Documentos de Académico
Documentos de Profesional
Documentos de Cultura
$XWKRUV5REHUW$[HOURG
6RXUFH7KH$PHULFDQ3ROLWLFDO6FLHQFH5HYLHZ9RO1R-XQSS
3XEOLVKHGE\$PHULFDQ3ROLWLFDO6FLHQFH$VVRFLDWLRQ
6WDEOH85/http://www.jstor.org/stable/1961366 .
$FFHVVHG
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
American Political Science Association is collaborating with JSTOR to digitize, preserve and extend access to
The American Political Science Review.
http://www.jstor.org
This article investigates the conditions under which cooperation will emerge in a world of egoists
without central authority. This problem plays an important role in such diverse fields as political
philosophy, international politics, and economic and social exchange. The problem is formalized as
an iterated Prisoner's Dilemma with pairwise interaction among a population of individuals.
Results from three approaches are reported: the tournament approach, the ecological approach,
and the evolutionary approach. The evolutionary approach is the most general since all possible stra-
tegies can be taken into account. A series of theorems is presented which show: (I) the conditions
under which no strategy can do any better than the population average if the others are using the
reciprocal cooperation strategy of TIT FOR TAT, (2) the necessary and sufficient conditions for a
strategy to be collectively stable, and (3) how cooperation can emerge from a small cluster of dis-
criminating individuals even when everyone else is using a strategy of unconditional defection.
Under what conditions will cooperation emerge ment for mutual defection). By assumption,
in a world of egoists without central authority? P > S, so it pays to defect if the other player de-
This question has played an important role in a fects. Thus no matter what the other player does,
variety of domains including political philosophy, it pays to defect. But if both defect, both get P
international politics, and economic and social ex- rather than the R they could both have got if both
change. This article provides new results which had cooperated. But R is assumed to be greater
show more completely than was previously pos- than P. Hence the dilemma. Individual rationality
sible the conditions under which cooperation will leads to a worse outcome for both than is possi-
emerge. The results are more complete in two ble.
ways. First, all possible strategies are taken into To insure that an even chance of exploitation or
account, not simply some arbitrarily selected sub- being exploited is not as good an outcome as mu-
set. Second, not only are equilibrium conditions tual cooperation, a final inequality is added in the
established, but also a mechanism is specified standard definition of the Prisoner's Dilemma.
which can move a population from noncoopera- This is just R>(T+S)12.
tive to cooperative equilibrium. Thus two egoists playing the game once will
The situation to be analyzed is the one in which both choose their dominant choice, defection,
narrow self-maximization behavior by each per- and get a payoff, P. which is worse for both than
son leads to a poor outcome for all. This is the the R they could have got if they had both coop-
famous Prisoner's Dilemma game. Two individu- erated. If the game is played a known finite num-
als can each either cooperate or defect. No matter ber of times, the players still have no incentive to
what the other does, defection yields a higher pay-
off than cooperation. But if both defect, both do
worse than if both cooperated. Figure 1 shows the Cooperate Defect
payoff matrix with sample utility numbers at-
tached to the payoffs. If the other player coop- Cooperate R3, R3 S=O,7T5
erates, there is a choice between cooperation
which yields R (the reward for mutual coopera-
tion) or defection which yields T (the temptation Defect T-5, S=0 P-1, -1
to defect). By assumption, T >R, so it pays to de-
fect if the other player cooperates. On the other T>R>P>S
hand, if the other player defects, there is a choice R > (S+T)/2
between cooperation which yields S (the sucker's
payoff), or defection which yields P (the punish- Source: Robert Axelrod, "Effective Choice in the
Prisoner'sDilemma,"Journal of Conflict Resolution
24 (1980): 3-25; "More Effective Choice in the
Prisoner'sDilemma,"Journal of Conflict Resolution
I would like to thank John Chamberlin,Michael 24 (1980): 379-403.
Cohen, BernardGrofman, William Hamilton, John
Kingdon,LarryMohr, John Padgettand ReinhardSel- Note: The payoffs to the row chooserare listed first.
ten for their help, and the Instituteof Public Policy
Studiesfor its financialsupport. Figure1. A Prisoner'sDilemma
306
to make some assumptions about micro-motives, utilities. The utilities already include whatever
and then deduce consequences for macro- consideration each player has for the interests of
behavior (Schelling, 1978). Thinking about the the other (Taylor, 1976, pp. 69-83).
paradigm case of a legislature is a convenience, Under these conditions, words not backed by
but the same style of reasoning can apply to the actions are so cheap as to be meaningless. The
emergence of cooperation between individuals in players can communicate with each other only
many other political settings, or even to relations through the sequence of their own behavior. This
between nations. While investigating the condi- is the problem of the iterated Prisoner's Dilemma
tions which foster the emergence of cooperation, in its fundamental form.
one should bear in mind that cooperation is not Two things remain to be specified: how the pay-
always socially desirable. There are times when off of a particular move relates to the payoff in a
public policy is best served by the prevention of whole sequence, and the precise meaning of a
cooperation-as in the need for regulatory action strategy. A natural way to aggregate payoffs over
to prevent collusion between oligopolistic business time is to assume that later payoffs are worth less
enterprises. than earlier ones, and that this relationship is ex-
The basic situation I will analyze involves pair- pressed as a constant discount per move (Shubik,
wise interactions.' I assume that the player can 1959, 1970). Thus the next payoff is worth only a
recognize another player and remember how the fraction, w, of the same payoff this move. A
two of them have interacted so far. This allows whole string of mutual defection would then have
the history of the particular interaction to be a "present value" of P+ wP+ w2P+ w3P. . . =
taken into account by a player's strategy. P/(1-w). The discount parameter, w, can be
A variety of ways to resolve the dilemma of the given either of two interpretations. The standard
Prisoner's Dilemma have been developed. Each economic interpretation is that later consumption
involves allowing some additional activity which is not valued as much as earlier consumption. An
alters the strategic interaction in such a way as to alternative interpretation is that future moves may
fundamentally change the nature of the problem. not actually occur, since the interaction between a
The original problem remains, however, because pair of players has only a certain probability of
there are many situations in which these remedies continuing for another move. In either interpreta-
are not available. I wish to consider the problem tion, or a combination of the two, w is strictly be-
in its fundamental form. tween zero and one. The smaller w is, the less im-
1. There is no mechanism available to the play- portant later moves are relative to earlier ones.
ers to make enforceable threats or commitments For a concrete example, suppose one player is
(Schelling, 1960). Since the players cannot make following the policy of always defecting, and the
commitments, they must take into account all other player is following the policy of TIT FOR
possible strategies which might be used by the TAT. TIT FOR TAT is the policy of cooperating
other player, and they have all possible strategies on the first move and then doing whatever the
available to themselves. other player did on the previous move. This
2. There is no way to be sure what the other means that TIT FOR TAT will defect once for
player will do on a given move. This eliminates the each defection by the other player. When the
possibility of metagame analysis (Howard, 1971) other player is using TIT FOR TA T, a player who
which allows such options as "make the same always defects will get T on the first move, and P
choice as the other player is about to make." It on all the subsequent moves. The payoff to some-
also eliminates the possibility of reliable reputa- one using ALL D when playing with someone us-
tions such as might be based on watching the ing TIT FOR TAT is thus:
other player interact with third parties.
3. There is no way to change the other player's V(ALL D I TFT) T + wP + w2P + w3P. . .
=T + wP (1 + w + w2. . .)
= T + wP/(l-w).
'A single player may be interacting with many others,
but the player is interacting with them one at a time. The Both ALL D and TIT FOR TA Tare strategies.
situations which involve more than pairwise interaction In general, a strategy (or decision rule) is a func-
can be modeled with the more complex n-person Pri- tion from the history of the game so far into a
soner's Dilemma (Olson, 1965; G. Hardin, 1968; R.
probability of cooperation on the next move.
Hardin, 1971; Schelling, 1973). The principal applica-
tion is to the provision of collective goods. It is possible Strategies can be stochastic, as in the example of a
that the results from pairwise interactions will help sug- rule which is entirely random with equal proba-
gest how to undertake a deeper analysis of the n-person bilities of cooperation and defection on each
case as well, but that must wait. For a parallel treatment move. A strategy can also be quite sophisticated in
of the two-person and n-person cases, see Taylor (1976, its use of the pattern of outcomes in the game so
pp. 29-62). far to determine what to do next. It may, for ex-
since the best A can do from move n onwards if Vn+1(A JB) < V(BIB)-
B defects is get P each time. Moving the second wn-1(T+ wP/( -w)) + wn-1T
term to the right side of the inequality gives the
helpful result that B has a secure position over Vn+1(AJB)< V(BIB) -w np/(I - W).
A on move n if Vn(AIB) is small enough,
namely if Therefore, B always has a secure position over
A, and consequently B is a collectively stable
Vn(AIB)< V(BIB)-wn-1P(l-w). (1) strategy.
The second part of the proof operates by
The theorem which follows embodies the contradiction. Suppose that B is a collectively
advice that if you want to employ a collectively stable strategy and there is an A and an n such
stable strategy, you should only cooperate that B does not defect on move n when
when you can afford an exploitation by the
other side and still retain your secure position. Vn(A IB) > V(BIB) - wn- (T + wP/(1 -w)),
i.e., when
THEOREM 3. THE CHARACTERIZATION Vn(A IB) + w n-I (T + wPI(I -w)) > V(BIB).
THEOREM. B is a collectively stable strategy if (2)
and only if B defects on move n whenever the
other player's cumulative score so far is too Define A' as the same as A on the first n-I
great, specifically when Vn(AIB) > V(BIB) - moves, and D thereafter. A' gets T on move n
wn- I (T+wP/( 1-w)). (since B cooperated then), and at least P
thereafter. So,
X- V(B B)
V,(AlB)
May Defect
Moven Time
Source: Drawn by the author.
To consider arrival in clusters rather than 9For related concepts from biology, see Wilson (1979)
singly, we need to broaden the idea of "inva- and Axelrod and Hamilton (1981).
of Washington Society. New York: Scribner's. Warner, Rex, trans. (1960). War Commentaries of
Snyder, Glenn H. (1971). " 'Prisoner's Dilemma' and Caesar. New York: New American Library.
'Chicken' Models in International Politics." Inter- Wilson, David Sloan (1979). Natural Selection of Popu-
national Studies Quarterly 15: 66-103. lations and Communities. Menlo Park, Calif.: Ben-
Taylor, Michael (1976). Anarchy and Cooperation. jamin / Cummings.
New York: Wiley. Young, James Sterling (1966). The Washington Com-
Trivers, Robert L. (1971). "The Evolution of Recipro- munity, 1800-1828. New York: Harcourt, Brace &
cal Altruism." Quarterly Review of Biology 46: World.
35-57.