Social Traps

Social Traps JOHN PLATT Mental Health Research Institute University of Michigan 1 A new area of study is the field that some of us the problem of competitive extermination of the are beginning to call social traps. The term refers last great whales. to situations in society that contain traps formally A converse type of situation might still be re- like a fish trap, where men or organizations or garded as a generalized trap, but perhaps is more whole societies get themselves started in some di- accurately called a countertrap. The considera- rection or some set of relationships that later tion of individual advantage prevents us from doing prove to be unpleasant or lethal and that they see something that might nevertheless be of great bene- no easy way to back out of or to avoid. fit to the group as a whole. It is, so to speak, a Two recent descriptions of traps of this kind social fence rather than a social trap. have already become widely quoted and discussed. A famous, or infamous, example of this kind The first is Garrett Hardin's (1968) article en- was the Kitty Genovese murder in New York City titled "The Tragedy of the Commons." The a few years ago, in which a girl was raped and title refers to situations like that of the Commons, killed in an areaway while more than 30 neighbors or public grassland, of the old New England vil- watched out the windows—and none of them called lages, where anyone could graze his cows freely. the police. This apparent failure of concern and Since this is a "free good" for the owners of cows, action produced a national wave of horror, as well every owner can make money faster by increasing as much recrimination afterward among those in- the number of cattle that he grazes there. But as volved. Yet, in such a situation, it is clear that everyone's number of cattle increases, the grass there is a certain individual barrier against calling gets scarcer until finally it is destroyed entirely, the police. Not only must you tear yourself away and the owners collectively wind up with a loss from the spectacle, but you face the probability of rather than a gain. The trap is that each individ- having to testify in court and even a chance of ual owner continues to do something for his in- being hunted down by the murderer or his friends. Each observer may have felt a strong prick of dividual advantage that collectively is damaging to social conscience at the time, but simply hoped the group as a whole. that someone else would make the troublesome Hardin saw this as the prototype and formal phone call first. analogue of the world population problem, where Many contrasting cases of this kind have been each family may find pleasure and advantage in discussed in a fascinating article by Thomas more babies; and the problem of competitive con- Schelling (1971), "The Ecology of Micromotives." sumption of nonrenewable natural resources; and Schelling cataloged several dozen type situations where individual actions or inactions controlled by immediate personal goals or self-interest, even 1 This article is a report of research in progress by rather weak self-interest, produce long-range so- John Cross, Mel Guyer, Gardner Quarton, and the author, all of whom are affiliated with the Mental Health Research cietal effects which are to almost no one's self- Institute of the University of Michigan. The article was interest. an invited address presented at the annual meeting of For example, he demonstrated how a population the American Psychological Association, Honolulu, Hawaii, September 1972. of red people and green people distributed at Requests for reprints should be sent to John Platt, who random over a chessboard who move from time to is on leave for the 1972-1973 year. His address is Center for Advanced Study in the Behavioral Sciences, 202 Juni- time to new sites can become sharply segregated pero Serra Boulevard, Stanford, California 94305. by color very quickly even if the individuals have AMERICAN PSYCHOLOGIST • AUGUST 1973 • 641 only a mild preference for neighborhoods of the menter or the environment creating a situation same color. or stimulus, S, in which an organism or subject Another example is the decay of railroad service, emits some behavior, B, which is followed by some as people begin to prefer using their cars. As the reinforcement or result, R. We find it helpful to railroad begins to go downhill, still more people write this S-B-R formulation on two lines, as prefer cars. The process is self-accelerating, end- follows: ing up with no one riding the trains, while there are traffic jams on the highways in which everyone B B B S R . S R . S R involved would prefer—too late—to be using the railroad. The process of inflation is likewise self- where the top line represents the actions or outputs accelerating, with every increase in inflation pro- of the organism and the bottom line represents the ducing new demands for raises in wages and actions or inputs of the environment, with an on- profits, which drive up inflation faster. When we going transactional relation between organism and begin to look at such examples, we see that many environment. The S-B-R sequence, even if writ- of our really troublesome social and political prob- ten only once, is thought of as being repeated over lems today are made difficult, not by stupidity or and over again in learning or maintaining behavior. avarice or immorality but by a certain trap com- A positive reinforcement or result, R+, is defined ponent of this kind. as an environmental consequence that makes an Our group at the University of Michigan be- initial behavior, B, more probable when that par- came interested in these questions in the course of ticular type of S occurs again. An aversive conse- studying Skinnerian mechanisms of reinforcement quence or a negative result, R", is one that makes of behavior, examining how they apply to personal a given B less probable the next time S occurs. self-control and to social transactions. We were This is equivalent to a feedback formulation, trying to make formulations of the behavioral re- in which S is an input field to the organism, B is sults that might be applicable to several disciplines, the motor output from the organism back to the my own interest being that of general systems environment, and R is the "reafferent stimulation," theory; John Cross, an economist, was concerned or change of input field from the environment with microeconomics and bargaining; Mel Guyer, which gives the changed "error signal" to the or- a game theorist, was interested in the locked-in ganism and closes the feedback loop, as seen in conflict and cooperation modes of non-zero-sum Figure 1. games; and Gardner Quarton, a psychiatrist, was A frequent objection to the Skinnerian formula- interested in the explanatory and therapeutic pos- tion is that the definition of reinforcement or of a sibilities of behavioral reinforcement. positive reinforcer is "circular," as in saying, "be- After reading the Hardin article and later the havior is made more probable by positive reinforce- Schelling article, we suddenly saw that a number ment." But this is very similar to the useful Dar- of their social trap and countertrap situations winian phrase, "the survival of the fittest," which could be formalized in a reinforcement language. This quickly led to a useful classification of dif- Organism ferent kinds of traps, with interesting parallels between what had seemed to be very different /B problems. This formulation led in turn to several S' 1 R suggestions of personal and social methods of self- control for getting out of these traps. Here I Environment want to outline some of these new findings. Reinforcements and Behavior First, however, it may be helpful to describe our FIG. 1. Feedback loop where S = input field to the organism, B = motor output from the organism back way of formalizing the Skinnerian results showing to the environment, and R = the change of input field the effects of reinforcement on behavior. Skinner from the environment which gives the changed "error uses a three-term formulation, with the experi- signal" to the organism and closes the feedback loop. 642 • AUGUST 1973 • AMERICAN PSYCHOLOGIST is likewise a circular definition of fitness. In Here, the immediate punishment (or its expecta- fact, Skinner seems to think of the evolution of tion after some experience) tends to block behavior behavior in an organism from childhood as in- B, even though there would be a long-run reward. volving a similar "natural selection" of behaviors In these formulations, we have a behavioral defi- that "work" and therefore "survive" in a given nition of the exact traditional meaning of the words stimulus situation, presumably because there are trap and fence. internal neurophysiological loops of response that A similar concept of reversal oj reinforcers can are reinforced while others are eliminated. He be applied to individual-group traps, where it is sees the evolution of a repertoire of behaviors in not a question of shorter and longer times so much an organism as like the evolution of species by as the fact that the personal reward or punishment, survival in an ecosystem. "Reinforcers" are in RP+ or Rr-, is in opposition to the collective or + fact defined much better than the Darwinian "fit- group advantage or disadvantage, RG or RG~. ness" of species, and reinforcers such as food, Again we can have traps or fences depending on water, sex, petting, praise, and random jackpots whether the initial personal result is positive or of food or money operate similarly and predictably negative.

Load more