trusted by those 2nd-order connections.
How large is that 2nd-order trust network, in people? If the the average organization is 25 people and the average family can handle 250 social-business connections, every family has a 1st-level trust with 250*25 =6250 people. Each of those individuals are also a part of families and organizations in relationships of trust. Thus, because we individuals in a family know one of our own who knows someone who personally knows any one of those 6250 and who who can personally vouch for them, we can trust all of them. All the information needed for those connections is easily available as we add the connections to our own private facebook-type information base, in which we each own our data.
We can conclude that the number of social and business groups and networks in trusting relationships is not a limit on scaling a society built of family-clan-business units.
10B / 6240 = 1.6M such social networks. Adding one more level of indirection means a network of 6250 * 25 = 64000 networks in a population of 10B people. Probably people’s trust won’t extend more than 3 levels of indirection, although Sexbots and other mechanisms will likely correct the situation. Also, these are conservative assumptions, the least-social person in most families will handle more than one outside relationship.
Note that many small units is optimal for evolution, our original goal.
***A friend read a book perhaps 20 years ago and recounted it to me in several different discussions. I cannot find the book, Google and other online are increasingly useless for small details the further back you look.
From my memory of his interpretation : The Catholic Church had several reasons for preventing cousin marriages, one of which was reducing the power of clans and therefore increasing the relative power of the church. Cousin and 2nd-cousin marriages were a common way of ‘keeping it in the family’, as land holdings were the source of most wealth at the time. By preventing even 2nd-cousin marriages, land was broken into smaller holdings and more opportunities existed for it to be willed to the Church in return for masses for the departed’s soul in perpetuity.
Other thoughts on the subject, which don’t fit in anywhere above :
If people never voluntarily allow themselves to group in units larger than they can personally affect the direction of sufficiently to maintain primary loyalty to that group, generally less than 100, we can maintain humanity in our organizations, if we are careful in other ways about bring up people ‘right’. If more than 150, not. Certainly there are quite powerful arguments against being part of larger groups, which must be dominated by a few individuals. Groups small enough to know each other personally and to have largely-shared interests should, generally speaking, be decision makers, as a group. Careful attention to rules of evidence and open discussion, everyone’s opinion solicited, the group leading itself, is the way to make decisions and grow individuals. Read O’Neill’s “Summerhill” for how he ran a very successful school and brought kids up to be in control of themselves, but that is just one of many equivalents. Iroquois Nation people raised children who were tough and brave, yet allowed everyone great independence in thought, e.g. whether or not individuals should go to war on behalf of the French vs the English or not at all. People nearly uniformly do well at managing themselves in groups, if they know each other well and communicate with neighbors so as to avoid conflict.
Few people have faced the problems of running a large family unit in a modern society with modern technology and communications. ‘Dysfunctional family’ no doubt has many dimensions, and training and advice about how to avoid it will be a big industry for a while. Families with the longest, most successful experience will lead, as usual.
Big advantage is that large families will make us human again by making us familiar, intimately, with human flaws. It will not so much soften our judgement as give us a better basis for them, improved metrics. Dealing with the same people throughout your life, hearing the gossip about everyone and learning all of the dimensions of flaws in your brother-in-law tends to make us more accepting of remote others.
Few times in human history have such very good conditions for the average man been used so well. Sad to see it ending.
One view of the problem we currently have with trust in institutions is that we have no way of connecting our trust in individuals to those institutions.
Entertainment is a major way of learning about everything. Entertainment need not be mindless, unless created by the mindless. Games have already improved human minds, Hollywood has no clue, its products systematically degrade people’s understanding of everything.
The proper group size fosters social evolution, has evolved to support that evolution. Dunbar’s number must foster civilization. If so, society has been wrong to abandon clans. The Catholic Church did that to increase its power, a central-planner decision, not a bottom-up evolution.
Smaller groups can be smarter groups :
Another advantage of large, enduring families is decisions are made locally by local criteria, yet can be tuned to the world. Modern communications can enhance a modern family.
Re-reading this, on what I swear will be the last pass edit, I realize you could think I was promoting some communalist philosophy. Not at all, many forms of everything is the goal, many combinations of forms. That is the goal. I think capitalist is the best general form, but probably far from the one we have. I hope there are much better versions of Constitution and Law and government in general, there are some bad aspects of this one, e.g. Wall Street and the CIA. So far, the global summary would be the more public the less well if functions, is protected, valued. Major cities now have many ‘public-private’ places to go, and they are kept clean and neat. The wild animals of the world are killed off in the best run large reservations, gone extinct in nations that depend on tourist income for a large part of their GDP.
Sell the species, if you want to preserve them! Maybe, maybe not, but this model of civilization is not working very well. It has to change for a myriad of reasons.