The most recent and refined version of the evolutionary worldview that was first presented in Evolutionís Arrow can be found in the 34 page document The Evolutionary Manifesto which is here
Chapter 5. Organising Cooperation
Cooperation enables living processes to do better in evolutionary terms. Whatever evolutionary challenges they face, organisms will do better by cooperating. And the wider the scale of the cooperation, the more effective it isthe greater the power the cooperators will have over their environment, living and non-living. When evolution discovers ways to increase the scale of cooperation, life progresses in evolutionary terms.
But cooperation does not evolve easily. Self-interest stands in the way. Wherever cooperation has evolved, this impediment has had to be overcome. If we are to demonstrate that evolution can progress by organising cooperation over wider and wider scales, it is not enough to show only that wider-scale cooperation delivers evolutionary advantages. We must also show that evolution is capable of organising cooperation to exploit the advantages. How has evolution got around the barrier to the evolution of cooperation? How has it built cooperative organisations out of self-interested components? And can evolution continue to progress in this way in the future, building cooperative organisations of even larger scale?
Until recently, when evolutionists have attempted to answer these questions, they have focused on the obvious and simple instances of cooperation that we see around us. They have concentrated on the limited cooperation we see amongst multicellular animals such as mammals, birds and insects. It is only in the last decade or so that much work has been done on the evolution of the more complex cooperation found within cells, within multicellular organisms, and within human society.
Evolutionary theorists have been very successful at explaining the limited cooperation that exists amongst some multicellular organisms. Theorists have shown that there are a number of mechanisms that will enable simple cooperation to evolve amongst these animals. And they can explain why the mechanisms have produced cooperation only amongst a relatively few species. Current evolutionary theory is good at showing why most of the animals we see around us do not cooperate much.
As might be expected, the mechanisms are not so good for explaining the evolution of the spectacular cooperation we see within cells, multicellular organisms, and human society. The mechanisms do not disclose a general method for organising cooperation amongst self-interested components that would work for any living processes in any circumstances. The mechanisms are therefore not much help in showing how to build human organisations that are more cooperative.
However, the discovery of these mechanisms has been an important step towards a comprehensive theory of how evolution can overcome the barrier to cooperation. While the mechanisms are not a comprehensive answer, they point to the direction in which a general answer might lie.
The first major breakthrough in understanding how cooperation might evolve between multicellular animals came with the publication of William Hamiltons theory of genetical kin selection in 1964. The central idea in the mechanism he described is very simple: consider a cooperator who shares its food with others who have not been able to gather enough. If the cooperator shares food only with other individuals who contain copies of its cooperator gene, the benefits of cooperation will go only to other cooperators. Only those who carry the cooperator gene will benefit. Non-cooperators will be excluded from any of the advantages of cooperation.
Provided the benefits of cooperation outweigh the costs, cooperators will end up in front, and will out-compete non-cooperators. Even if an individual cooperator ends up worse off because it pays the costs of cooperation and gets no benefits, cooperator genes will be better off overalla copy of the gene will also be carried by the individuals who get the benefit of cooperation. The full benefits of cooperation will be captured by the cooperators, the barrier to the evolution of cooperation will have been overcome, and the potential advantages of cooperation can be exploited.
Of course, the catch is that individuals cannot tell easily which other individuals carry the cooperator gene. So they are unable to direct their cooperation only towards others who carry the cooperator genes. Some of the benefits will leak to free riders and other non-cooperators, enabling them to do better than cooperators.
Hamilton saw that there was a partial way around this. If an individual carries the cooperator gene, its relatives are more likely to carry the gene. So if you are a cooperator, and you also know who your relatives are, you know individuals who are more likely to be cooperators. And you do not even have to know exactly who your relatives might be. If you are an organism that tends to live its life near where it is born, you can be sure that those around you are more likely to be your close relatives.
But the less certain you are about who else carries a copy of your cooperator gene, the less effective is kin selection at evolving cooperation. For example, if you carry a cooperator gene, there is only a 50 per cent chance that your brother or sister also has the same gene due to relatedness. So if you provide some resources to a brother or sister, you are denying your cooperator gene the value of those resources, in exchange for only a 50 per cent chance that the resources will be going to a related cooperator gene. If the cooperator genes are to end up in front, you would help a brother or sister only if the benefits of the cooperation outweigh the costs by at least two to one. When you help a brother or a sister, there is a fifty-fifty chance you will not be helping a related cooperator gene. So on the 50 per cent of occasions when the brother or sister does carry the cooperator gene, you have to get double the return. You have got to make up for the 50 per cent of times when the cost of cooperating is wasted.
To take a starker example, consider a cooperator who sacrifices its life to help its brothers and sisters. On average, it would have to save at least two of their lives to be sure that it was saving at least one related cooperator gene to replace its own.
The more distantly-related the individuals, the worse this gets. The pay off from cooperation has to be even higher for cooperation to come out in front. For example, it is worth helping a first cousin only if the benefits produced by the cooperation outweigh the costs by at least eight times. On average, there would be one related cooperator gene in eight first cousins. An individual cooperator should sacrifice itself for a first cousin only if it saves at least eight of them.
As a result, kin selection is a very poor mechanism for evolving cooperation. Because relatedness is an imperfect indicator of common genes, kin selection can evolve cooperation only where the benefits from cooperation far outweigh the costs. This leaves an immense amount of beneficial cooperation that cannot be evolved by kin selection. If a mechanism is to be able to fully explore the benefits of cooperation, it must be able to establish any cooperation in which the benefits outweigh the costs. Kin selection falls far short of this. Without these severe limitations, kin selection would have been much better at organising cooperation amongst the animals around us. The potential benefits of cooperation would have driven the evolution of a complex division of labour within each population of animals, similar to what we see within our societies, our bodies, and our cells.
But as a theory, kin selection has been very successful. The ineffectiveness of the kin selection mechanism is consistent with the sparse and limited cooperation found amongst multicellular animals. And where there is cooperation, the way it is organised is consistent with kin selection. Much of the cooperative organisation we find amongst the animals around us is between relatives.
The most common is where parents provide food and protection to their young. But parents are not the only individuals that help to bring up young in some species. In a number of bird species, eggs are incubated and young are fed by birds that are often not their parents. An example that has been studied at length is the Mexican Jay that lives in flocks of five to fifteen in pine and oak woodlands. Each flock defends a territory as a group, actively excluding other flocks. At each nest most or all of the members of the flock help bring food to the young. Birds that are not parents typically bring about half of the food.
Why do these birds cooperate in this way? Surely they would be better off in evolutionary terms by eating the food themselves, and putting their time and energy into starting up their own families? Kin selection provides the answer to this. The individuals within each flock are highly related, and the cooperators are in fact helping others that have a higher likelihood of carrying their cooperator genes.
Similar patterns are found in other birds that breed communally, including Australias superb blue wrens, and white winged choughs. High relatedness is also a universal within the cooperative hunting groups of African lions, wild dogs, and spotted hyenas, and within the groups of savannah baboons and many species of monkeys that collectively defend against predators. And the highly cooperative colonies of ants, termites, and bees are invariably formed of individuals that are closely related.
The second mechanism that enables cooperation to evolve between animals is reciprocal altruism. Again, the basic idea is very simple: an individual will not be disadvantaged if it helps another provided the other helps it in return. The individuals initial costs of cooperating will be recouped when the other returns the favour. If the benefits of cooperation outweigh the costs, both individuals can end up in front. The cooperators will therefore out-compete individuals that dont reciprocally help each other in this way.
For example, when an individual is successful in collecting food, it might share it with others who have not been so lucky. On occasions when others are more successful, they can return the previous favour by sharing their food. Over time, all will be better off than if they did not share.
Exactly this sort of cooperation has been found in South American vampire bats. Each night the bats leave the hollow trees where they spend the day to search for large animals. If a bat finds an animal, it will attempt to feed on the animals blood. But there is no guarantee that an individual bat will successfully find and feed off an animal every night. Young inexperienced bats are unsuccessful about one night in three. If they have a few bad nights in a row, they can be in danger of starving. This danger will be avoided if the bats that hole up together share blood. Those that have been successful regurgitate some blood for others that have missed out. Because all bats risk having nights on which they fail to get blood themselves, they can all benefit from this cooperation.
But closer study of the bats showed that any particular individual would not regurgitate blood for all of the other individuals. It would give blood to some, but not to others. And those who got the blood were not necessarily close relations, as would be expected if kin selection were operating. It turned out that those who were given blood were bats who had given the individual blood in the past. And those who were refused blood were those who had previously refused the individual blood. The bats knew each other as individuals, knew who were likely to return favours and who were not, and chose to give blood only to those who were likely to reciprocate.
A closer understanding of reciprocal altruism and of its limitations shows that the bats have to be this choosey. In fact, reciprocal altruism cannot be established in a group of organisms unless they are capable of refusing to cooperate with individuals that fail to return favours. If an individual continues to do favours for others that do not return the favour, the non-reciprocators will win out. An individual that cheats by receiving favours but not repaying them will gain all the benefits of cooperation without paying any of the costs. And if a cheat is within a large population of cooperators, it will do extremely well, collecting favours and never returning them. In these circumstances, cheats will prosper, increasing in numbers in the population until all cooperators die out. In such a group, the most competitive strategy will be to accept favours but not reciprocate. Reciprocal altruism will not survive, even if its benefits far outweigh its costs.
However, all this changes if the individuals live together for a long time, repeatedly have the opportunity to exchange favours, and have the capacity to remember which individuals return favours, and which do not. If these conditions are met, cheats will not be able to continually take benefits from the cooperators, because they will be recognised as cheats and excluded from further cooperation. Once this happens, they will not gain any of the benefits of cooperation. In contrast, cooperators will end up in front because they will limit their exchange of favours to those who have shown they will return the favours. If they cooperate in this way whenever the benefits outweigh the costs, cooperators will out-compete cheats.
But it will rarely be this simple. It might pay cheats to behave like cooperators for a while to extract favours, then to quit while they are in front. Or when they have gained a reputation for cheating in one group, they may move on to another group where they are unknown. To avoid being cheated, cooperators will have to get very good at detecting cheats.
Furthermore, the whole system of reciprocity works only if there is a fair balance between the favours that are exchanged. Individuals whose favours cost more than the individuals receive in return will be out-competed. But most animals have little ability to judge whether reciprocal favours are similar in value. And having to do so further complicates the critically important job of telling the difference between cheats and cooperators.
All these challenges become almost impossible when acts of cooperation benefit many others, and impact differently on each of them. If a cooperative act benefits a number of others in the group, rather than a specific individual, it is much harder to work out who owes who what, and how much. But unless individuals can keep track of all this, cheats will flourish
The great advantage of reciprocal altruism over kin selection is that it can evolve cooperation between individuals who are not related. However, the conditions under which it works are so limited that it has not been very significant in establishing cooperation between animals. Not many animals live together for long periods, repeatedly have the opportunity to cooperate, are able to recognise each other as individuals, can remember how each individual behaved each time it had the opportunity to cooperate, and can keep track of favours owed and favours earned. Cases of reciprocal altruism have been found only in a few species of animals such as vampire bats, dolphins, and elephants, and in some species of monkeys and apes.
Of course, the species that meets these conditions to the highest degree is us. And the theory of reciprocal altruism fits very well with the way we relate and cooperate with people that we meet repeatedly. It works well for explaining key aspects of the cooperative relationships between people in tribal groups, closely knit communities, and in friendship networks and other stable social groupings.
The theory predicts that successful reciprocal altruists will have two great concerns in their relationships with others. First, they will not want to be cheated by others, so they will want to know who can be trusted to return favours. Second, they will want others to choose them as cooperative partners, so that they can share in the benefits of cooperation. They will want others to see them as trustworthy cooperators. They will not want to be seen as a cheat that should be excluded from beneficial cooperation.
Consistent with this, we humans have many behaviours that seem to serve the function of protecting us against cheating. For example, we tend to have an extraordinary interest in judging other peoples trustworthiness and honesty. A major part of gossip and other social communication involves swapping information that helps us to make and update these judgements. We are also reluctant to risk cooperating with strangers when we have no knowledge of their reputation and character. We are even more reluctant when we know the stranger does not depend on us for future cooperative opportunities. In these circumstances, they will lose little if they destroy their reputation with us.
We are also very concerned to protect our own reputation for trustworthiness and honesty, particularly amongst those that we spend a lot of time with. We often react instantly and emotionally to any attack on these aspects of our character. Most of us are also concerned to ensure that we are seen to treat our friends fairly whenever there are things to be shared, and that we repay our obligations to people we know. And we do not want to enter into obligations with friends that we might not be able to repay.
But even amongst humans, the capacity of reciprocal altruism to establish cooperation is severely limited, even where cooperation produces substantial benefits. As we have seen, reciprocal altruism will work only where cheats can be successfully excluded from future cooperation. This will not be the case if cheats can easily find other individuals to deal with who know nothing of their previous actions. It is only where individuals have no option but to continue living and interacting together for long periods that cheating will not pay. If an individual earns a reputation for cheating in these circumstances, it will not find other cooperators to deal with.
This condition is not generally met in modern large-scale economic markets. Individuals can readily leave a bad reputation behind them, and find new individuals to deal with who are unaware of their previous actions. Reciprocal altruism is therefore not the mechanism that has produced the large-scale markets that organise economic exchanges across human societies.
The other great limitation of reciprocal altruism is that it is only effective where the benefits of cooperation can be neatly parcelled up, precisely valued, and swapped between individuals. Where the benefits of a cooperative act impact on many individuals, reciprocal altruism cannot operate easily. This is a significant limitation. In the complex, differentiated cooperative organisation that are cells, organisms, and human societies, most cooperative acts impact on many others in the organisation. It would be impossible to separate out the effects and keep track of them as favours that have to be returned. And if favours are not returned with acts of similar value, reciprocal altruism fails.
So genetical kin selection and reciprocal altruism are not the mechanisms that have enabled evolution to explore the potential benefits of cooperation within cells, multicellular organisms, and human society. And they are not the mechanisms that will change humans from partial cooperators into full cooperators. They can achieve only limited cooperation under restricted conditions. If kin selection and reciprocal altruism were the only mechanisms that evolution could use to organise cooperation, evolution would not be progressive. It would not be able to exploit the benefits of cooperation by progressively building cooperative organisations of wider and wider scale.
But kin selection and reciprocal altruism do hint at a more general way of building cooperative organisation out of self-interested components. Where they are effective, both mechanisms establish cooperation in the same way. They ensure that cooperators benefit from their cooperation. The mechanisms work to the extent that cooperators are able to capture the benefits created by their cooperation. Kin selection achieves this to the extent that the individual helped by the cooperator also carries the cooperator gene. If it does, the cooperator gene captures the benefits of the cooperation. Particular cooperators may not capture the benefits of their cooperation, but cooperators as a class will. If the cooperation is worthwhile, its benefits will outweigh the costs, and the cooperator genes end up in front.
Reciprocal altruism works to the extent that the recipients of cooperation return the favours. If they do, cooperators capture the benefits of their cooperation. They benefit from all of their cooperative acts, because each will earn a return favour. If the cooperation is worthwhile, cooperators again end up in front.
Where the mechanisms fail, both do so for the same reason. They fail to the extent that cooperators do not capture the benefits of their cooperation. Kin selection is undermined when the individual helped by the cooperation does not carry the cooperator gene. When this occurs, the cooperator gene will not capture the benefits of cooperation, only the costs, and will be out-competed. Reciprocal altruism is undermined when a recipient of cooperation cheats by failing to return the favour. When this occurs, a cooperator will not capture the benefits of its cooperative act, only the costs, and will end up worse off than the cheat.
This analysis of kin selection and reciprocal altruism points to what a mechanism must do if it is to be capable of establishing cooperation whenever cooperation is worthwhile. It must ensure that cooperators capture the benefits of their cooperative actions. Free riders, cheats and thieves who contribute nothing to the cooperation must capture none of the benefits created by cooperation.
To overcome the barrier to cooperation fully, a mechanism must also ensure that those who harm others capture the effects of any harm they cause. Harming others must be harmful to those who cause it. And to organise cooperation across generations, a mechanism must also ensure that all the effects of actions on others must be captured, no matter how distant in the future the effects may be. This is necessary to ensure that cooperation that benefits future generations will also be profitable to those who produce it.
A mechanism that ensures individuals capture all the effects of their actions would completely overcome the barrier to the evolution of cooperation. If individuals capture the effects of their actions on others, self-interest would no longer stop an individual from helping another. Helping the other would be as profitable to the individual as helping itself. As a result, the individual would treat the other as it would treat itself. And it would do so without giving up its self-interest in any way. It would be in the interests of the individual to treat the other as self. Anything it could do to help the other would help itself. Anything it could do to harm the other would harm itself. The individual would benefit as much from discovering new ways to help others as it would from finding new ways to help itself directly. All this holds true whether it is individual cooperators or cooperators as a class who capture the effects on others of their actions.
At this point it is useful to look briefly at the difference such a mechanism would make for the examples of failed cooperation that we looked at in the previous Chapter. Taking each example in turn, it would ensure that: corporations that pollute would capture the harm they visited on others, causing them to be out-competed by corporations that introduced less efficient but less harmful manufacturing processes; businesses that trained employees would capture the benefits of producing skilled employees for their industry; only employees who contributed to the bargaining process with their employer would obtain any improvements in wages and conditions that were won through the bargaining; cells that specialised to perform useful functions for the group of cells would out-compete cells that contributed less to the group; mitochondria that harmed their host would also harm their own interests, and would be competitively disadvantaged against other mitochondria; and protein enzymes that contributed more to the autocatalytic set as a whole would in turn have their production catalysed by the set.
Wherever complex cooperation has been able to evolve, it is because cooperators have been able to capture the effects on others of their actions. As a result, self-interest has driven them to treat the other as self. A strong example that may not be immediately obvious is an economic market. The market will reward an individual who develops a new product that benefits others by satisfying their needs better than existing products. Selling the product enables the individual to capture the beneficial effects that the product has on others. In the limited areas where economic markets work effectively, individuals benefit in this way from actions that benefit others. Where the market enables individuals to capture the benefits of their effects on others, the market aligns the interests of individuals with the interests of others.
In contrast, where the barrier to the evolution of cooperation has not yet been overcome, cooperators do not capture the beneficial effects on others of their actions. For example, in our economic system, a person will not capture the benefits of providing food to people who are starving in an African famine. It would be far more profitable to invent an improved mousetrap. Or a new weapon of mass destruction. But when the barriers to cooperation in human society are overcome, and when all individuals capture in full their beneficial effects on others, the provision of food to the starving will be a lucrative way to make a living.
So a mechanism that ensures that cooperators capture the effects of their actions on others would allow cooperation to be established wherever cooperation is worthwhile. Organisms that pursue only their own interests would act cooperatively. The mechanism would establish cooperation between organisms whose basic nature is greedy and selfish. It could produce cooperation between lawyers and real estate salesmen. The mechanism would enable the potential benefits of cooperation to be fully explored and exploited amongst any living processes. It would enable evolution to progressively exploit the advantages of cooperation, organising cooperative organisations of wider and wider scale.
This is the ideal. But is there a mechanism that can achieve this? Is there a general form of organisation in which individuals capture the effects on others of their actions? Or is the ideal unattainable?
One general way in which cooperators could capture the effects of their actions on others is if another individual takes action to ensure this happens. The other individual could do this by providing cooperators with benefits that reflect the useful effects of the cooperators on others, and by punishing those who harm others.
To see how this might work, imagine a population of organisms that are formed into a number of unorganised groups. Within each group, the barrier to cooperation would apply. Individuals who use resources to help others without benefit to themselves would be less competitive than non-cooperators. A group that contains cooperators might do better than other groups for a while, but this would not help the cooperators. Within such a group, the cooperators would still be less competitive, and they would still die out. Specialisation and division of labour could not evolve within a group no matter how beneficial this would be to the group as a whole.
What would be the outcome if a group contained an individual who acted to ensure that all others in the group captured the effects of their actions on others? The individual would provide resources and services to cooperators to reflect the benefits of their cooperation. And the individual would punish free riders, cheats and thieves who would otherwise take the benefits of cooperation without contributing anything to others in the group.
Within such a group, cooperators who produced net benefits for others would end up in front. They would out-compete free riders, cheats and lesser cooperators. Whenever a more effective form of cooperation was discovered, it would flourish within the group. Specialisation and division of labour would be established wherever it was beneficial.
But there is a fundamental flaw in this.
An individual who was a normal and typical member of the group would be disadvantaged if it carried out this task of rewarding cooperators and punishing cheats. An individual who used its time and resources in this way would be out-competed by those who instead used their resources for their own benefit. Although the individual would be carrying out a function that would greatly benefit the group as a whole, it would disadvantage itself in the process. It would run into exactly the same barrier as any other individual who does things for the benefit of the group. And there would be no other individual to reward it so that it captured the beneficial effects of its actions on others.
The problems do not end there. As a typical member of the group, the individual would also be unable to gather all the resources needed to reward cooperators for their contributions. When it attempted to collect resources for this purpose the individual would be in competition with all other members of the group. It would not be able to gather more than others.
And the individual would not have the capacity to control and punish cheats and free riders in the group. It would not be in the evolutionary interests of cheats to be controlled and punished. They would have a very strong incentive to escape control, and to exploit the growing cooperation in the group. A typical member of the group would not be able to control cheats, free riders and thieves, particularly if they ganged-up to resist control.
A typical member of the group cannot carry out the function of organising the group so that members capture their effects on others. How could this difficulty be overcome? What sort of entity could successfully organise the group in this way, and enable the immense benefits of cooperation to be exploited?
If the job of organising a group of humans in this way was on offer, no one would be likely to accept the job unless they also had the power and authority needed to do it. How they got the power would not be of critical importance. The job would probably be easier if they got the authority by the agreement of the members of the group. But the job could also be done even if the authority was achieved by coercion, without agreement.
This points to a general way in which these problems could be overcome, enabling the group to be organised to promote cooperation. The entity or team of entities who organise the group must have control over the group. It must have the power to control and manage the individuals who make up the group. This controller, who I shall call the manager, must have the capacity to take from the group whatever resources it needs to carry out its tasks. And it must be able to punish others without risk to itself.
Importantly, a manager who has this power would not be in competition as an equal with the other members of the group. Other individuals in the group would not have the capacity to out-compete the manager because the manager has control over them and their resources.
With this power, the manager has the ability to take whatever resources it needs to give to cooperators to ensure that they capture the benefits of their effects on others. The manager will be able to redistribute resources within the group to ensure that cooperation pays. And the manager will have the power to control and punish cheats and free riders throughout the group, even when they gang up against the manager. It can overcome all the problems that stop a normal member of the group from organising the group to support cooperation.
I will now develop a number of broad examples that will illustrate more concretely how this form of organisation is able to support cooperation. I will begin with an example in which cooperation is sustained within human organisation by the actions of a manager. Then I will look at how the management of molecular processes and mitochondria has been able to produce the complex cooperation found within cells.
About ten thousand years ago, humans began to domesticate plants and animals and form the first agricultural communities. Potentially, the communities could benefit greatly from specialisation, division of labour, and other forms of cooperative activity. For example, it would be more efficient if some individuals specialised in making farm implements or weapons, if others trained animals for use in transport and farming, and if others specialised in defending the community. And the welfare of the community would be greatly enhanced if individuals cooperated together to construct irrigation systems for farming, or to fight off enemies.
But the establishment of these and other forms of cooperation had to overcome the barrier to the evolution of cooperation. If the barrier was not overcome, cheating and stealing would make it difficult for specialists to get a proper return for what they produced. And free riders could avoid making contributions to community activities such as defence and irrigation, but could still share in the benefits produced by community projects.
As a result, cooperators would not capture all the benefits of their cooperation. And it would not be in the interests of any individuals in the community to initiate many forms of cooperation that could produce significant benefits to the community. The community could not exploit the immense benefits of cooperation. However, this could be completely turned around if the community was managed by a powerful ruler. The ruler could collect taxes off all members of the community, and redistribute these to support useful cooperation. The ruler could pay for specialist services and fund other activities that the ruler judged were of benefit to the community as a whole. And the ruler could punish thieves and cheats.
A particularly wise ruler might discover a better alternative than using tax to support specialists such as toolmakers. The ruler might find that specialists would be able to make a living by exchanging their products for food and other services, provided they were protected from theft and from cheats who do not reciprocate in exchanges. The ruler could provide this protection. Cheating and theft could be punished, exchange agreements could be enforced, and disputes could be heard by the ruler and settled fairly. With these protections, a system of market exchange would emerge.
Once a market system existed, the ruler would not have to decide the types of goods and services that were to be produced by specialists, nor their quantities and prices. Instead, these issues would be decided directly by producers and consumers through their interactions. The advantage is that producers and consumers would be in a much better position than the ruler to make these decisions. The consumers would be far better able to assess the value to them of what the specialists produced. And the specialists would be far better placed to decide whether it was worthwhile for them to produce particular goods and services, given the costs of doing so and the price they could get.
However, such an exchange system could not work effectively without a ruler who could manage cheating and stealing. Without this, all that could emerge is a system of reciprocal altruism, with all the flaws and limitations of such a system. The essential role of the manager is to patch up the limitations in the mechanism of reciprocal altruism so that cooperators are able to capture the benefits of their acts. And the individuals involved in the exchanges are unable to do this patching. Only managers who have power and control over individuals can do it. As we shall see in detail later, a manager is as essential to a market economy as it is to a centrally planned economy.
In these ways, a manager can produce a highly cooperative organisation of humans, even if the individual members of the organisation are purely self-interested.
The barrier to the evolution of cooperation in autocatalytic sets can also be overcome by management. As we have seen, a protein enzyme that catalyses reactions that greatly improve the efficiency of an autocatalytic set would not necessarily be catalysed in return by other members of the set. In fact, such a cooperative protein could be out-competed by a similar protein that did not help the set as a whole. This would be the case if the non-cooperator protein used its time and resources to catalyse other proteins that in turn catalysed the production of the non-cooperator.
The management of a set by one or more RNA molecules could change this. RNA molecules could take control of an autocatalytic set of proteins because of their larger size, greater stability, and their superior ability to catalyse a diversity of proteins. The RNA could use its superior catalytic ability to cause the formation of proteins that otherwise would not exist in the set. It could use this ability to produce cooperative proteins that are beneficial to the set as a whole. A specialist protein could be supported by the RNA to the extent needed to reflect the beneficial effects of the protein on others in the set.
The RNA could also produce proteins that further enhanced the ability of the RNA to catalyse other proteins, improving its ability to control the set. And it could control cheats and free riders that take resources from the set and are catalysed by the set without contributing anything in return. The RNA could do this by catalysing reactions that take resources and catalysis away from the cheats and free riders. The RNA could have as much control over what happens in an autocatalytic set as a human ruler who controls an agricultural community.
The RNA manager could use its control over the set to ensure that each type of protein captures the effects of its actions on the set as a whole. The manager could ensure that its support for the production of a particular type of protein reflects the contribution of the protein to the organisation.
Management was also essential for the evolution of the close cooperative relationship that developed between cells and the bacterial ancestors of mitochondria that began to live inside the cells. As we saw in the previous Chapter, bacteria that put all their time and effort into helping the cell in which they lived were not likely to have been successful. They would have had to compete with other bacteria who instead plundered the resources of the cell, used them to breed up, then left the cell to find other cells to plunder. These non-cooperative bacteria would free ride on any help given to the cell by cooperative bacteria.
But cells could overcome this by controlling the bacteria inside them. A critical step was for cells to prevent the bacteria from leaving them. Once this option was cut off, the fate of the bacteria was much more closely linked to the fate of the cell. Free riders that were a drain on the cell would now capture their harmful effects because their future was totally dependent on the success of the cell and its descendants. Free riders would lose out. Cooperators could also now capture the benefits of their cooperative effects on the cell. The better the cell and its descendants went, the better the cooperators and their descendants went.
Multicellular organisms had to introduce additional controls to prevent mitochondria competing within them and undermining cooperation. A key control was to prevent the transmission of mitochondria from males to their offspring. As we saw earlier, mitochondria are transmitted only through the egg, not through sperm. So all the mitochondria in the cells of a multicellular organism such as yourself are descendants of mitochondria that came from the mother. And the mothers mitochondria are all descendants of those that came from her mother, and so on. There is no mixing of mitochondria from different individuals or different lineages.
This is important for ensuring that cooperative mitochondria win out: if mitochondria are prevented from mixing with mitochondria from other lineages, they do not have the opportunity to free ride off them. If males did transmit mitochondria, and mitochondria from different lineages did mix, free riders could flourish and undermine cooperation. They could free ride off any cooperative mitochondria they were mixed with, out-competing them in the process. Then they could move on when the host reproduced, to be mixed with a fresh lot of cooperators, and so on. The result would undermine cooperation amongst mitochondria, and cooperation between mitochondria and the cells that house them.
There is clear evidence that different lineages of mitochondria would compete and undermine cooperation in this way if they were mixed: often, not all mitochondria are excluded from some sperm cells. A few include a small number of mitochondria. When such a sperm fuses with an egg cell, the mitochondria and other organelles in the sperm are usually destroyed immediately by their more numerous counterparts within the egg.
But locking mitochondria into separate lineages so that they eventually capture their effects on their hosts is not enough. Mitochondria within an individual host will still compete amongst each other to get into the egg cells that produce the hosts offspring. And, as ever, mitochondria that put all their time and energy into helping the hosts cells will be out-competed. Mitochondria that instead put their time and energy solely into getting into egg cells will win out.
A host organism will do better in evolutionary terms if it is able to manage this competition by supporting cooperators and preventing non-cooperators from gaining any competitive advantage. The host can use its power over its mitochondria to ensure that cooperators capture the benefits of their effects on the host, and that non-cooperators capture the harm they cause. The relationships that now exist between mitochondria and the cells they inhabit seem harmonious, consensual and mutual. But a closer examination shows that this has been achieved only by the strict control of mitochondria by the cells and organisms that house them. They can use controls to support cooperators, suppress cheats and free riders, reduce the possibility that non-cooperators will be able to get any advantage over cooperators, and even to restrict the likelihood that non-cooperators will arise by mutation.
Specific examples of how mitochondria are controlled by the cell that houses them include: the rate at which mitochondria release energy for the cell is controlled not by the mitochondria, but by factors outside them in the cell; proteins produced by mitochondria cannot function independentlythey work only in the presence of proteins produced by the cell; a large part of the genetic material that was contained in the bacterial ancestors of mitochondria no longer remains in the mitochondriait is now contained in and controlled by the cell nucleus, leaving mitochondria with little control over their own functions and, more importantly, with little ability to change them; and finally, the genetic material that controls the operation of mitochondrial genes and their replication is no longer in the mitochondriathe genes of the cell control these critically important functions.
Cells have had to strictly control and manage mitochondria to ensure that the only way they can pursue their own interests is by helping the cell. To make mitochondria manageable, their independence has been greatly restricted, and their capacity to independently innovate and evolve has been suppressed by the removal of much of their genetic material. So that they serve the interests of their hosts, mitochondria have been enslaved and lobotomised.
These three examples illustrate how a manager can control a group of individuals in ways that enable the potential benefits of cooperation to be explored. Left to themselves, a group of individuals that interact as equals is unable to establish complex cooperative organisation. It will not be in the interests of any member of the group to initiate this cooperation, no matter how great the benefits that it can provide. This is because individuals will not capture the benefits of cooperation that they initiate, particularly in the face of cheating, free riding and theft.
This all changes when an extra layer of organisation is added to the group. This extra layer is the manager, made up of one or more entities who are able to control the group. The addition of this extra layer converts the original organisation of equals into what can be called a vertical organisation. Extending this terminology, the original organisation that had only one layer of organisation can be referred to as a horizontal organisation.
The manager can use its power to organise the horizontal group in whatever way is necessary to ensure that cooperators, free riders and cheats capture the effects of their actions on others in the organisation. To the extent that the manager is successful in doing this, the interests of all individuals will be aligned with the interests of the group, and individuals will treat others as they treat themselves. Self-interest will be the same as group interest, and the vertical organisation will have successfully created cooperative organisation out of self-interested components.
The manager of an ideal vertical organisation will be able to support and sustain whatever cooperative arrangements are best for the group. By determining which individuals succeed in the group and which dont, an ideal manager will be able to construct whatever organisation is needed for evolutionary success. The ideal manager will be able to make it in the evolutionary interests of members of the group to undertake whatever forms of cooperation are best. There will be no form of cooperative organisation that the manager cannot construct. Management will be able to produce any pattern of specialisation and any form of division of labour amongst the members of the group. The barrier to the evolution of cooperation will be completely overcome.
It is worth emphasising that the manager must have the capacity to control the group if the manager is to be able to organise complex cooperative organisation. If the manager is to stop cheats, free riders and thieves, it must have power over them. It must be able to make them do things that would not otherwise be in their interests, and they must not be able to avoid this control by influencing the manager in return. If they can, they can escape control, and continue to undermine cooperation. This is what happens when criminals use bribery to influence police, and when friendships with staff undermine the ability of an executive to manage staff in a corporation.
Individuals that are equals within a horizontal organisation obviously cannot impose the level of control needed to stop cheats, free riders and thieves. Individuals that band together to act as a group can achieve some control over cheats. But when they do this they are no longer acting as individuals. They are forming a manager and a new level of organisation that can control individuals.
On closer examination, we find that even the limited forms of cooperation found between multicellular organisms cannot arise unless cheats and free riders are restrained somehow. If cheats and free riders are free to take advantage of cooperators, cooperation will be undermined. In the situations where cooperation occurs amongst multicellular organisms, we invariably find that there is something about the structure of the situation that restrains cheating and free riding. For example, in vampire bats cooperation would fail if cheats and free riders could move from hollow tree to hollow tree to continually exploit fresh cooperators who are unaware of their reputation. They are prevented from doing this because the bats within each hollow tree exclude other bats from using their tree. Bats need to defend their resting-places because good ones are in short supply, and other bats will try to take them over.
Similarly, kin selection would be undermined in birds that breed communally if cheats and free riders could easily join a group and take advantage of the cooperation. But, communal birds usually defend territories to obtain exclusive access to food and other resources. So cheats and free riders are prevented from easily moving between groups, and closely-related birds will remain together. And finally, computer simulations show that reciprocal altruism can evolve easily amongst colonial animals that live fixed to rocks. With no mobility, cheats and free riders are prevented from moving freely from cooperator to cooperator, exploiting them as they go.
In all these cases, the factors that control cheats and free riders have not evolved because they help cooperation get established. They have evolved for other reasons, and the restrictions they place on the opportunities for cheating and free riding are chance side effects. So the restrictions are not tuned by evolution to be as effective as possible at promoting cooperation, and they will not adapt to stay effective when things change.
What vertical organisation does is take the control of cheats out of the hands of restrictions that have evolved for other reasons. As we shall see in the next Chapter, the controls established by managers are tuned by evolution specifically to promote cooperation. Controls that are best at promoting cooperation will win evolutionary struggles. And evolution will adapt the controls to find better ways to support cooperation and to prevent cheats and free riders from escaping control as they evolve.
For all these reasons, only vertical organisation has the capacity to support the extraordinary level of cooperation we find within cells, multicellular organisms, and human society.
But all we have shown to this point is that management has the potential to promote cooperation amongst others. Will it take up this potential? Will evolution establish managers that use their power to organise cooperation in the way I have described? Or will evolution favour managers that use their powers to plunder resources and services from other organisms, trampling over their interests in the process? Why would they use their power to promote cooperation between organisms? These are important issues for the case I am building in favour of progressive evolution. If it is not in the evolutionary interests of managers to organise cooperation, the barrier to cooperation cannot be overcome. And unless evolution is able to find comprehensive ways to overcome the barrier, it will not be able to progressively build cooperative organisations of wider and wider scales.
To a person at the turn of the 20th century, the answer to these questions might seem obvious. There is plenty of evidence from recent human history to show that if you give a person power over others, the power is likely to be abused. Power corrupts, and absolute power corrupts absolutely. In the 20th century individuals such as Stalin and Hitler who have had the most power seem to have been responsible for the most human deaths and misery.
Time and time again we have seen rulers and governments sacrifice the interests and often the lives of their people in order to stay in power. Many governments have been guilty of persecuting minorities on the basis of their religious beliefs, ideology or race. White South Africans and Australians are notorious for their exploitation and abuse of the indigenous peoples of their countries. And those who have the wealth to influence governments have never been shy about getting governments to use their immense power to favour them over others.
And there is abuse of power all the way down, at all scales. An extraordinary number of catholic priests have used their position to sexually abuse children, and, for most of the 20th century, the hierarchy of their church used its power to hide these practices and to protect the guilty. Even within families, individuals with greater power will often abuse others physically, sexually, emotionally or economically. And anyone who has worked in a large corporation knows that executives use their power over employees for many purposes that have little to do with the efficiency of the organisation.
These abuses have obscured the fact that the extraordinary level of cooperative organisation found within cells, organisms and human society has been achieved only through vertical organisation. But does this level of abuse mean that the ability of vertical organisation to establish cooperation is fundamentally limited? Is this abuse the price we have to pay for the benefits of cooperation that can be produced only by vertical organisation? Or is it simply that current human organisation and its obvious flaws should be seen as part of an evolutionary progression in which the flaws will eventually be overcome? Is this just an evolutionary stage we are going through? And if so, how will the flaws be overcome?
These are questions I will begin to answer in the next Chapter. We will see that vertical organisations of molecular processes and organisations of cells both passed through a phase in their evolution in which managers exploited the organisation, rather than promoted cooperation. But these phases ended when organisations emerged that were controlled by management that has discovered how to support cooperation. Management that used its power to organise cooperation proved superior in strictly evolutionary terms to management that used its power to exploit the organisation. We will see that competition between organisations aligns the evolutionary interests of the manager with the organisation. Managers that promote cooperation do better than those that do not, because the organisations they manage are more competitive.
But competition between human societies is no longer strong enough or incessant enough to align the interests of human rulers and governments with those of their citizens. When human groups were small and numerous, competition was strong and continual. But as human societies grew larger and fewer, competition between them has decreased. We will see that in the absence of strong and continual competition, the key to the evolution of cooperative human societies is the discovery of new organisational structures. These must align the interests of the ruler or government with those of the members of the society. When this is achieved, a government driven purely by self-interest will manage in the interests of the society. Abuse and exploitation will end, and the benefits of cooperation can be fully exploited. In the later Chapters of this book we will see in detail how human societies can be reorganised to align the interests of their government (management) with the interests of the members of the society. This will unleash the immense advantages of cooperation for the benefit of all.
. For the first attempts to build a comprehensive theory of the evolution of cooperation across all these levels, see Maynard Smith, J., and E. Szathmary (1995) The Major Transitions in Evolution. Oxford: W. H. Freeman; Stewart, J. E. (1995) Metaevolution. Journal of Social and Evolutionary Systems 18: 113-147; and Stewart, J. E. (1997) Evolutionary Progress. Journal of Social and Evolutionary Systems 20: 335-362.
. Hamilton, W. (1964) The Genetical Evolution of Social Behaviour. Journal of Theoretical Biology. 7: 1-52.
. For example, see Chapter 9 of Brown, J. L. (1975) The Evolution of Behaviour. New York: W. W. Norton.
. Brown, J. L. (1970) Cooperative breeding and altruistic behaviour in the Mexican jay. Animal Behaviour 18: 366-378.
. Heinsohn, R. G. (1992) Cooperative enhancement of reproductive success in white-winged choughs. Evolutionary Ecology 6: 97-114.
. See Corning, P. (1998) The Cooperative Gene: On the Role of Synergy in Evolution. Evolutionary Theory 11: 183-207.
. For example, see Wilson, E. O. (1975) Sociobiology: The New Synthesis. Cambridge, MA: Harvard University Press.
. Trivers, R. (1972) The Evolution of Reciprocal Altruism. Quarterly Review of Biology 46: 35-57; and Axelrod, R. and D. Dion (1989) The further evolution of cooperation. Science. 232: 1385-1390.
. Wilkinson, G. S. (1984) Reciprocal food sharing in the vampire bat. Nature 308: 181-184.
. Axelrod, R. (1985) The Evolution of Cooperation. New York: Basic Books.
. Martinez-Coll, J. C. and Hirshleifer, J. (1991) The limits of reciprocity. Rationality and Society 3: 35-64.
. Dugatkin, L. (1999) Cheating Monkeys and Citizen Bees: The Nature of Cooperation in Animals and Humans. New York: Free Press.
. See Dunbar, R. (1996) Grooming, Gossip and the Evolution of Language. London: Faber and Faber.
. Stewart, J. E. (1997) Evolutionary transitions and artificial life. Artificial Life 3: 101-120.
. Stewart: Metaevolution. op. cit.; and Stewart: Evolutionary transitions and artificial life. op. cit.
. Stewart: Metaevolution. op. cit.; and Stewart: Evolutionary transitions and artificial life. op. cit.
. This impediment to the evolution of cooperation is termed the second order problem by Axelrod, R. (1986) The evolution of norms. American Political Science Review 80: 1095-1111.
. Stewart: Metaevolution. op. cit.
. Stewart: Metaevolution. op. cit.; Stewart: Evolutionary transitions and artificial life. op. cit.; and McGuire, M. C. and M. Olson (1996) The Economics of Autocracy and Majority rule: The Invisible Hand and the Use of Force. Journal of Economic Literature 34: 72-96.
. Hodgson, G. (1988) Economics and Institutions. Cambridge: Polity Press.
. Frank, S. A. (1997) Models of symbiosis American Naturalist 150: S80-S99.
. Eberhard, W. G. (1980) Evolutionary consequences of intracellular organelle competition. Quarterly Review of Biology 55: 231-249.
. Blackstone, N. W. (1995) A units-of-evolution perspective on the endosymbiont theory of the origin of the mitochondrion. Evolution 49: 785-796.
. Stewart: Metaevolution. op. cit.
. Wilkinson: Reciprocal food sharing in the vampire bat. op. cit.
. Brown, J. L. (1987) Helping and communal breeding in birds: ecology and evolution. Princeton: Princeton University Press.
. Nowak, M. A., May, R. M. and K. Sigmund (1995) The Arithmetics of Mutual Help. Scientific American June 1995: 50-55.