Chapters 1 & 6
By David Gauthier
In Morals By Agreement, Gauthier puts forth a theory to show why an individual, reasoning from non-moral premises, would accept the constraints of morality on his choices. He defines two types of dispositions: a straightforward maximizer and a constrained maximizer. In prisoners-dilemma type games, the straightforward maximizer chooses the solution that maximizes his individual benefit. The constrained maximizer, if she believes she is interacting with other constrained maximizers, chooses the joint action if, in the case that everyone chooses the joint action, the outcome is fair and the benefit is greater than if every person, including herself, chose the individual action. If people’s dispositions are transparent, the CMs would have a benefit, because they could always achieve the increased benefits of mutual cooperation, while SMs could only achieve the individual benefits. In reality, people are not transparent, but translucent, so the rationality of being a CM increase as the ratio of the probability of all-CM interactions to CM and SM interactions increases, and as the ratio of the benefits from joint action to the benefits of defection increase. In this theory, the ability to distinguish people’s true disposition would be very important.
Chapter 1: Overview of a Theory
Some philosophers have aimed to show that acting morally is in each individual’s best interest. However, if this were the case, there would be no need for morals, as people would do the ‘right thing’ no matter what. Gauthier suggests that rational choice theory – the same theory that is used in economics, game theory, and other social sciences, can be applied to moral theory. This was done to some extent by John Rawls and John Harsanyi, however, they only show ethics as part of the theory of rational behavior. Gauthier aims to prove more than that connection, and to show that “morality can be generated as a rational constraint from the non-moral premises of rational choice. Moral principles are a subset of rational principles for choice.” Gauthier shows why an individual, reasoning from non-moral premises, would accept the constraints of morality on his choices
Gauthier describes two conceptions of rationality. There is the maximizing conception of rationality, in which the rational person seeks the greatest satisfaction of her own interests. You may take into account about other people’s interests, but only to the extent you personally care about their interests. This is a nearly universally accepted definition used in economics and other social sciences. Another idea is the universalistic conception of rationality, which suggests that what makes it rational to satisfy an interest does not depend on whose interest it is, thus the rational person seeks to satisfy all interests. You take into account other people’s interest, simply because they are interests. This definition would make it easy to show a connection between rationality and morality, but you have to defend this conception of rationality, since it isn’t the most widely accepted. Therefore, Gauthier sticks with the more commonly accepted version: the maximizing conception of morality.
An individual first understands what he can and cannot do, but it is not clear how he decides what he may or may not do. Morals by agreement is based on a contractarian rational; it suggests that moral principles are the result of a fully voluntary ex ante agreement among rational persons. Of course, this agreement, taking place before morals exist, is hypothetical, but there it is true that in the real world, people acknowledge a distinction between what they can and cannot do. The problem with the contractarian theory is not the idea of moral agreement, but the problem of moral constraint: it makes sense to have everyone agree not to harm each other, but why does a person actually stop themselves from harming someone after the promise is made? This issue has its roots in the earliest philosophical debates. In Plato’s Republic, Glaucon argues that it seems that being secretly unjust (stealing, lying, etc.) would be more beneficial than acting just, and asks Plato to refute this idea. In his book, Hobbes argues that society must be constrained by a political sovereign. However, Hobbes’ ideas form the conceptual underpinning for Gauthier’s theory. Gauthier theorizes that if social institutions and practices can benefit all, then some set of social arrangements should be acceptable to all as a cooperative venture However, there is an incentive, in some cases, to defect – it may be in your interest to make an agreement, but then break it. (i.e. set up a scheme for everyone to pay taxes, but secretly don’t pay your own taxes). Gauthier aims to use rational choice theory to show how a system of moral constraint might exist.
Chapter 6: Compliance: Maximization Constrained
Hobbes provides the underpinning for Gauthier’s theory, so it is important to review Hobbes’ theory. He argued that in a state of nature, everyone has the right to everything – even to one another’s body – therefore, there can be no security (it is a constant state of war). To get out of this situation, Hobbes’ argues that reason could be used to identify “Laws of Nature.” The first law of nature is that everyone should endeavor for peace, if it is possible. The second law is that a man should be willing, when others are too, to create peace, and he should be willing, for this reason, to give up his right to liberty if others will do the same. If others won’t lay down their right, then he shouldn’t either. The third law, argues that a person who keeps the agreements he has rationally made is just. This agreement-keeping is what allows society to have morals by agreement. Hobbes’ writes that the fool may say that “there is no such thing as justice,” because a person will break the agreement if it is in their best interest. Hobbes argues against this criticism by saying that a person disposed to break their agreements will not be allowed into society, so he won’t get the security of the society. However, Hobbes’ still makes use of the absolute monarch to keep society in line. Gauthier argues that this is a costly solution, and that it would be less costly if, like market forces, society were able to keep agreements on its own just by acting rationally.
Gauthier presents two types of dispositions, a straightforward maximizer and a constrained maximizer. The straightforward maximizer (SM) is a person who seeks to maximize utility given the strategies of those with who he interacts. (This is the normal strategy used in game theory and other rational choice theories.) The constrained maximizer (CM) is a person who seeks in some situations to maximize her utility, given not the strategies but the utilities of those with whom she interacts. She is conditionally disposed to base her actions on a joint strategy as long as the utility she expects were everyone to base his action on joint strategy would be no less than what she would expect were everyone to employ individual strategies. She is willing to cooperate in ways that, if everyone does it, the outcome is beneficial and fair. Someone who actually acts on this conditional disposition should have an expected utility greater than what she would expect were everyone to employ individual strategies. It’s important to note that the CM is not just a person taking a longer view of straightforward maximization – she constrains her options to the situations where everyone cooperates or no one cooperates. The main difference between these two dispositions is seen in prisoner’s dilemma type situations, where cooperation by all is more beneficial than everyone taking an individual strategy, but where defection from mutual cooperation is more beneficial for the individual that defects than is cooperation.
Normally, straightforward maximization by everyone results in everyone getting the basic level of utility, u, from the individual action (as opposed to joint action or defection). This is the normal dominant strategy conclusion in game theory. However, if you assume that people’s disposition is transparent, so that everyone knows whether each person is a straightforward maximizer or a constrained maximizer, the game changes. If people are disposed to constrained maximization, then it is possible for constrained maximizers to get a higher level of utility, u’, by cooperating. It is not possible for the person disposed to straightforward maximization to get any benefit higher than u, because those disposed to constrained maximization would not choose cooperation if they know they are interacting with a straightforward maximizer. However, the assumption of transparency is not very realistic. It may be more realistic to assume that people are translucent – i.e. we can sometimes tell whether a person is disposed to constrained maximization or not.
If people are translucent, then the likelihood that being disposed towards CM is more beneficial than being disposed to SM if the ratio between the probability that an interaction involving CMs will result in cooperation and the probability that an interaction involving CMs and SMs will involve exploitation and defection, is greater than the ratio between the gain from defection and the gain from cooperation. Basically, all else constant, as the proportion of CMs in the population increases, the likelihood of increased benefit of being a CM increases. And, all else constant, as the ratio between the gain from defection and the gain from cooperation decreases, the benefit of being a CM increases.
We earlier assumed that the agreement a CM would agree to need be not only beneficial, but also fair. A broadly compliant CM that would agree to terms that offer her only a small benefit compared to others, would likely only find herself in situations where a small benefit is possible. This would make it less beneficial for this person to be disposed towards CM (since the ratio between the benefit from defection and the benefit from cooperation would be large). Therefore, it is better to be a narrowly compliant CM, and limit cooperation to fair situations.
There are a few important considerations that follow from this theory. The theory doesn’t need to apply to all social institutions. For example, if you’re paying taxes, but believe your tax dollars are being used to increase the chances of nuclear war, then tax evasion may seem rational. However, this does not show that it is irrational to comply with fair and beneficial social arrangements, but instead that it is irrational to allow yourself to be exploited. Also, it is only rational to dispose oneself to constrained maximization if one also is sure to exclude straightforward maximizers from the benefits from cooperation. Otherwise, there is no added benefit to being a CM rather than an SM.
The idea of translucency will have important implications for society. Since CMs want to be recognized by other CMs, it will be difficult to hide their disposition from SMs. Therefore, a CM will rely on a well-developed ability to tell the dispositions of others, so they aren’t exploited by SMs. As CMs and SMs both get better at ‘disposition detection’, the SMs will benefit because they are better able to detect CMs and exploit them, but the SMs will be hurt by others’ ability to detect them. However, the CMs will benefit because they are better able to detect and will benefit from others CMs’ ability to detect them.
In general this theory looks as a ‘maximizing disposition’ rather than a maximizing choice, and that is why it comes to different conclusions than more traditional rational choice theories.