The Prisoner’s Dilemma

A Stepping Stone To The Refutation Of Pure Self-Interest And A Guide To Political And Moral Obligation

---

The prisoner’s dilemma is a logical problem game whose invention is attributed to A. W. Tucker.¹ It operates in the following manner: Two prisoners, who cooperated in a crime together, are brought before an attorney general, one after the other so that each is unable to learn of the events that take place with their partner. They are told that if they confess to their crimes (i.e. turn their partner in) they themself will be allowed to walk free while their partner will spend ten years in jail. This is only true however if it is just one of the prisoners who confesses; if both confess then they will each have to spend five years in jail. Finally the third alternative, if neither of the prisoners confesses, results in each spending one year in jail.

Taking the problem from a strictly individualistic and rationally self-interested point of view the solution is obvious: to confess. In this way you, the hypothetical prisoner, are assured of going free which is something that you would want. However the problem is not nearly as simple as that because there is still the factor of the other prisoner to consider. Obviously they would also choose to confess as it would be in their own best interest also. But if both of these situations occur then both of you would be sentenced to five years which is certainly something that you don’t want.

Let’s try analyzing the problem in another fashion. If you do confess then either you will be let off or have to spend five years in jail. If you don’t confess then you’ll either spend one year or ten years in jail. Looked at in this fashion, from the sum-total of the results of confession vs. non-confession, it is apparent that the most logical thing to do is turn your partner in, as the least unpleasant overall results would occur. Instead of grouping the results on a confession/non-confession level they could also be grouped according to simple results in order of preference. Such a grouping would be of the form: no time, one year, five years, and ten years. Obviously the first choice preference is the one that is desired but there is no certain way for either prisoner achieving it because its results rest on what the other prisoner chooses to do. However, to allow for the possibility of such a result a confession must again be given. So, from all of this the best decision that can be made by either prisoner is to confess to the crime, hoping to get off but being assured of at least avoiding the worst case scenario of ten years incarceration.

The problem is set up in such a way that the prisoners are unable to confer with each other, and that the second prisoner is unable to know the decision of the first.² If this were not so then in the latter case the decision process would instantly be resolved for the second prisoner. However, the alteration of the problem to include the opportunity for discussion brings up further possibilities which deserve attention.

It would be unreasonable to think that one of the prisoners (A) would agree to allow the other (B) to confess and not also confess himself. Such an action on the part of prisoner A would not be in accordance with a doctrine of (rational) self-interest, and so far this is the only standpoint that this essay has been concerned with. The only agreement that could rationally be made by the prisoners is for them both to remain silent, since this would be best case scenario for the two of them as a whole.

This solution falls apart though once the prisoners are separated and allowed to make their decisions, since they are once again on their own. There is no reason why either prisoner should want to keep an agreement that they might have made with the other, when it would be in their own best interest to break it. With this realization it is to be observed that the opportunity for discussion can have no weight in the actual decision making process itself³, and that the logical solution of confessing to the crime still remains.

As has been mentioned before this problem has been approached from the view of rational self-interest, and in fact it is with this approach in mind that the problem was created. If this weren’t the case then there would be other solutions, and more readily arrived at ones. One such approach would be from the standpoint of utilitarianism where the first choice would be to do the most good for the greatest number of people and have neither prisoner confess.⁴ Alternatively the confession of the crime may be part of a moral imperative in itself, regardless of the effects of such. In addition another such case would be where one of the prisoners feels that it is their duty to help others as much as possible regardless of the consequences. For such a person, particularly if they know that the other prisoner was motivated by self-interest, the only course of action would be to remain silent, letting the other go free at their own expense.

In any case, since the problem was designed as an exercise in rational self-interest it is best to regard it as such. In light of what has been seen it is apparent that either self-interest is an unproductive motive for behaviour or that its tenets must be altered slightly. The reason for this conclusion is that as it stands, if everyone were to act purely out of the immediate circumstances in such a way to further their goals then no-one would be able to live in as beneficial a way as they could were they to do otherwise. In the prisoner’s dilemma this is exemplified by the fact that if both parties act out of pure self-interest⁵ then they will both be imprisoned for five years rather than for only the one if they had followed other dictates.

An alteration of pure self-interest commensurate with a more productive motive for behaviour would be a consideration of the behaviour of others. In order to create such a proposed system it is necessary that there be some force in existence that binds an individual to keep the agreements that have been made with another individual.⁶ In the case of the prisoner’s dilemma there would then be no question of either party breaking their agreement of silence and so each would only be in jail for one year instead of the five that would occur otherwise.

Certainly one way of ensuring that agreements are kept is if there is some physical or psychological pressure imposed upon those included in the contract such that it is literally impossible for anyone to break an agreement. The classic example of such a form of psychological control can be found in Anthony Burgess’ A Clockwork Orange, but since this is of questionable moral worth (even from the standpoint of self-interest), applicability and feasibility, it would be better to find an alternate solution to the problem.

The actual solution lies in an analytic interpretation of the idea of contracts (and pure self-interest) itself. As Kant applies the Categorical Imperative to the validity of lying, so should a similar train of thought be used as regards pure self-interest. The Categorical Imperative states with respect to lying that such a practice is invalid because if everybody lied then there would be no point in doing so, as the purpose of lying is to make others believe you and this would surely not be the case if everyone always suspected an untruth.

This procedure, while applicable to the moral considerations of certain practices, can also be used to test if an action is logically self-defeating or not. As the analogy goes, if nobody were obliged to keep their promises (obey the contracts that they have made) then there would result a world where people were living in a less beneficial way then if they did keep them. Since the thrust of this argument is centered around the idea of self-interest it is plainly obvious that the conclusion must be that it is not in anybody’s self-interest to act out of pure self-interest, and that contracts with others must be honored in order to preserve the world in a more beneficial state.

Perhaps a more concrete example of the need for contracts would help to show to a greater degree the importance of this concept. In a state of nature man is left to his own devices for his survival and comfort.⁷ In such a state he is also acting according to the impulses of pure self-interest, in that what he does has no connection with anything other than himself. This state is one where everyone must watch out for his neighbors because there is no assurance that they will not do to him what he would do to them. From this state of nature a need for a society develops because of the benefits that it would incur to the individual as they would no longer have to be concerned with their own welfare⁸ through their personal diligence; such things are taken over by the instruments of the government. In entering such a society a contract is agreed to with the sovereign power that will rule the people and it is with this first step that the disadvantages of a pure self-interested motivation are seen and the philosophy discarded.

In following, and maintaining, the system that they have now become a part of they are obligated, through their consent to the contract, to uphold their end of it and obey those instructions that the sovereign gives to them in their own best interest. In turn it is the sovereign’s responsibility to see that he⁹ upholds the part of his contract that proclaims that he will give instructions to the people that are in their own interest.

It is an interesting point that if the prisoner’s dilemma is to be considered in light of the citizen’s duty to the state, they are in fact obligated to confess their crimes. This is apparent by the fact that all citizens in contracting to enter society are agreeing to withhold from those criminal acts that the prisoners are responsible for. Once the prisoners committed their crimes they broke their contract with the state and were placed in an at least temporary¹⁰ state of nature. If they wish to be released from prison, and so re-enter society, they must do so by contracting again and doing everything that they can to preserve it. In doing so they must acknowledge their crimes so that the authorities are able to act in such a way that the wrong done is corrected.¹¹ Ironically the very same argument that says the prisoners should act out of self-interest by keeping their agreement of silence, compels them to confess.¹²

In any consideration of political obligation it is then necessary to look for the terms of agreement in the contract between the individual and the state. If an action were to break this agreement then it is the obligation of the individual to not perform that action. There may be a time in which a question of political duty cannot be answerable by only one solution, and in such cases it is necessary to look for another system of behaviour that would determine what should be done, given that the highest rule for behaviour, political contract, is inapplicable. In such a situation another viable system is that of moral obligation, where certain actions may have no more (discernable) political worth than another, but may be judged as to their worth according to moral principles.

Before going on to describe the principles discussed so far as they relate to morality it should be mentioned that by placing moral obligation under political obligation it is not a proclamation of politics being better than morality as a motivation for action. It is simply that when considering self-interest as it relates to politics, the political aspects must be considered first. Similarly in the following discussion on morality the moral aspects will be considered first, and the political second. While this may seem inconsistent, it is unavoidable. The question as to which system of obligation actually takes precedence (if any) for the self-interested, although important to consider, is beyond the scope of this essay.

Upon first thought it may seem that (rational) self-interest and morality have nothing to do with each other¹³, for morality is generally thought of as relating to metaphysical questions regarding what is right and what is wrong rather than how the individual should act to benefit the most. However, even if there is no motivational connection between the two, there is at least a connection in the similarity of actions that are motivated by each. Additionally the concept of abiding by contracts (keeping agreements, promises) is a focus point of this essay in relation to self-interest, and is dealt with to a great extent by both morality and rational egoism.

Utilitarianism has already been cited in this essay. It deals with the greatest good for the greatest number, and the similarities between it and self-interest are obvious. Perhaps even more obvious is John Stewart Mill’s “Greatest Pleasure Principle”, or that of Existentialism as recognizably first propounded in a revolutionary way by Sartre. “Contractarianism” is also a moral system, and its principles are in fact grounded in rational self-interest.

But the most interesting of moral systems to look at is that of moral justice, where the assumptions of motivation are clearly different from that of self-interest. An act is either wrong or right, not because of its effects on the individual in a specific manner, but because of a universal and metaphysical rule. Putting it in a different way the precepts of moral justice can be exemplified through specific, empirical situations but would be valid regardless of this; whereas self-interest retains its applicability only through such exemplification. Further, although moral justice is a universal concept, applying to all equally, self-interest is not – as one individual’s needs differ from those of another.

The teachings of morality (others than those of moral justice can be included here where applicable) which advise a certain action, or lack of action, are in many ways similar to the religious commands of various faiths. Take as an example, “Thou shalt not kill.” This is a tenet common to all¹⁴ morality and religion.¹⁵ The tenet holds that you should not kill rather than that you cannot. In fact the strength of morality lies arguably in the fact that the moral agent is free to choose between right and wrong.¹⁶

Since there is this free choice of action, the moral decision comes under the heading of an obligation¹⁷ rather than a necessity. As an obligation it can be considered as analogous to the ideas of self- interest and political duty. This determines how we must act¹⁸ if we are to remain moral agents.¹⁹ Notice the distinction between the self- interested doctrine, used in the prisoner’s dilemma (i.e. the agreement to certain action or inaction because you don’t want others to do to you as you would do to them otherwise) and that of moral teaching. The former is agreed upon because of individualistic concerns, whereas the latter (“Do unto others as you would have them do unto you.”) is motivated by a universal code. The moral agent is judging how others would feel if they were treated in a manner disliked by the agent. Yet even though the motives are different the end results are the same so, from a practical standpoint of how people are actually treated, there is no difference.²⁰

The idea of pure self-interest is difficult to adopt even from the viewpoint of the rational egoist, as has been demonstrated, but it is impossible for it to be accepted from a moral viewpoint. Not only do the motives of the two philosophies not match, but neither would the actions (and hence apparent motives).

As the prisoner’s dilemma serves as a groundwork for the refutation of pure self-interest, and as a basis for the need of obligation to contracts, it also demonstrates a further fact which can finally be claimed after the analysis of moral obligation. This is the observation that those members of a society, other than the individual concerned, must be considered, if not as persons in their own right²¹, at least as if they were such. The prisoner’s dilemma is then a very good basis for the understanding of political and moral obligation, as well as a means of understanding social interaction in forms that have not been mentioned here.

---

Endnotes

1. John Rawls, A Theory Of Justice, p. 269.
2. Note also that this problem allows for no sense of individuality on the parts of the participants. Were one prisoner to be aware of a certain philosophical, religious, or moral disposition in the other that would determine their decision then the choice of that prisoner would no longer depend on an unverified assumption of the other’s choice.
3. In the act of discussion one prisoner could of course use emotional pressure on the other to ensure that a certain choice is made, but I look at the problem from a point of view where the only points of consideration come from the solution to the problem itself.
4. With the assumption of course that there are no other factors to consider than those of the problem itself, and that it is in the prisoners’ well-being to have them released.
5. I use this term to mean self-interest that is concerned only with what the individual should do regardless of the effects that others’ decisions might have. In the prisoner’s dilemma the pure self-interested decision is to confess.
6. One individual with another, or with whoever else is involved in the contract.
7. When I say state of nature I borrow from Hobbes, Locke, and Rousseau, and in doing so I adopt their own terminology and mean the word man to apply to all humans.
8. Rousseau would argue that such concerns are not motivated only by self-interest but also by the need to respect others as beings in themselves. Such views are further elucidated in his Discourse On Inequality.
9. The ruling party of the society, whatever form that that might take.
10. Temporary in the sense that their “exile” will last as long as their prison term.
11. An interesting point to consider here is that if it is in the state’s best interest to not keep their promise of releasing the prisoners then they will not do so. Although it would be an empty agreement between the state and the prisoners, it would nonetheless serve its purpose of getting the prisoners to confess.
12. This situation takes place because the contract of the individual to the state takes precedence over any other contract that would violate the former. Only if the former did not exist could the latter be agreed upon. If the prisoner’s dilemma problem concerned two wandering individuals who had contracted only among themselves, and not with the state whose authorities accused them, then the seeming paradox would be resolved and they would be acting quite correctly by keeping their silence.
13. I acknowledge the fact that rational self-interest is a form of morality but for the sake of argument I want to remove it from an analysis of such. Perhaps it would be best to make the statement that rational self-interest has nothing to do any other form of morality.
14. I cannot think of a single instance of moral or religious belief where killing simply for the sake of killing is acceptable. I exclude in this analysis cases of self-defence, mercy-killing, and the like.
15. I am considering religion only as a formal ground for moral teachings. The idea of religion being a forum for the dictates of a supreme, or wiser, being is something that deserves to be looked at separately.
16. I say this with the understanding that there are those who would argue that it is the acts of the individual, and not their will, that determines their moral worth. However, for the sake of argument and because I think that it is the general belief, I wish to consider here only the question of the will.
17. A corollary to this is that the moral agent is under some form of contract with the laws of morality, for without this the idea of an obligation would be weakened if not lost entirely.
18. In accordance with whatever moral code we follow. I’m not going to raise the question of subjective vs. objective moral codes.
19. This raises the question of whether we can give up our status as moral agents because of an immoral act, or whether we must always be considered as capable of moral decisions and responsibility.
20. With this in mind it would be a difficult (if not impossible) task to conduct a sociological survey of the philosophical motivations behind any group interaction. This is so because personal (emotional) feelings are difficult to separate from intellectual beliefs by the individual concerned at the best of times, let alone by an objective observer.
21. As Kant would say, considered as an end rather than merely as a means.