Social Traps

JOHN PLATT Mental Health Research InstituteUniversity of Michigan 1

A new area of study is the field that some of usare beginning to call social traps. The term refersto situations in society that contain traps formallylike a fish trap, where men or organizations orwhole societies get themselves started in some di-rection or some set of relationships that laterprove to be unpleasant or lethal and that they seeno easy way to back out of or to avoid.

Two recent descriptions of traps of this kindhave already become widely quoted and discussed.The first is Garrett Hardin's (1968) article en-titled "The Tragedy of the Commons." Thetitle refers to situations like that of the Commons,or public grassland, of the old New England vil-lages, where anyone could graze his cows freely.Since this is a "free good" for the owners of cows,every owner can make money faster by increasingthe number of cattle that he grazes there. But aseveryone's number of cattle increases, the grassgets scarcer until finally it is destroyed entirely,and the owners collectively wind up with a lossrather than a gain. The trap is that each individ-ual owner continues to do something for his in-dividual advantage that collectively is damaging tothe group as a whole.

Hardin saw this as the prototype and formalanalogue of the world population problem, whereeach family may find pleasure and advantage inmore babies; and the problem of competitive con-sumption of nonrenewable natural resources; and

1 This article is a report of research in progress byJohn Cross, Mel Guyer, Gardner Quarton, and the author,all of whom are affiliated with the Mental Health ResearchInstitute of the University of Michigan. The article wasan invited address presented at the annual meeting ofthe American Psychological Association, Honolulu, Hawaii,September 1972.

Requests for reprints should be sent to John Platt, whois on leave for the 1972-1973 year. His address is Centerfor Advanced Study in the Behavioral Sciences, 202 Juni-pero Serra Boulevard, Stanford, California 94305.

the problem of competitive extermination of thelast great whales.

A converse type of situation might still be re-garded as a generalized trap, but perhaps is moreaccurately called a countertrap. The considera-tion of individual advantage prevents us from doingsomething that might nevertheless be of great bene-fit to the group as a whole. It is, so to speak, asocial fence rather than a social trap.

A famous, or infamous, example of this kindwas the Kitty Genovese murder in New York Citya few years ago, in which a girl was raped andkilled in an areaway while more than 30 neighborswatched out the windows—and none of them calledthe police. This apparent failure of concern andaction produced a national wave of horror, as wellas much recrimination afterward among those in-volved. Yet, in such a situation, it is clear thatthere is a certain individual barrier against callingthe police. Not only must you tear yourself awayfrom the spectacle, but you face the probability ofhaving to testify in court and even a chance ofbeing hunted down by the murderer or his friends.Each observer may have felt a strong prick ofsocial conscience at the time, but simply hopedthat someone else would make the troublesomephone call first.

Many contrasting cases of this kind have beendiscussed in a fascinating article by ThomasSchelling (1971), "The Ecology of Micromotives."Schelling cataloged several dozen type situationswhere individual actions or inactions controlledby immediate personal goals or self-interest, evenrather weak self-interest, produce long-range so-cietal effects which are to almost no one's self-interest.

For example, he demonstrated how a populationof red people and green people distributed atrandom over a chessboard who move from time totime to new sites can become sharply segregatedby color very quickly even if the individuals have

AMERICAN PSYCHOLOGIST • AUGUST 1973 • 641

only a mild preference for neighborhoods of thesame color.

Another example is the decay of railroad service,as people begin to prefer using their cars. As therailroad begins to go downhill, still more peopleprefer cars. The process is self-accelerating, end-ing up with no one riding the trains, while thereare traffic jams on the highways in which everyoneinvolved would prefer—too late—to be using therailroad. The process of inflation is likewise self-accelerating, with every increase in inflation pro-ducing new demands for raises in wages andprofits, which drive up inflation faster. When webegin to look at such examples, we see that manyof our really troublesome social and political prob-lems today are made difficult, not by stupidity oravarice or immorality but by a certain trap com-ponent of this kind.

Our group at the University of Michigan be-came interested in these questions in the course ofstudying Skinnerian mechanisms of reinforcementof behavior, examining how they apply to personalself-control and to social transactions. We weretrying to make formulations of the behavioral re-sults that might be applicable to several disciplines,my own interest being that of general systemstheory; John Cross, an economist, was concernedwith microeconomics and bargaining; Mel Guyer,a game theorist, was interested in the locked-inconflict and cooperation modes of non-zero-sumgames; and Gardner Quarton, a psychiatrist, wasinterested in the explanatory and therapeutic pos-sibilities of behavioral reinforcement.

After reading the Hardin article and later theSchelling article, we suddenly saw that a numberof their social trap and countertrap situationscould be formalized in a reinforcement language.This quickly led to a useful classification of dif-ferent kinds of traps, with interesting parallelsbetween what had seemed to be very differentproblems. This formulation led in turn to severalsuggestions of personal and social methods of self-control for getting out of these traps. Here Iwant to outline some of these new findings.

Reinforcements and BehaviorFirst, however, it may be helpful to describe ourway of formalizing the Skinnerian results showingthe effects of reinforcement on behavior. Skinneruses a three-term formulation, with the experi-

menter or the environment creating a situationor stimulus, S, in which an organism or subjectemits some behavior, B, which is followed by somereinforcement or result, R. We find it helpful towrite this S-B-R formulation on two lines, asfollows:

BS R . .

BS R . .

BS R

where the top line represents the actions or outputsof the organism and the bottom line represents theactions or inputs of the environment, with an on-going transactional relation between organism andenvironment. The S-B-R sequence, even if writ-ten only once, is thought of as being repeated overand over again in learning or maintaining behavior.A positive reinforcement or result, R+, is definedas an environmental consequence that makes aninitial behavior, B, more probable when that par-ticular type of S occurs again. An aversive conse-quence or a negative result, R", is one that makesa given B less probable the next time S occurs.

This is equivalent to a feedback formulation,in which S is an input field to the organism, B isthe motor output from the organism back to theenvironment, and R is the "reafferent stimulation,"or change of input field from the environmentwhich gives the changed "error signal" to the or-ganism and closes the feedback loop, as seen inFigure 1.

A frequent objection to the Skinnerian formula-tion is that the definition of reinforcement or of apositive reinforcer is "circular," as in saying, "be-havior is made more probable by positive reinforce-ment." But this is very similar to the useful Dar-winian phrase, "the survival of the fittest," which

Organism

S'/B

1 R

Environment

FIG. 1. Feedback loop where S = input field to theorganism, B = motor output from the organism backto the environment, and R = the change of input fieldfrom the environment which gives the changed "errorsignal" to the organism and closes the feedback loop.

642 • AUGUST 1973 • AMERICAN PSYCHOLOGIST

is likewise a circular definition of fitness. Infact, Skinner seems to think of the evolution ofbehavior in an organism from childhood as in-volving a similar "natural selection" of behaviorsthat "work" and therefore "survive" in a givenstimulus situation, presumably because there areinternal neurophysiological loops of response thatare reinforced while others are eliminated. Hesees the evolution of a repertoire of behaviors inan organism as like the evolution of species bysurvival in an ecosystem. "Reinforcers" are infact defined much better than the Darwinian "fit-ness" of species, and reinforcers such as food,water, sex, petting, praise, and random jackpotsof food or money operate similarly and predictablyover a wide range of organisms.

What we realized when we began to considerour various social traps from this reinforcementpoint of view was that the trap depends on thedifference between the personal or short-term re-inforcements for a given B and the group conse-quences or long-term consequences of that B.Skinner has always emphasized that behavior isshaped more by its rapid consequences, within onesecond or a few seconds, than it is by what happensin five minutes or an hour—or at the end of thequarter when the student receives his grades. Im-mediate reinforcement singles out some particularrecent behavior, while long-run reinforcement isambiguous, not indicating which of thousands ofprevious behavioral acts is responsible for it.

In the Federalist Papers, Alexander Hamiltonlikewise emphasized as a central problem in de-signing a governmental structure the fact thatmen's behavior is more affected by immediate con-siderations of personal advantage than by the long-run public interest.

A social trap occurs, then, when there is an op-position between the highly motivating short-runreward or punishment, Rg

+ or Rs", and the long-run consequences, RL* or RL~. In our notation, atrap then has the following form:

Trap: BS Rs

Conversely, a countertrap or a fence would bewritten as follows:

Fence:(countertrap)

BS Rs- . . Rr,+

Here, the immediate punishment (or its expecta-tion after some experience) tends to block behaviorB, even though there would be a long-run reward.In these formulations, we have a behavioral defi-nition of the exact traditional meaning of the wordstrap and fence.

A similar concept of reversal oj reinforcers canbe applied to individual-group traps, where it isnot a question of shorter and longer times so muchas the fact that the personal reward or punishment,RP+ or Rr-, is in opposition to the collective orgroup advantage or disadvantage, RG

+ or RG~.Again we can have traps or fences depending onwhether the initial personal result is positive ornegative.

Types of Traps

Using these ideas, our group has now studied some40 cases and subcases and examples of varioussorts, where the relation between Rs and RL, orRP and RG, differs in one way or another. In thisarticle, I discuss only the broadest general types.There seem to be three major classes: the one-person traps or self-traps; the group traps of theKitty Genovese type or missing-hero type, whereone person is needed to act for the group; and thegroup traps of the Commons type, where the com-mon pursuit of individual goods leads to collectivebads, because of scarcities, overcrowding, and thelike. There can be both traps and countertraps inall three classes, although only a few of the pos-sible subcategories will be illustrated here.

For clarity and simplicity, the different caseswill be identified here by fairly abstract formula-tions and type anecdotes or mnemonic labels.However, I think in each case the reader will beable to see that these designate the trap aspectof several real social problems.

ONE-PERSON TRAPS

We thought it was important to study the variousone-person traps first, to get their main featuresstraightened out before going on to the grouptraps. The most important subgroup of one-persontraps seems to involve the simple reversal of rein-forcers after a time delay.

Such delayed reversals, where Rs+ changes toRr7, are exemplified in the smoking of cigarettes,where there is both a biochemical reinforcementand perhaps a social reinforcement in the short

AMERICAN PSYCHOLOGIST • AUGUST 1973 • 643

run, but which may lead to lung cancer in the longrun. Similarly with overeating, where there arethe pleasures of the food and perhaps of a mother'sapproval in the short run, but an increased riskof heart attacks in the long run.

Countertraps of this type, with simple delayedreversals, where Rs~ changes to RL

+ in the longrun, are exemplified by the difficulty of saving forChristmas, or for old age, because of the depriva-tion of present pleasures—even though the savingscan eventually lead to a considerable reward, withinterest.

A second subgroup of the one-person traps isthat in which the problem is not simple delay, butrather ignorance of the unexpected or reversedoutcome. The fish swimming into the fish trapdoes not know that he cannot get out. In thelong run, ignorance is as lethal as evil. This is thecase of the man handling a gun who shoots himselfor his friend because he "didn't know it wasloaded."

Another subgroup is that of sliding reinf&rcers.These are rein forcers that change steadily as yougo on repeating a behavior, so that it becomes lessand less rewarding and in fact punishing. Yet youmay go on for a long time with the habit, or youmay keep trying, in the hope that the results willsometime again be as good as they once were.This is one aspect of drug addiction, where theoriginal kick and the fun you had with yourfriends turn into a frightful necessity which youregret for most of your waking hours.

The general public has a similar problem in thedeterioration of old pleasures, such as the taste offood. Bacon these days has a label that reads"artificially smoked"; but it doesn't taste likesmoked bacon to me, and I might never havegotten in the habit of breakfast bacon if it hadtasted like that SO years ago.

Today our global changes are confronting uswith many sliding reinforcers. Once, large fami-lies with more babies were good for survival, andthey were a delight, but now excessive babies haveturned into an expense and have contributed toovercrowding for everyone. At one time, moreconsumption of natural resources and of electricpower gave us consumer goods and liberation, butnow we see them turning into a destruction of ournatural heritage, with pollution and overheating.

THE MISSING HERO

When group profit, RG+, is blocked by RP- for any

personal action, we have the missing-hero trap.A type case is the mattress problem, which is en-tertainingly described in Schelling's (1971) article.Consider the situation, on a summer Sunday eve-ning, when thousands of cars are coming backfrom a Cape Cod weekend on a two-lane road anda mattress falls unnoticed from the top of a stationwagon and lies in the northbound lane. All of thecars behind, being uncertain, go around the mat-tress, waiting for the cars in the southbound laneto go by, and the result is a traffic jam that backsup for miles.

Now who moves the mattress? The answer is,generally, no one. People far back in the line donot know what the trouble is and cannot help.And the drivers close to the mattress are thinkingonly of how to get around it quickly—and afterthey have spent so long in line they are damnedif they will spend another several minutes, per-haps endangering themselves, to stop to move thething. Those who have gone past, of course, nolonger have any incentive for moving it.

In such a situation, it is true that sometimes ahero does come forward. Once, when I told anEnglishman this story, he said, "Hah! That's onlya problem for Americans! If there had been asingle Englishman in that line, he would havegotten out and moved the mattress, because we aretrained in childhood to take leadership in a caselike that." This reminded me of another groupthat also would not have had such a problem—theMormons. I was once at Utah State University ina snowy February, and we went to an under-graduate party in the mountains. The students,who were mostly Mormons, almost automaticallyformed a 14-car caravan up the icy winding road.They kept looking up and down the line to see ifthey were still together, and the whole caravanstopped several times with all the men getting outto push a car that had lost traction or was slidingoff the road. So perhaps a Mormon would alsohave moved Schelling's mattress.

These examples immediately show the roleplayed by moral or ethical training in preventingor getting out of this kind of group trap. Never-theless, the willingness even of people of greatgoodwill to come forward and play the hero insuch a case obviously depends a great deal on the

644 • AUGUST 1973 • AMERICAN PSYCHOLOGIST

level of personal difficulty or danger. We see thisin the reluctance of anyone, either in Sicily orAmerica, to testify against the Mafia. The KittyGenovese case, which belongs in this missing-herocategory, may not indicate so much a lack ofcharacter in Americans as a different perception ofpersonal hazard in "getting involved." We needcontinuing positive reinforcers for brave and in-telligent initiative to help keep up our "character"in cases of this kind: combat ribbons and awardsfor valor on the civilian front, so to speak. (Ifit is not the Star of the Order of Lenin, perhapsit could be the Star of John Lindsay!)

Individual Goods and Collective Bads

The third major category is that of purely collec-tive traps, like the Tragedy of the Commons,where RG" follows only because of the excessivenumber of RP

+ practitioners. The problem cannotbe solved by one or two heroes volunteering notto graze their cows on the commons, although sucha course is frequently advocated by men of good-will. And the problem is not the result of anysingle person doing anything that is unethical orbad, for if the number of persons involved werekept small, one can imagine that the collective goodwould be well served by the sum of all the per-sonal RP

+ rewards. It is when the number is ex-cessive that the difficulty arises.

A famous type problem which can be fitted intothis classification is the Prisoner's Dilemma. Thisis one of the types of two-person non-zero-sumgames which Anatol Rapoport (1966) has studiedso extensively for many years. The type situationis that of two prisoners who have been caught bythe police in some misdemeanor but who are sus-pected of worse crimes. They are held incom-municado from each other and each is questioned.The police offer a pattern of rewards such that ifthey both "talk" or "defect" on each other, theyget the standard sentence for their crime; if they"cooperate" with each other, so that neither talks,they get off lightly for their misdemeanor; but ifone talks and the other does not, the first gets areward and goes free, while the second gets adoubly severe sentence.

In this situation, the payoff matrix is designedby the police so that each man benefits by de-fecting, no matter what his partner does. If hispartner defects, the first gets only the standard

sentence, while if his partner makes the mistakeof cooperating, the first man gets off with a re-ward. Yet if both cooperate with each other bynot talking, they do much better than if they bothdefect. So individual rationality is at odds withcollective rationality.

What do human beings do in such a situation?In our Institute, we have had thousands of suchnon-zero-sum games played by student volunteersfor real money with various payoff matrices.Sometimes pairs or groups of students play againsteach other repeatedly for hundreds of trials, andsometimes they play without knowing it againsta stooge or a computer which is programmed torespond in one regular pattern or another. Gen-erally, in prisoner's dilemma situations, it is foundthat the opposing players tend to lock into eithersteady cooperation or steady conflict with eachother. Which pattern is obtained seems to dependcritically on the outcome—or should we say the"reinforcements"?—of the first few plays. Some-times a pattern of cooperation is quickly experi-enced as mutually profitable and is kept. But ifsuch a pattern is not started early, it seems to bealmost impossible for anyone to continue to cooper-ate when his opponent is continually defectingon him and making money at his expense. It ishard to keep working for RG

+ when the otherparty's behavior keeps turning it into Rp~ for you.

As Rapoport (1971) has emphasized, this di-lemma and these alternative outcomes are remark-ably parallel to some aspects of international rela-tions in the non-zero-sum situations of either mu-tual economic dependence or mutual nuclear threat.The United States and Canada have had locked-incooperation; the United States and Russia havehad 25 years of locked-in hostility and arms es-calation.

Another example of individual goods leading tocollective bads, which can also be fitted mathe-matically into this same classification, is the Sell-A-Dollar game, which was invented by MartinShubik a few years ago. This has some formalresemblance to the Prisoner's Dilemma, but it hasan additional escalation feature. At a dull party,just to make things lively, I offer to auction off adollar. You may laugh at this as absurd, for whyshould I auction it off for less than a dollar, andwhy should anyone pay more? But you agree toplay just to see what will happen.

AMERICAN PSYCHOLOGIST • AUGUST 1973 • 645

However, to keep the game interesting, I makesome simple rules. The first is that bidding startsat a nickel. The second is that bidding must goup by 50 per bid. The third is that biddingmust not go over $50. (Everybody laughs.) Thefourth and last rule is that since you will all tryto get my dollar for so little, I want the two high-est bidders both to pay me their bids, althoughonly the highest bidder will get the dollar. (Andthat is where the kicker is.)

So the bidding starts—and when it gets up toabout 25$, you begin to wonder how high itwill go. In a short time, you know. For thereare two moments of truth in the game. The firstone is when the bidding passes 500, Someone hasbid 500, you have bid 450. And if you go on tobid 550, then this SOB with the dollar (me) willbe getting back more than the dollar. But if youdon't raise, you will lose 450. So you raise.

The second moment of truth is when the biddingpasses $1. The other fellow has bid a dollar,while you have bid 950. And if you raise to $1.05,even if you win, you will lose money. But ifyou don't raise, you will lose a lot more. So youraise.

At this point, of course, I can lean back andwatch you two clobber yourselves to death. Theonly limit on how high the bidding goes dependson how much money you have, or what your wifewill think, or how furious you are with yourself.It is a dangerous game to play, because this seriesof escalating reinforcements and pressures canchange the friendships or the relationships ofeverybody involved, perhaps permanently. Iwould not play it with children, even for pennies.

The only case I know where the bidding hasbeen actually tested was when Layman Allenplayed Sell-A-Dollar with play money, with someof the people in his academic games groups. Inone of his tests, the bidding for the dollar went upto $4 before they tired of it.

Why describe this theoretical game in such de-tail? The reason is that it appears to representin simplified monetary form some of the escalationaspects of such problems as drug addiction. Thefirst little injection with your friends—the firstbid—starts off easy and light, as a game, but thebidding then gets higher and higher until you arelosing more than you are gaining. Yet withevery shot you are getting again and again that

little RP+ whose influence outweighs that great

big long-run RP- you are steadily losing, as wellas the RG- that your family and society are losing,too.

The Sell-A-Dollar game also throws further lighton the escalation aspects of the international armsrace. As Rapoport (1971) has pointed out in hisrecent book The Big Two, the military-industrialcomplexes of the United States and Russia arelike two con men, who are actually working to-gether in "selling a dollar" to our two govern-ments—only the governments are raising their bidsby $10 billion each time, instead of a nickel. Theimmediate reinforcers of the bidding situationcontinue to be reinforcing to each group, therebyproducing and maintaining behaviors of both coun-tries which are extremely damaging to both ofthem in the long run, not only in expense and interror, but finally, all too likely, in annihilation.

I hasten to say that there are ways you canprevent getting into, and can sometimes get out of,the sell-a-dollar traps. One first thing that canbe done is to tell everybody about the game, sothat people will be less likely to get into this kindof trap. A little preeducation and dramatic warn-ing always help. In addition, the process of see-ing the game as a whole—getting a metapicture,so to speak, of the competitive processes and theoutcome—helps prevent a person from becomingquite so entrapped by the immediate reinforcersof each bidding step.

But what is most important is to see that it ispossible to change the character of the game withside agreements and side payments. When thebidding passed 250, if you had said to your op-ponent, "You take the dollar and split the profitwith me," you would both have made money, ifhe had had the sense to do so. The United Statesand Russia made a side agreement of this sortwith the Atmospheric Test Ban Treaty, which hasbeen to the advantage of all of us for 10 years now.

LOCKED-IN ASPECTS OF COLLECTIVE BEHAVIOR

It is worth digressing for a moment to note thelocked-in behavior that is characteristic of manyof these social traps. Immediate small reinforce-ments, or lack of them, lead to self-maintainingor stereotyped behavior in the mattress problem,in the conflict and escalation games, and in manyother social situations.

646 • AUGUST 1973 • AMERICAN PSYCHOLOGIST

Of themselves, of course, locked-in social be-haviors and relations are not inherently either goodor bad. A working society requires thousands ofthem to maintain the behavioral patterns that sup-ply food or goods, just as a biological organism re-quires thousands of locked-in and repetitive enzymecycles to maintain its metabolism. We are usedto such contracts and networks, but what is sur-prising to us is the unexpected locked-in patternsthat appear to arise from the free interactions ofmany independent individuals. And we still needto learn how to produce helpful ones and how toprevent or correct damaging ones in these collectiverelationships.

There are three distinct types of locked-in pat-terns in collective behavior that need much morestudy from a reinforcement point of view to seethe microstructure that gives rise to a kind of socialthermodynamics. The first type is what AdamSmith referred to as the invisible hand of the mar-ketplace. He used this term to emphasize theabsence of any overt or mechanical causal mecha-nism in the stabilization of prices or wages aroundsome median value in a free economic market ofcompeting individuals. A similar invisible handtends to equalize and centralize the political partiesin systems with majority (rather than proportionalrepresentation) elections.

The second type might be called by contrast theinvisible fist, where the competition of numerousindividuals does not produce agreement on a medianvalue, but instead runs away from the median, witheither escalation or elimination past some point ofno return. This happens with Gresham's law ineconomics, where "bad money drives out good."Several of our current crises have this characteristic,as with the escalation of arms races or unrestrainedpollution or the elimination of good railroad service,as we have noted. The urban crisis is made almostunsolvable by multiple complex escalations of thiskind, as slum clearance drives slum dwellers else-where, and the poor are migrating into the city toget welfare money, while the rich are moving tothe suburbs to escape taxes.

The third type of locked-in pattern could becalled the invisible chain. This signifies a loop oftransactional relationships among two or morepeople, forming self-maintaining systems that aresometimes very damaging and very hard to get outof. Married couples frequently get locked intorepetitive disagreements over sex or money or the

temperature of the ropm or whether to go to theshow early or late. Eric Berne (1964) has dis-cussed various locked-in networks of this kind inhis book Games People Play. In his game of "al-coholic," for example, he shows how the alcoholicis trapped in a self-maintaining game with three orfour other people, such as the long-suffering wife,the best friend, and the corner bartender, witheach of their responses reinforcing the others fortheir responses.

Over a lifetime, our originally accidental roles inmany such chains, beneficial and damaging, maycome to create and maintain the responses thatfinally appear as our personal or social "character."The self-maintaining character of federal bureausor of the military-industrial complex come fromlarge-scale invisible chains of the same sort. Acareful analysis of such spontaneous lock-ins couldbe crucial today.

Ways Out

It is to be emphasized that this type of formalanalysis, classification, and explanation of manysocial problems as arising from reversal of rein-forcers and the like is radically different from theusual explanations by moralists and social philos-ophers. For example, it is more immediate andpractical than the explanations of some anthropolo-gists and ethologists who imply that our socialproblems of conflict and disorganization are dueto our evolutionary inheritance of "aggression" or ofa "territorial imperative" (ignoring the fact thatsome societies have orders-of-magnitude of lessaggressiveness or territorial demands than others).It passes by the explanations of therapists whointerpret our personal and social problems as dueto childhood frustrations or the Oedipus complexor to our cultural quest for power or the denial oflove or the death wish. It is more behavioral andrealistic than the countercultural claim that thelocking in of our economic system to technologyand consumerism and the increasing use of powerand resources is due to Bacon or to "Newton'ssingle vision" and the ignoring of Blake's fourfoldvision.

Others have looked for explanations of socialproblems within the individual psyche. Koestler(1968) sees our problems today as basically due toa conflict between the lower instinctive brain andthe recently evolved higher rational brain, and he

AMERICAN PSYCHOLOGIST • AUGUST 1973 • 647

has called for some drug which we could take toharmonize the two. Kenneth B. Clark (1971) hasseriously recommended that antiaggressivenessdrugs be given to presidents and public officials,although he did not discuss who will control theinjections.

These two-level "mind-body" individual ap-proaches to problems are like those of St. Paul.Like the other Judeo-Christian moralists, Paul in-terpreted lack of self-control and conflict andcatastrophe as due to sin, or to the "old Adam"inside:

The good that I would, I do not; the evil that I wouldnot, that I do ... I see another law in my members,warring against the law of my mind. . . . Oh wretched manthat I am, who shall deliver me from the body of thisdeath? . . . with the mind I myself serve the law of God;but with the flesh the law of sin [Romans 7:19-25].

This mind-body dichotomy is what has led to theantisex attitudes and the asceticism and self-flagel-lation of Christianity, in its attempts to control andpunish the lower members and drive out the devil.Niebuhr (1932) is not far from this same traditionin seeing social problems as being due to a kindof collective rather than individual wickedness.

But it is clear that Paul's "good" and "evil"could also be formalized as long-range or socialconsequences, RL+ or RL", of behavior which his"lower members" push him into doing or not doingbecause of the immediate gratification aspects ofRg- or Rs

+. And it is clear that a change in therelationship between Rs and RL to prevent this re-versal of reinforcers can create easy self-control, ora society in which it is "easy to be good" withoutself-flagellation or repression. In spite of all of ourserious problems and traps today, the poor aremore or less fed, the children are taught, and thegarbage is disposed of almost automatically by thereinforcements and feedback mechanisms of oursociety without the Christian effort and charity thatwere once necessary to solve such problems. Wehave learned to convert long-range social goodsinto daily wages, RS+, for social workers, teachers,and garbage collectors. Paul's view of the humancondition turns to punishment and actually blocksthis planned conversion of reinforcers that can giveimproved self-management or the correction ofsocial traps.

SOLUTION BY CHANGEDREINFORCEMENT RELATIONS

In contrast to these usual prescriptions, the reversal-of-reinforcers approach suggests a number of spe-cific changes of reinforcers or policies that can getus out of various social traps and that are alreadyin effective use today for solving one problem oranother. Some of these methods have been in usefor hundreds or thousands of years, and much ofour fashionable despair today comes from a kindof willful blindness to the methods that society de-veloped long ago.

For example, the Tragedy of the Commons isessentially a problem of the allocation of scarce re-sources. And a half-hour's thought will turn up adozen mechanisms that we use every day for dealingwith such problems. In various societies, scarceresources of various kinds may be allocated byforce, by tradition, by inheritance, or by election.When they are to be distributed to many people,they may be distributed by lot, or to the loudestvoices, or by first-come-first-served, or by auction,or by selling tickets, as to a World Series baseballgame. Fish and game commissions are set up, oftenwith the support of the fishermen and huntersthemselves, to sell licenses, set bag limits, and limitthe hunting season so as to maintain the ongoingresource undiminished. Hardin (1968) made outhis New England cattle owners to be a good dealmore stupid than they actually were. The prob-lem is not a problem of thoughtless competition, butrather the problem of setting up a superordinateauthority to handle the reinforcement mechanisms—the tickets and bag limits—for getting out ofthese traps.

In fact, when we look at possible reinforcementchanges, we can make a fairly exhaustive formalismof ways to prevent or get out of various social traps.Five major ways stand out immediately:

1. Change the delay to convert long-range con-sequences into more immediate ones. Or, as Skin-ner (1969) put it, "Bring the consequences to bearon behavior." This is part of what we do whenwe put warning labels on cigarette packages, orwhen, in "deconditioning" methods to stop smoking,we put unpleasant-smelling transparent tape aroundeach cigarette.

On a larger scale, the highways of Indiana andOhio were once jammed and ugly, and the problemseemed hopeless until some social entrepreneurs per-suaded the legislatures to set up toll road corpora-

648 • AUGUST 1973 • AMERICAN PSYCHOLOGIST

tions, which sold bonds and paid the constructioncompanies and the workers to build new highways.The short-range pay and return on invest-ments, Rs

+, was a conversion of the eventual bene-fit, RL

+, that would accrue to the state and all ofthe drivers—who were indeed glad in the end topay the toll for their immediate pleasure, Rs

+, indriving on the improved highway. This method ofsolving the unsolvable highway problem wasmuch easier and more successful than any directattempt to change the old highways by the laboriousmethods of appeals to goodwill or unpaid com-munity spirit or coercion, all of which are full ofRs". Obviously a superordinate authority or cor-poration was a useful intermediate step. The well-established investment mechanism is a powerfullyeffective device for getting over any short-run bar-riers, Rgf, of habit or conflict or complexity be-cause it brings the long-range benefits closer in theform of an RS+.

Education is another mechanism that gives animmediate payoff to the student—at least when itis good education—in the form of attention orgrades or job-training pay or intellectual satisfac-tion, RS+, all to make the large long-run payoffsmore immediately visible and effective. (This isseparate from the question of education in generalas a method of avoiding social traps. It is obvi-ously a good and needed method for anticipating allour social problems, but our modern problems in ahigh-education society demonstrate clearly thateducation is not enough, unless it is combinedwith, or used to design, specific reinforcementmechanisms in each case.)

2. Add counterreinforcers, such as social incen-tives or punishments, to encourage or discourage be-haviors by their immediate Rs

+ or Rg~. This issupposed to be the main function of punitive laws,but they obviously have little preventive effect inmany areas, except in things like traffic control,where the probability of detection is high and pay-ment is relatively quick. The incentives providedby administrative law and contract law are muchmore effective in large-scale problems (as might beexpected from the Skinnerian effectiveness of posi-tive reinforcement), and a combination of taxesand incentives for institutions and corporations israpidly changing such problems as pollution, civilrights, and women's rights in the United Statestoday.

3. Change the nature of the long-run conse-quence, RL'. One way to do this is by new inven-tions. Once upon a time it was a sin to make loveto a girl you were not married to, and God wouldpunish you, both with syphilis and with a babythat would kill her in childbirth. Edward Gibbonnoted the injustice of God in not giving the Romanssyphilis for their immoralities; and the Victoriannovels as well as Hollywood movies until recentlyhave had this theme of necessary punishment forsexual sins (though not other sins). But with theinvention of penicillin to stop syphilis, with anti-septic methods to stop childbirth fever, and witheasy contraceptives to prevent having a baby inthe first place, suddenly it is no longer a sin tomake love. State legislatures are decriminalizingextramarital intercourse, and even many religiousleaders are now emphasizing the sacred value,rather than the sin, of lovemaking.

There are many large-scale social problems whereimproved design and planning is what can changethe nature of the long-range consequences. Today,because of thought and design, social security ismade a law; new cities are designed and built; aninternational monetary system is set up; and ahundred old problems are transformed.

4. Add Rs+ for competing behavior, which will

not lead to the bad long-range consequences. Drinka diet cola with saccharine instead of fatteningsugar; smoke a pipe instead of cigarettes. This isan essential component in the Skinnerian methodsof self-control. To avoid the card games everynight in the dormitory, give yourself goodies forstudying instead: for example, some treat such ascandy or a phone call to your girl friend only afterso many pages of study, or a mark on your chartwhen your study alarm clock rings after every sixhours of work, with 10 of these marks entitling youto have an afternoon in the woods or a dinner outin your favorite restaurant. If one reinforcerdoesn't work, another one may, and so progresscan indeed be made, until the larger reinforcementsof grades or parental approval or the "naturalreinforcers" of your own growing competence beginto maintain regular study habits.

Such methods have been used with considerablesuccess to improve school performance and familyrelations of delinquent children in Tharp andWetzel's (1969) study, Behavioral Modification inthe Natural Environment. Larger social exampleswould include the revitalization of an ailing auto

AMERICAN PSYCHOLOGIST • AUGUST 1973 • 649

assembly plant by new management methods (par-ticipatory reinforcements) (see Guest, 1962).

5. Get outside help in changing the reinforcementpatterns of locked-in loops. This is the main newcomponent of Tharp and Wetzel's (1969) methods.The delinquent child gets reinforced for his be-havior by the attention he gets in being scolded, theexcitement of being chased by the police, or theadmiration of his friends; and the teacher, parents,police, friends, and the child are caught together inan invisible chain of self-maintaining reinforcementtransactions. Tharp and Wetzel tried to find"mediators"—a teacher or adult friend who couldsee the child's daily behavior and give him immedi-ate reinforcers, such as marks in a book for in-creased attention or reading, with so many marksentitling him to extra television viewing or horse-back riding on weekends. As the child's behaviorbegan to change within a few weeks, teacher, police,and parents changed their attitudes, and his friendsbegan to admire him for different things (and werereinforced by the network for their change of valuesalso). At this point, Tharp and Wetzel, as theoutside "therapists," were able to withdraw becausethe system had been flipped to a new self-main-taining mode.

In many of our personal and social traps, we can-not easily see or change the locked-in reinforcernetwork ourselves, and a skilled outsider can helpstart the change to a new pattern. Our "lowermembers," for example, our taste buds, get immedi-ate rewards from the cookies; and it takes ourhigher cortex to examine the situation and to putthe cookies in the refrigerator instead of leavingthem on the table. Likewise, it is easier to savefor Christmas if the local bank, as the outside"therapist," has made a contract with you to de-posit your Christmas savings every month and towrite it in your book. Industrial safety laws, in-surance, and social security deductions all repre-sent outside agencies that we have called in toprotect improvident man.

6. Set up superordinate authority to present en-trapments, to allocate resources, to mediate con-flicts, and to redirect immediate reinforcement pat-terns to more rewarding long-range goals. The or-ganization of fish and game commissions againstthe exhaustion of the game, the Sherman Anti-TrustAct against the escalation of monopolies, and asheriff system with mayors and courts in a westernfrontier town all represent something more than

just an outside therapist. They represent the demo-cratic creation of new superordinate authority ableto manage and correct social traps that were lead-ing to collective bads.

This has happened over and over in humanhistory, as in the creation of the European Com-mon Market; the Special Drawing Rights, the newinternational money managed by an internationalcommission; and the International Whaling Com-mittee, even though it does not have any teeth init at the moment capable of controlling the Rus-sian and Japanese competition for the last remain-ing whales. It may be easier to set up super-ordinate authorities when there are manycompetitors than when there are only two or three,because the special pleading or self-interest of astrong individual can be more easily dealt with bythe rest of the group when they are numerous.Nevertheless, the process is never very easy, andit would be important to make theoretical and his-torical studies to see for what kind of social trapssuperordinate authorities can and should be set upand how it can be done most easily and effectively.

Nested Traps

Finally, it is important to note that there are mixedtraps and, in particular, nested traps that are muchharder to solve than any of the simple traps wehave discussed so far. Traps of this kind includethe locked-in violence of United States communica-tions media, books, and drama; delinquent gangbehavior; and drug and alcohol addiction.

In the United States media, the methods andhabits of violence and violence as a communityexcitement are demonstrated daily and weeklymany orders of magnitude more often than, say,human affection, or daily problem solving, or even(horrors!) sexual love. The media are in theinvisible fist of the competition for more sales orhigher audience ratings, so as to get more adver-tising or profits. (The invisible fist is proven bythe self-accelerating elimination of hundreds ofleading newspapers and magazines over the last 20years.) This media violence locks in, in turn, toa multiplication of violent acts and violent indi-viduals in the community. Headline reports ofhijackings or any other special type of crime pro-duce immediate imitations. In addition, it locksin the children and the older consumers to the self-maintaining idea that this is the only important

650 • AUGUST 1973 • AMERICAN PSYCHOLOGIST

kind of news or drama, and conversely that theviolence of real war is just another television spec-tacular. No cure of these nested lock-ins to violencein our society may be possible without a super-ordinate authority that can change all three of theseself-maintaining loops simultaneously.

Gang behavior is also locked in at several levels.Each member of the gang is reinforced by his gangpartners, as a subsociety, for slashing tires or takingdrugs or escalating to more daring things. In addi-tion, the gang entity is reinforced by the non-zero-sum conflict situation with other gangs or theexcitement of avoiding the police. It is also oftentacitly supported by an adult neighborhood sub-culture that protects the members and applaudstheir daring and sees them as expressing suppressedresentments against the larger society. And finally,the gang, with growing experience, makes contactwith and comes to be supported by, and lockedinto, the larger criminal subculture, which is inturn serving needs and demands of the larger so-ciety. All of these loops would have to be inter-rupted and changed to produce any general changein the gang problem. It may not be possible, ex-cept through an ideological revolution, or a totalchange in city and neighborhood structures andlegal-criminal relationships.

Likewise for drug and alcohol addiction. Thealcoholic, for example, is locked in first by his ownbiochemical need for alcohol. He is also caught inthe invisible chain of reinforcements with familyand friends, as illustrated in Games People Play(Berne, 1964). In addition, he is an importantcog in the network of the corner bar and the com-petitive liquor industry. And finally the liquorindustry itself—producing a certain percentage ofalcoholics—is serving a cultural need which ourcivilization can probably no longer do without, withour social contacts and meetings and most of ourmajor business and government and military de-cisions being lubricated by alcohol. In this nest of

traps, the individual alcoholics come to be liketraffic casualties, which can be reduced in numberbut not eliminated if our society is to continue tofunction in the only way it knows how.

But whatever solutions may eventually be foundin these more complex nested cases, it is clear thatthe approach by analysis of reinforcements and re-inforcement loops offers important new clarifyingexplanations and new tools for any ameliorationthat may be possible. Social traps are not the onlykind of social problems, of course. For example,traffic accidents are not traps, nor are many fightsand conflicts of interest, or business failures wherethere was an expectation of risk from the begin-ning. But the social traps represent all of ourmost intractable and large-scale urban, national,and international problems today. And it seemspossible that the study of social traps from thisreinforcement point of view may be opening thedoor on a whole new discipline that could do morethan almost any other academic study to illuminateand solve these locked-in collective problems.

REFERENCES

BERNE, E. Games people play. New York: Grove, 1964.CLARK, K. B. The pathos of power: A psychological per-

spective. American Psychologist, 1971, 26, 1047-1057.GUEST, R. H. Organizational change: The effect of suc-

cessful leadership. Homewood, 111.: Irwin, 1962.HARDIN, G. The tragedy of the commons. Science, 1968,

162, 1243-1248.KOESTLER, A. The ghost in the machine. New York: Mac-

millan, 1968.NIEBUHR, R. Moral man and immoral society. New

York: Scribner's, 1932.RAPOPORT, A. Two-person game theory: The essential ideas.

Ann Arbor: University of Michigan Press, 1966.RAPOPORT, A. The big two, Soviet-American perceptions

of foreign policy. Indianapolis: Pegasus, 1971.ScHEHtNG, T. The ecology of micromotives. Public

Interest, 1971, No. 25, 61-98.SKINNER, B. F. Contingencies of reinforcement: A theo-

retical analysis. New York: Appleton, 1969.THARP, R., & WETZEL, R. Behavior modification in the

natural environment. New York: Academic Press, 1969.

AMERICAN PSYCHOLOGIST • AUGUST 1973 • 651