본문 바로가기

[연구] Research

[IR-Anarchy and Cooperation] Axelrod, Robert (1984) The Evolution of Cooperation

Axelrod, Robert (1984) The Evolution of Cooperation 관련 유용.



The Evolution of Cooperation: Revised Edition

저자
Axelrod, Robert 지음
출판사
Basic Books | 2006-12-04 출간
카테고리
과학/기술
책소개
Updated for the first time, the cla...
가격비교 글쓴이 평점  

 

ü  Axelrod, Robert (1984) The Evolution of Cooperation, chs. 2 & 9

 


Review:

Axelrod’s experiment is a profound work in the application of game theory to political science, shedding new lights to how states can cooperate in the neorealist setting. Based on the use of theory, computer simulation, and historical examples, Axelrod argues here for the possibility of cooperative strategies in the world of perpetual rivalry (anarchy). Unlike the realist tenet that power maximization - zero-sum game (at the expense of others) is the best way of survival, Axelrod’s notion of reiterated interactions (reiterated prisoner’s dilemma) enlightens how states’ encounters are rarely one-time based in reality, how they are able to learn the consequences of their strategies in long term (shadow of future), which in turn create the incentives for egoist actors to cooperate rather than defect despite the absence of any central authority. Following Axelrod’s logic/perspective of the world where actors interact in multiple times, a system of reciprocity becomes a possible cooperative solution: Viewed in this way – reiterated games – evolution of cooperation is viable among these rational egoist states as “The foundation of cooperation is not really trust, but the durability of relationship... Whether the players trust each other or not is less important in the long run than whether the conditions are ripe for them to build a stable pattern of cooperation with each other (p. 182).” Furthermore, “The individuals do not have to be rational: the evolutionary process allows the successful strategies to thrive, even if the players do not know why or how (p. 174)”; “Likewise, there is no need to assume trust between the players: the use of reciprocity can be enough to make defection unproductive. Altruism is not needed: successful strategies can elicit cooperation even from an egoist. Finally, no central authority is needed: cooperation based on reciprocity can be self-policing (p. 175).”

As a sidenote, Axelrod’s discovery of reiterated game (shadow of future – anticipation of continued interaction, that states’ encounters are not one-time based) counters Mearsheimer’s central argument (offensive realism) that states possess the incentives to use offensive strategies to maximize their rate of survival. Instead, Axelrod’s notion on “limited provocability” is in agreement with Waltz’s proposal of the defensive realist theory, that it is “actually better to respond quickly to a provocation” but not defect first (TIT FOR TAT strategy).

Overall, Axelrod’s work is a highly honorable approach to social science, in which he devotes to go beyond mere explanation of the world and create a system where cooperation and peace become possibility (solution-oriented approach) – a respectable question, experiment, and creativity on how to promote cooperation and avoid a Hobbesian war of all against all.

 

Yet, some criticisms:

-    Applicability to real world – Prisoner’s dilemma with non-interruptable communication: unlike the designed setting of the game, the real world involves multiple channels of communication among the two parties(states). Also, the real world prospect of cooperation depends not only on the pairs of cooperation but also on multiple parties and overlapping pairs of relationships (Games).

-    Individual game to collective game decreases the solidity of Axelrod’s results

-    Additionally, some scholars have questioned Axelrod’s use of WWI trench example…

-    Note on the possibility of misperception: TIT FOR TAT proved to be the best rule, although “it got into a lot of trouble when a single misunderstanding led to a long echo of alternating retaliations, it could often end the echo with another misperception” (p. 183) -> In reality, states’ dislike or fear of the costs of trials-and-errors are in many cases insurmountable and/or detrimental – can states afford to practice?

-     Implications are in no doubt strong, but implementation prospect should be less optimistic than how Axelrod describes.

 

Table of Contents (Axelrod)

Preface

PART I. Introduction

1.       The Problem of Cooperation

PART II. The Emergence of Cooperation

2.       The Success of TIT FOR TAT in Computer Tournaments

3.       The Chronology of Cooperation

PART III. Cooperation Without Friendship of Foresight

4.       The Live-and-Let-Live System in Trench Warfare in World War I

5.       The Evolution of Cooperation in Biological Systems (with William D. Hamilton)

PART IV. Advice for Participants and Reformers

6.       How to Choose Effectively

7.       How to Promote Cooperation

PART V. Conclusions

8.       The Social Structure of Cooperation

9.       The Robustness of Reciprocity

APPENDIX A

Tournament Results

APPENDIX B

Proofs of the Theoretical Propositions

Notes

Bibliography

Index

 

Essense:

-          Problem: Iterated Prisoner’s Dilemma = both mutual gains from cooperation and also exploitation are possible.

-          To find the best strategy: A computer simulation - Computer Prisoner’s Dilemma Tournament (1st round: invited game theorists in economics, psychology, sociology, political science, and mathematics for entries -14 entries in total)

-          Best strategy: TIT FOR TAT – “merely the strategy of starting with cooperation, and thereafter doing what the other player did on the previous move” (preface p. viii)

-          2nd round of computer hobbyists and other professors in evolutionary biology, physics, and computer science, in addition to the fields in the 1st round, again proved TIT FOR TAT as the best strategy (RECIPROCITY)

-          Results extended to the question of fostering cooperative conditions among individuals, organizations, and nations… 3 part MAIN Arguments:

1)       TIT FOR TAT is the effective strategy in iterated prisoner’s dilemma (begin by cooperation then do what your opponent did in the previous step)

2)       This strategy as the best strategy in important real-life settings (WWI trench example- start by cooperating and this continues)

3)       Evolutionary: this strategy is natural and hence expect this to prevail in nature

-          Evolutionary Perspective on cooperation (reiteration, shadow of future, tit for tat): Unlike the common one-time play of the prisoner’s dilemma of the time, Axelrod’s reiterated version showcased how straightforward cooperation can outcompete the benefits of defection.


Summary

Chapter 1. The Problem of Cooperation

*Main Question:

“Under what conditions will cooperation emerge in a world of egoists without central authority?”

Ø  “a world of egoists”: state as the actor, selfish (Hobbes, Morgenthau)

Ø  “without central authority”: anarchy (structural condition – Waltz)

= A question for cooperation in the neorealist setting of the world

=Solution: design a system of cooperation (tit-for-tat – reiterated games) = Neoliberal Institutionalism.

 

“The approach of this book is to investigate how individuals pursuing their own interests will act, followed by an analysis of what effects this will have for the system as a whole. Put another way, the approach is to make some assumptions about individual motives and then deduce consequences for the behavior of the entire system” (p. 6) 

“to develop a theory of cooperation that can be used to discover what is necessary for cooperation to emerge” (p. 6)

*Chapter devoted to introduction and discussion on the original Prisoner’s Dilemma

“What is best depends in part on what the other player is likely to be doing. Further, what the other is likely to be doing may well depend on what the player expects you to do” (Chapter II. p. 27) - *Both parties end up in a mutually despised outcome but cannot commit to the better result due to their selfish individual incentives.

 

Chapter 2. The Success of TIT FOR TAT in Computer Tournaments

*Limitations of previous literatures on Prisoner’s Dilemma:

1)       Do not reveal very much “about how to play the game well” (p. 29)

2)       Mostly one-time, first-time played game analysis

3)       “All together, no more than a few dozen choices” have been analyzed… (p. 29)

4)       Even the ones that focus on “strategic interaction” are limited, studying the dilemma that is “designed to eliminate the dilemma itself by introducing changes in the game” (p. 29)

 

*New experiment needed = a computer tournament among game theorists in the above mentioned five disciplines (1st round)

- New setting (non-zero-sum setting) should account for the “Two important facts about non-zero-sum settings” (p. 30)

Ø  Information on other’s strategies - “what is effective depends not only upon the characteristics of a particular strategy, but also upon the nature of the other strategies with which it must interact” (p. 30)

Ø  Information on the history – “An effective strategy must be able at any point to take into account the history of the interaction as it has developed so far.” (p. 30)

 

*New Experiment:

-          Round robin tournament (each entry paired with each other entry) + RANDOM (a program that randomly cooperates and defects with equal probability)

-          Each game: 200 moves

-          Mutual cooperation: 3 points for both players

-          Mutual defection: 1 point for both players

-          One defect, the other cooperate: defector 5 points, cooperator 0 points

-          Overall, benchmark for good performance is 600 points (always cooperating) and very poor performance at 200 points (always defecting)

-          Result: TIT FOR TAT, submitted by Professor Anatol Rapoport (Univ. Toronto) won – “playing what the other player did on the previous move” (p. 31)

 

*Important Findings:

 

-          The value of cooperating first: all of the eight top-ranking entries were “nice”; that is, they never defected first, at least not until near the end of the game. The “meanies,” which tried to take advantage of the programs that cooperated, often by clever and devious methods, were defeated by a wide margin.

-          The value of forgiving: “One of the main reasons why the rules that are not nice did not do well in the tournament is that most of the rules in the tournament were not very forgiving.”

-          The long-term danger of defection (“echo effect”): “A major lesson of this tournament is the importance of minimizing echo effects in an environment of mutual power. When a single defection can set off a long string of recriminations and counterrecriminations, both sides suffer.” (p. 38)

 

*Interesting strategies:

 

1)       DOWNING – “outcome maximization” principle – deliberating to attempt to understand the other player and then make a choice (probability of other player cooperating after one cooperates and probability of other player cooperating after one defects – each move “updating the estimate of these two conditional probabilities and then selects the choice which will maximize its own long-term payoff”) (p. 34)

Ø  Yet doomed to defect on the first two moves for the deliberation process…

 

2)       FRIEDMAN – lowest scored, least forgiving – “unforgiving rule that employs permanent retaliation. It is never the first to defect, but once the other defects even once, FRIEDMAN defects from then on.” (unlike TIT FOR TAT which “lets bygones be bygones” (p. 36)

 

3)       JOSS – “a sneaky rule that tries to get away with an occasional defection… it always defects immediately after the other player cooperates. But instead of always cooperating after the other player cooperates, 10 percent of the time it defects after the other player cooperates. Thus it tries to sneak in an occasional exploitation of the other player” (p. 36)

 

*2nd round (62 entries), many attempted to develop a better and more complex program – yet the original simple TIT FOR TAT still proved to be the best.

 

*Better Rules do Exist:

Ø  TIT FOR TWO TATS rule: defecting only if the other player defected on the previous two moves -> against our common perception, more gains were made from being even more forgiving (p. 39)

Ø  LOOK AHEAD

Ø  Slight modification of DOWNING (now assuming that other players would be responsive rather than unresponsive)

 

*Three levels of analysis on choice:

1)       First level = analysis of the direct effect of a choice

2)       Second level = analysis of the indirect effects

3)       Third level = Tertiary effects… 부메랑 에펙트와 유사… (“echo effects”)


Overall,
Axelrod draws significant conclusions on the prospect for cooperation: “Mutual cooperation can emerge in a world of egoists without central control by starting with a cluster of individuals who rely on reciprocity.” Furthermore, considering the overall success of TIT FOR TAT, the value of being “nice” and “forgiving” must be noted. Yet, it is also noted that too forgiving strategies (TIT FOR TWO TATS), not retaliating immediately, are unable to survive the game.

 

Chapter 9. The Robustness of Reciprocity

Main Concluding Remarks: The value of reciprocity in the prospect of cooperation

 

*Value of a system, institution that fosters reciprocity:

“The main results of Cooperation Theory are encouraging. They show that cooperation can get started by even a small cluster of individuals who are prepared to reciprocate cooperation, even in a world where no one else will cooperate. The analysis also shows that the two key requisites for cooperation to thrive are that the cooperation be based on reciprocity, and that the shadow of future is important enough to make this reciprocity stable. But once cooperation based on reciprocity is established in a population, it can protect itself from invasion by uncooperative strategies” (p. 173)

 

*Only required info:

1) History of interactions

2) Information on others’ strategies (shadow of the future deliberation)

 

But what is most interesting is how little had to be assumed about the individuals or the social setting to establish these results. The individuals do not have to be rational: the evolutionary process allows the successful strategies to thrive, even if the players do not know why or how. Nor do the players have to exchange messages or commitments (p. 174)… Likewise, there is no need to assume trust between the players: the use of reciprocity can be enough to make defection unproductive. Altruism is not needed: successful strategies can elicit cooperation even from an egoist. Finally, no central authority is needed: cooperation based on reciprocity can be self-policing. (p. 175)”

 

*Overall Cooperation Solution: Institution based on the norm of reciprocity:

-           Examples:

Ø  Diamond markets (“The key factor is that the participants know they will be dealing with each other again and again” p.178)

Ø  Ordinary business transactions (“based upon the idea that a continuing relationship allows cooperation to develop” p. 178)

 

-           “The importance of future interactions can provide a guide to the design of institutions. To help promote cooperation among members of an organization, relationships should be structured so that there are frequent and durable interactions among specific individuals.” (p. 180)

-           “The foundation of cooperation is not really trust, but the durability of relationship. When the conditions are right, the players can come to cooperate with each other through trial-and-error learning about possibilities for mutual rewards, through imitation of other successful players, or even through a blind process of selection of the more successful strategies with a weeding out of the less successful ones. Whether the players trust each other or not is less important in the long run than whether the conditions are ripe for them to build a stable pattern of cooperation with each other” (p. 182).

 

*Note on the possibility of misperception: Still, TIT FOR TAT proved to be the best rule, although “it got into a lot of trouble when a single misunderstanding led to a long echo of alternating retaliations, it could often end the echo with another misperception” (p. 183)

 

*Note on “limited provocability” and the “speed of response”

(in line with Waltz’s defensive realism):

-           “one of my biggest surprises in working on this project” (p. 184)

-           “one should be slow to anger” – the results demonstrate that “it is actually better to respond quickly to a provocation. It turns out that if one waits to respond to (p. 184) uncalled for defections, there is a risk of sending the wrong signal. The longer defections are allowed to go unchallenged, the more likely it is that the other player will draw the conclusion that defection can pay. And the more strongly this pattern is established, the harder it will be to break it… By responding right away, it gives the quickest possible feedback that a defection will not pay.” (p. 185)

-           yet, the danger of provocability – echo effect and the continuation of the conflict

ð  “Limited provocability is a useful feature of a strategy designed to achieve stable cooperation (p. 187)”

 

“It is precisely when this anticipation of future interaction, breaks down that an external authority is invoked.” (p. 179)

 

*As concluding remark, Axelrod admits the slow and painful process of instilling reciprocity in the international system:

“The core of the problem of how to achieve rewards from cooperation is that trial and error in learning is slow and painful. The conditions may all be favorable for long-run developments, but we may not have the time to wait for blind processes to move us slowly toward mutually rewarding strategies based upon reciprocity. Perhaps if we understand the process better, we can use our foresight to speed up the evolution of cooperation” (p. 191)