Over at Felicifia, Alonzo Fyfe, blogger at Atheist Ethicist and The Secularite, and I had a debate on “Which is a better theory of utilitarianism – Alonzo Fyfe’s desirism or Peter Hurford’s two-level utilitarianism?”. I wanted to reprint that debate here. Here’s the first entry (of eight). This entry was written by Alonzo Fyfe.
I want to defend the claim that desirism (desire utilitarianism) is the best form of utilitarianism - better even than Peter Hurford’s version of RM Hare’s two-level theory.
I wish to start by explaining its differences from other utilitarian theories and suggesting how these make desirism stronger than those other theories.
Types of Utilitarianism
There are two ways to cut utilitarian theories.
Types of Utility: On this first cut, we distinguish what counts as utility. Candidates include pleasure and freedom from pain (Jeremy Bentham), happiness over unhappiness (John Stuart Mills), preference satisfaction (Peter Singer), and the well-being of conscious creatures (Sam Harris).
Objects of Evaluation: This second cut looks at primary objects of evaluation. Act utilitarianism makes the evaluation of actions primary - do that act that maximizes utility. Rule utilitarianism makes a rule set primary - do that act that conforms to the best rule set, where the best rule set maximizes utility.
These two cuts go across each other, allowing for possibilities such as happiness maximizing rule utilitarianism (Mill) or preference satisfaction act utilitarianism (Singer).
Desires as the Primary Object of Evaluation
Desirism (desire utilitarianism) holds that desires are the primary object of evaluation. On this account, the right act is the act that conforms to the best desire set, and the best desire set is the desire set that maximizes utility.
Many who heard the term “desire utilitarianism” interpreted it as a candidate for the first cut – the right act is the act that maximizes desire fulfillment. Somebody making this mistake may be tempted to argue: “What about a case where a lot of people desire the torture of a young child? Your theory would say that torturing the child is not only right, but obligatory.”
Desirism would not make the act obligatory. It would look at the desire to have a young child tortured and see it as a desire that tends to reduce utility. Imagine that the desire did not exist – there will be little to no reason to introduce it and many and strong reasons not to. As a desire that tends to thwart other desires, it is does not have a place in the best desire set.
Desirism as Modified Rule Utilitarianism
More specifically, desirism takes rule utilitarianism and physically writes the rules into the structures of brains in the form of desires. In doing this, it avoids the most common objection to rule-utilitarianism - that it collapses into act utilitarianism.
Imagine a situation where an act that follows the rules (saving an innocent person) produces less utility than an act that breaks the rules (killing an innocent person). These are your choices: Perform the act that conforms to the rules, or perform the act that produces the most utility.
In many cases, a rule utilitarian can avoid this objection by pointing to the efficiency of following a rule of thumb. Yet, that response does not work in cases where we know which act will maximize utility - and yet still hold that the moral requirement is to do something other than maximize utility. When the child is tortured for the satisfaction of 10,000 sadists, we know which option maximizes utility.
Considerations of utility alone demand performing the act-utilitarian best act. In order to defend obeying the rules in these circumstances we must introduce a principle that says that following the rules has moral value even when it produces less utility. Requiring that rule-following have value independent of utility abandons utilitarianism.
However, lets write the rules into brains in ways that do not allow this type of choice. A person can forego fulfilling a particular desire, but only if there are more or stronger desires that outweigh it. To perform the act-utilitarian best act, one must have a different desire set - there is no other option. From here, we can ask what is the overall effect of the desire set that would permit the act-utilitarian best act.
Specifically, desirism takes rule utilitarianism and adds three elements:
(1) The rules are not abstract entities - they refer to the potential structures of physical matter (brains). Physical actions can change these structures, and different structures cause different actions. Our question is: Which structures do we have reason to promote?
(2) The rule sets are inviolable. An individual rule/desire may be broken, but only if outweighed by more or stronger rules/desires stacked against it.
(3) The rules/desires are persistent, though changeable. The desires present during any action will motivate an agent’s actions through a wide variety of circumstances. More importantly, the desires we have reason to promote will exist both across agents and across time.
Ethics in Likely versus Exotic/Impossible Situations
In evaluating the utility of a rule/desire set, we will need to take into account how likely particular situations are. A desire that produces non-utilitarian outcomes in rare or impossible situations would not count against it.
This allows us to sidestep many proposed “tests” of utilitarianism. They ask us to imagine an unlikely or impossible situation with an obvious act-utilitarian best act that still seems objectionable. Examples include pushing a fat man onto the path of a runaway trolley thus preventing it from killing five people further down the track, or a doctor killing a healthy patient to harvest five organs to save the lives of five patients.
Desire utilitarianism says that the utility of desires is to be determined by looking at the situations in which they are likely to play a causal role. Morality concerns refusing to drink and drive, keeping promises, paying debts, leaving the money your co-worker has in her purse, refraining from aggressive violence, abstaining from sex without consent, and not killing or maiming people just because you think that some good might come of it.
A look at the news tells us that we have a far bigger problem with people killing others when they think some good can come of it than we have with people not being killed - be they political or religious ends or simply punishing somebody the killer thinks deserves to be killed. Weakening the aversion to killing based on fantastic unrealistic scenarios would be foolish. That is exactly the effect of saying that no condemnation is due the person who kills when he thinks good can come of it.
What Is ‘Utility’?
Desirism not only makes desires the primary object of moral evaluation, it has something to say about what counts as utility.
Every desire that P, for some proposition P, provides an agent with a motivating reason to realize states of affairs where P is true.
The desires for pleasure or happiness provide agents with motivating reasons to realize states where “I am experiencing pleasure” or “I am happy” are true. A desire that one’s children are safe provides a motivating reason to realize states where, “My children are safe” is true. An aversion lying or breaking promises motivates agents to avoid realizing states where, “I am lying” or “I am breaking a promise” are true. An aversion to having any child go hungry provides a motivating reason to realize states where “no child is hungry” is true.
The list of propositions P that define our potential desires is as unbounded as the list of propositions P that define our potential beliefs. This means that the list of things that count as utility is not limited to one or two things. The list may be endless.
Most utilitarian theories that compete against this model fall into one of three groups.
Group 1: Brain-state theories.
Brain-state theories identify utility with particular brain states - pleasure, happiness, contentment.
Brain-state theories fail Robert Nozick’s experience machine test. Ask parents whether they would prefer to be put into an experience machine that will feed them sensations that their child is healthy and happy - while the child suffers great agony in fact - and the vast majority would refuse. They are not seeking a brain state, they are seeking a world state where their child is, in fact, healthy and happy. The experience machine does not create a state where P is true, so the experience machine does not hold value for the parent with a desire that their child be healthy and happy.
Group 2: Value-laden terms
Another family of utilitarian theories identifies utility with some value-laden concept such as “well-being” or “the good life”.
The problem with these theories is that they don’t really answer the question, “What is good?” They simply push the question back to “What is well-being?” or “What is the good life?”
Desirism can actually answer questions like, “What is well-being?” However, understanding well-being means understanding that it is not the sole content of utility.
Desirism identifies well-being is a state of being in which an agent’s self-regarding desires are fulfilled. Because such a state realizes some of the things an agent desires, an agent has reason to realize such a state. However, an agent’s self-regarding desires are only a subset of the agent’s desires. Therefore, well-being is only a subset of what counts as utility.
Many parents would prefer are a greater loss of well-being for themselves to a lesser loss of well-being for their child. This is because their other-regarding concern for their child is something that parents care about in addition to well-being, so that their child’s lesser well-being has more value than their own greater well-being. Such a person would reduce overall well-being, but it does not follow that they are immoral.
Group 3: Intrinsic Value Theories
Intrinsic value theories tell us that some states have intrinsic value, and intrinsic value must be maximized while intrinsic badness minimized. Group 1 and Group 2 are intrinsic value theories.
Desirism holds that intrinsic value does not exist. A desire that P creates a motivating reason to realize P, not its intrinsic merit.
An intrinsic value claim may take the form of saying, for example, all preference satisfaction has equal value, so it is wrong to put one person’s preferences above another’s. One must consider one’s child’s preferences as equal to those of another child - as equal to one’s own.
Desirism can give us reason to deny that a preference for the well-being of one’s family and friends is immoral.
A large company seeking only profits will divide its operations into regions, telling each regional director, “You are responsible for this region. In assigning bonuses, for example, do not look at the benefits of giving bonuses to employees in other regions - that is not your job. Look only to your district.”
Similarly, it makes sense to assign each child to an adult and say, “This child is your responsibility. In accepting this responsibility, you are not to treat the interests of each child equally but to concern himself primarily with the interests of this child - leaving others to care for the children assigned to them.”
This would not be an argument against redistribution. An argument can also be made for taking the surplus resources of some families (regional directors) and investing them in a different region where they can do more good. However, is not inconsistent with the demand that individuals take responsibility for their own children, family, friends, and neighbors.
The value of a desire is not found in its intrinsic merit, but in the reasons that exist for promoting it. In this sense, desires have the same type of value as everything else has - according to its tendency to fulfill or thwart (other) desires.
My goal here has been to explain the differences between desirism and other forms of utilitarianism in order to highlight its strengths. It is a form of rule utilitarianism that does not collapse into act utilitarianism. As such, it can avoid the objections to act utilitarianism based on exotic cases. Its rules are potential physical entities in the real world. It does not require postulating some type of intrinsic value or any other type of exotic entity. It handles Nozick’s experience machine and provides a real-world answer to the question, “What is value?”