How Does Cooperation Evolve from Competition?
The following essay was published 11/3/2008 to state a larger purpose for OpenBEDM, which was then called the “Pygar Project”.
Motivation in our society comes from the competitive instinct: get ahead, get rich, get recognition. However, success always depends on effective cooperation. The rich, the famous, the well-known belong to networks of cooperating individuals. The top competitor is never a lone individual but always a member of a clique or network.
To understand how society and the economy actually work, we need to explain how success is driven by competition but accomplished through cooperation. I believe this will be a matter of national survival. Our society must compete against other societies with different values and political systems. If our networks fail us by undercutting the cooperation within our group, our values and uniqueness will by lost. My thesis is that the Internet is failing us because it is impeding important modes of cooperation. A new Internet method is essential.
True, the Internet offers endless on-line information resources. But where vital interests are concerned – national security, medical situations and treatments, financial dealings – the data are locked up in secure off-line databases. We are plugging into the web in great numbers. So are our doctors, our government and our security forces. But, too much of the critical information is locked up in secure databases and unusable. Data security is necessary; but it prevents essential cooperation between problem-solvers who want to increase our society’s health, security, and well-being.
Forward looking technology experts lament the fact that people still rely on traditional face-to-face communication when the topic is really vital. Such experts believe the ordinary users are just backward. Actually, ordinary people are exhibiting a wisdom based on millions of years of social interaction. They realize we should distrust the Internet! Evolution taught our species to avoid a bad deal even before we invented writing. There is something wrong with the Internet that is obvious at the gut level. Forget technology, we need to address the gut issue of risk to life and property.
To see what needs fixing on the Internet, we need to return to the question: why does cooperation emerge as the successful strategy in a world based on individual competition? That question was answered brilliantly by Dr. Robert Axelrod in his work on the evolution of cooperation. Briefly the answer is that cooperation is a highly effective competitive strategy. People who adopt a pattern of cooperation succeed against people who adopt alternatives like pure competition and hierarchical control. Under the proper conditions, cooperation evolves in a competitive landscape, drives out unproductive patterns of competition, and everybody wins. Under poor conditions, competition reduces a society to a subsistence level economic state and political subservience to a foreign power with a better model for social interaction.
In the Pygar Software Project, we will take Axelrod’s ideas, apply them to the Internet, and create conditions that foster cooperation and ensure a prosperous, secure society. To explain the idea takes several steps.
- First, we need to discuss the research on cooperation,
- then look at the successful strategies,
- next look at how the current Internet undermines and prevents those successful strategies
- and finally explain how Pygar fixes the problems with the Internet.
Axelrod’s Research on Cooperation
The study of cooperation predates the Internet revolution. University of Michigan Political Science Professor Robert Axelrod propelled the subject into public discussion with the publication of his book The Evolution of Cooperation in 1984. This work provides the social engineering context for the Pygar Project.
In his study, Axelrod and his collaborators showed that humans — and indeed most social animals — exhibit an ability to cooperate with others as a means to maximize self-interest. The study explained how this cooperation arises spontaneously when conditions are right. For many people, the most surprising observation in the study was the volume of evidence to show that voluntary, elective, cooperation is essential within government, corporations and even military units even though those organizations are theoretically based on formal, legal obligations. Organizations whose structure and culture foster natural social cooperation tend to succeed while those that cobble it while relying on hierarchical authority eventually fail.
Axelrod’s notion of the Evolution of Cooperation is the product of both experiments and historical analysis. Reducing the historical patterns to idealized game-theoretical strategies, Axelrod showed that successful strategies follow closely a simple, idealized strategy called tit-for-tat. Tit-for-tat advises competitors to cooperate at first and then take the cue for their next action from how the other player acted on the previous move. The strategy is cooperative because cooperation is the default decision, but the strategy is also principled because it looks out for self-interest. A player following tit-for-tat will not continue cooperation if the other player fails to respond in kind. Lastly, the strategy is forgiving because it reciprocates cooperation in the present moment and forgets past transgressions.
Prof. Axelrod’s argument considers a multitude of issues and draws on many lines of evidence. Only the broad conclusions can be given here. Successful patterns of cooperation evolve during a competitive game when several assumptions are true:
- the players interact often and expect to remain players indefinitely,
- the players know each other or at least can attribute actions reliably to the responsible agent, and
- each interaction is brief and less consequential than the sum of repeated actions.
Axelrod’s analysis showed that not only are these conditions necessary for tit-for-tat to succeed, they are also essential conditions for all voluntary cooperation. For a detailed analysis, the reader should turn to Axelrod’s book or any of the subsequent studies. The three assumptions are enough to explain the limitations of established technology and the value of the new approach applied in Pygar.
Critique of Integrated Information Networks
Let us consider today’s attempts at cooperative activity on the Internet in the context of the three assumptions required for principled cooperation.
Current procedures leave servers accessible to individuals with no confirmed identity. It is difficult to attribute information, requests, or orders to a fixed identity; thus, the second assumption is violated.
The Pygar Project addresses the identity problem with digital IDs — that is not an novel idea of course — but Pygar adds another important change. Interactions take place with the aid of a broker. Over time, people can develop trust with an intermediary — for example a Google or E-bay — and that trust may alleviate some of the fears of dealing with less well-known parties.
The largest obstacle by far to successful widespread integration of information via the Internet is the violation of the third principle of cooperation: The risk from each interaction should be bounded and small. In contrast, a large open, database is exposed to serious exploitation with potentially grievous consequences. Consequently, the most important databases are closed and private. The people who guard such data do not cooperate; in fact, it may be illegal for them to do so.
The Advantages of Pygar’s Approach
The Pygar Project’s goal is to protect most of an agent’s information resources from misuse. Sharing is based on small quantities of information that are exchanged for a specific reason. Although the limited amount of shared data represents a risk, the bulk of the data are protected. The parties involved in the exchange have verifiable identities so that future cooperation can be extended or withheld based on the past behavior of the other party. The significance of the small quantity of shared data is ensured by a blind broker. The blind broker is prevented from reading any data at all, but the methods allow a blind broker to determine when hidden data items meet the search requirements of the parties who are engaged in the private cooperation.
Thus, with Pygar, the interaction between parties consists of an extended series of small, descretionary two-way interactions as opposed to large-scale, broadly-inclusive sharing agreements. The Pygar Internet method helps develop a long-term mutual interest in the relationship by preventing a major betrayal of trust and rewarding long-term, mutually beneficial cooperation.