| Chapter 1: | Conceptual Framework for Collective Action |
|
|
Table 1.1. Payoff matrix for actors in a game.
| Actor A | Actor B | ||
| Cooperate | Defect | ||
| a | b | ||
| Cooperate | a | -1 | 0 |
| Defect | b | 0 | 1 |
the minimum possible. Hence, if B defects and plays column b,b, and A cooperates in the row a,b, B gains 1, and A's gain is 0. In such a cautious environment, if each suspects the other might defect (remembering that A's gain is B's loss and vice versa), their strategy will be A = (a,b) and B = (a,b), and the gains for both will be zero in the absence of cooperation.8 Since, however, each expects the other to defect, they would both make similar decisions so as to gain the best possible outcome, and no matter what each player does, and even if they knew what the other would do, this will not necessarily change their initial position to defect.
The limitations of this model in defining cooperation are apparent in the potential to misrepresent the broader settings in which behaviours occur. The problem arises from the many and imprecise meanings of cooperation, and the rational actor model tends to assume that cooperation denotes forcing or expecting an actor to behave in a manner that favours one actor's preferences. Definitions like this introduce functional issues that lack precise explanation. Among these are issues which, Jervis observed, include how to classify a behaviour when one side desires a high degree of friction, and the other responds with the sought for hostility: ‘What do we say about cases in which neither side thinks about the impact on the other? Does the impact of defection imply something more than or different from, non-cooperation?’9 A concept of collective action does not necessarily ignore these issues; on the contrary, it acknowledges them but falls short of focusing exclusively on rational calculations of the units or the degree of hostility among actors.