Social Semantic Web Paper Summaries 1: The Anatomy of Cooperation by Peter Kollock


In 2018, I have taken a master’s course from Bogazici University, called Social Semantic Web (CMPE 58H), teached by Suzan Üsküdarlı. (

I have wrote summaries for a few papers there. Today, I have decided to share them with you. This is both to share my understanding on these papers and to show my approach on how to read papers. Summary of the first paper is below.


2018 yılında Boğaziçi Üniversitesinde Sosyal Semantik Web (CMPE 58H) diye bir ders almıştım, Suzan Üsküdarlı hocamız dersi veriyordu.(

O derste birkaç makale özeti yazdım. Şimdi, o özetleri paylaşmaya karar verdim. Bunu, hem bu makalelere bakış açımı yansıtsın, ilgililer varsa okusun diye yapıyorum, hem de kendi makale okuyuş yöntemimi paylaşmak için yapıyorum. İlk makalenin özeti aşağıda.


The Anatomy of Cooperation

Citation: Peter Kollock

About Author:

Peter Kollock was a sociologist that worked on the topics trust and collective action, particularly on virtual communities. He had an interest in Buddhism and motorcycles. He died in a motorcycle accident.


The paper discourses types of social dilemmas, their solutions, factors that contribute or prevent the dilemmas, challanges and key points to study social dilemmas.


1-Social Dilemma Types

a-Dyadic Dilemmas:

  • Prisoner’s Dilemma
  • Assurance Game
  • Chicken Game

b-N-Person Dilemmas:

  • Tragedy of Commons
  • Public Goods (also Production Functions)
  • Rival and Unrival (Commons-Public Goods) Goods.

Factors that differ between N-person and Dyadic Dilemmas:

  • Anonimity
  • Ability to impact others
  • Diffusion of the harm caused by defection

2- Solutions to Social Dilemmas


  • 1-Social Value Orientations: Cooperative, Competitive, Individualist, Altruist
  • 2-Communication: Gathering info about others’ choices, Commitments and Promises, Moral Suasion
  • 3- Creating a sense of group identity: Could be effective for strategic self-seeking of group members as well as altruist reasons of group members.


  • 1- Reciprocity: Tit for Tat (no envy, don’t be the first defector, reciprocate both cooperation and defection, make your partner understand this strategy you are using)
  • 2- Choosing Partners: Out for Tat (choose someone else to play with, when someone defects)
  • 3- Grim Triggers: For N-person dilemmas, make every single person determined about quitting contribution when there is one free-rider in the group. Not many real life applications of this method exist.
  • 4- Social Learning: When assumed penalty aversive actors rather than “Profit=Award-Penalty” actors, some tresholds(?) and imitation tendency can augment cooperation in N-person groups.
  • 5- Group Reciprocity: Effect that a group imposes to its members through potential reciprocal exhanges in the future, sometimes even without intention (group reciprocity is a deep heuristic in humans).Could be the main reason behind the effects of motivational solution of creating a group identity.
  • c-Structural:
  • 1-Iteration,Identifiability: Make interactions more durable and frequent, make it easier for group members to identify each other, increase information of participants about other participant’s past actions.
  • 2-Payoff Structure: Make benefits of cooperation higher and defection lower.
  • 3- Efficacy: Form a subgroup of members that are responsible for contributing. Make the production function step-levelled. Perception of efficacy is actually enough, rather than its true self.
  • 4- Group Size: Effective because of diffusion of the harm, ease of shaping others’ behaviour, anonimity, ease of communication and coordination, efficacy, visibility of actions and ease of monitoring members’ actions. Although these factors exist, group size does not always harm cooperation.

Why it is difficult to measure “Group Size Effect”: So many parameters change when group size changes. Logistic problems occur when experimenting with big groups.

Benefit of Heterogenity: If a group has enough variety to house many types of people, and if a public good is unrival enough, a subgroup of people which like serving the good may pop up inside the group.

  • 5-Boundaries: Make some entity a regulator. Can be done by creating a “Leviathan” or Privatizing the good or making the locals that know the good well regulate the good.
  • 6-Sanctions: Inducements for cooperation, punishments for defection (such as ostracism or exclusion)


Author used an approach of analyzing and combining the results of past experiments via citations of different articles and via talks that he had done with articles’ authors. He gave real life examples of those results (such as feeding poor children example). He also mentioned the difficulties that occur when doing such experiments.

Author’s conclusions:

Author concludes that social dilemmas are results of people rationing for their own interests. However a better set of results could be reached in the light of collective rationality. Social dilemmas are unefficient equlibriums that could be resolved by creating conditions of recompensation for a party’s actions. Topic of social dilemmas is a hard-to-research domain which could be explained by the vast number of variables that effect cooperation rates.

My conclusions

Agreeing with the author, I am surprised by the number of solutions that are proposed for social dilemmas. I think there is an impressive amount of ways to compensate people’s self-interest-oriented actions. With more resources dedicated to research, even more number of techniques could be discovered. (Or probably some have been discovered since 20 years) Humans mainly seem to act by taking account of their own interest and of the interest of their kith and kin.

A perspective that I have formed on this topic after reading this article is that there are two ways of preventing social dilemmas, when the ways are categorized based on the parts they influence on mind. First is influencing any party’s ration, second is influencing any party’s instincts:

  • 1- Example for influencing ration: Making some human-like agent’s (firm, community, government, defector themself)interests to be put on the stake in lieu of compensation. This readjusts the rational equation of “benefit” for that particular party, making it act for compensation. Or, it makes the defecting parties to think again before defecting, because it changes the benefits and losses of that particular defection.
  • 2- Example for influencing instincts: Influencing defector agents psychologically by using their instincts of empathy, group pressure etc. even when it is actually benefitial to defect. This has no relation with rationing of the agents. It only deals with some heuristic formations in mind that influences behaviour. Defecting agents may not even realize the reason they’ve decieded to cooperate.


Being cautious about favoring any solution or analysis more than it is needed, author’s style of thinking seems to be great. Reviewing many articles and summarizing them all in one article is a great way of making research results become more avalible to people which are interested in using these results in practical life. Only because that it has been two decades after it has been published and because it possibly lacks many research that has been done ever since, my rating is 8/10 for this article.

NLP Engineer at PragmaCraft. Former Researcher at Bogazici University Medical Imaging Lab. twitter:ahmetmeleq ///