On this and subsequent pages, I will describe modifications to contentment utilitarianism that are bound to arise as my thinking develops. I could simply edit my original essay to reflect these changes, but I think there is some value in keeping a record of how my views change over time.
Update 1: Relevant people who do not yet exist
Although never stated explicitly, it is implied in my original essay that each of the relevant people in a given moral problem exists throughout the relevant time interval. However, I have begun to consider moral decisions that may affect the well-being of people who have yet to be born at the time the decision is made.
A rather fantastical example (provided by commenter Aristarchus at No Forbidden Questions) asks whether it is immoral to launch a deadly missile that will orbit the earth, inert, for several centuries before finally plummeting downward and wreaking destruction on a populous city. Our intuition tells us that such an action is obviously immoral: although no one alive today is in danger of being harmed by the missile, some future generation is, and it would be just as morally reprehensible to harm them as it would be to harm ourselves.
When I devised my definition of relevant people, I had no such contingency in mind. Now that it has captured my attention, though, I can see no good reason to exclude it. There is nothing inherently special about an action that occurs now rather than in the future. And there is no good reason to arbitrarily restrict relevant people to those who exist now, especially when we are confident that a relevant action will have an impact on the LOD of a person as yet unborn.
In this update, then, I offer the following revised definitions of relevance*:
Relevant people: Previously, I defined a relevant person as “one whose LOD would be changed by at least one action in X.” While this definition remains unchanged, I now state explicitly that it may include people who have yet to be conceived at the time the moral decision is made. Such people must, however, be expected to have their LODs changed by one of the relevant actions (i.e., we are not considering all non-existent people!).
Relevant desires: It is important that we define the LOD of a non-existent person. This is because the calculus of contentment utilitarianism depends on the comparison of each person’s LOD under different actions. If a person does not exist under one action, but does exist under another, then quantitative comparisons between the LOD changes for those actions are not well-defined.
Assigning such a value is difficult because there is no default, neutral value of LOD: a minimum value of 0 implies total contentment, and there is no well-defined maximum value. We must therefore assume that a non-existent person has, for the purposes of quantitative comparisons, some sort of typical or average LOD. This approach is similar to the use of the “reasonable person” in common law practice. In principle, we could use the mean LOD value in the global population, but in practice such a number would be virtually impossible to obtain, even if we had developed a way of measuring just a single person’s LOD.
Unfortunately, then, some judgement is required whenever considering a person’s LOD changes for cases in which he or she does not exist for one or more relevant action. Luckily, most moral problems involve people who exist for all possible actions, and no “reasonable” LOD for a non-existent person has to be determined.
Relevant time period: Previously, I defined the relevant time period as follows:
There exists an action in X that causes the longest lasting change to any desire in Y. The duration of this change is the relevant time period T.
I now provide the following more precise definition, which takes into account the possibility that no action will have an immediate effect on anyone’s LOD:
The earliest time at which any of the actions in X starts to affect any relevant desire, is denoted T1. Similarly, the latest time at which any of the actions in X still effects any relevant desire, is denoted T2. The relevant time period starts at T1 and ends at T2.
Having made these changes, let us consider some pertinent examples:
1. Delayed missile. Let us reconsider the delayed missile example described above. Here, the relevant actions are to launch the missile or not to launch it. The relevant people are all those who, centuries from now, will feel the effects of the missile strike. It does not matter particularly that we don’t know who exactly these people will be. What we do know is that there will exist people whose LODs will sharply increase centuries from now when the missile strikes (there may be the odd few people whose LODs will decrease, but I think it’s reasonable to assume that a missile strike has a predominantly negative effect on people’s contentment). Relevant desires include, at a minimum, the desire to live a long and happy life, the desire to avoid physical injury, and the desire to see one’s friends and family remain healthy and happy. The relevant time period only starts centuries from now, at the moment the missile strikes, and ends some indeterminate, but finite, amount of time later, when all injuries related to the strike are healed, or perhaps when relatives of lost loved ones are no longer grieving.
Having established relevance, we can look at LODs. If we do not launch the missile, then no changes will be affected to people’s LODs (over and above changes that occur for other reasons, of course). However, if we do launch the missile, then the LODs of the relevant people will overwhelmingly increase. It is quite clear then, in this example, that stopping the missile launch is the recommended action.
2. Should I have a child? There is perhaps no more relevant moral problem concerning future people than whether future people should exist at all. Let us consider this problem as posed by any particular couple who is considering having a child. The relevant actions are to have a child, or not to have a child. The relevant people are those whose LODs would be influenced by having a child. Not having a child is our “do nothing” action, and therefore causes no changes in people’s LODs in addition to those they will experience anyway.
There are many relevant people, with a range of possible LOD changes at stake. First and foremost is the future person whose creation is being considered. Next are the two people considering having the child, and then come other family members and close friends. On their heels are those people more distantly affected, like school teachers and mail deliverers. Finally, there is the rest of the population of the world, whose well being may be influenced in some small way by the creation of another person. There is a myriad of relevant desires involved in this problem, including every desire that the future child will ever experience, since the very existence of these desires depends on a particular action being performed (i.e. the creation of the child).
At this point, let’s start looking at the moral calculus as it pertains to the child, before we look at other people and their desires. The couple considering having the child will have some educated guess about how happy their child is likely to be. For instance, they may live in a rich western country that provides many useful services, tax breaks, etc., for people with children. They may be relatively wealthy, with sound moral principles, and with good intentions to love and look after their child. If this is the case, then it seems likely that the action of having a child will cause a net decrease to the total LOD in this problem, because the LOD of a happy child will be lower than the default average value for a non-existent child. In casual language, the world will be a happier place with the child in it. The remainder of the problem can therefore be solved by asking if the LODs of any other relevant people are likely to increase as a result of the creation of the child and if so, whether these increases will counterbalance the decrease caused by the child’s own LOD.
There are few situations in which a couple of the sort described above would, if they had a child, be likely to produce a net decrease in the LOD of all the relevant people. Friends and family member LODs would surely decrease, while small decreases are likely for people like teachers and mail deliverers whose livelihoods depend on a population of people to whom they can sell their services.
However, there is one highly pertinent exception to this general conclusion, namely overpopulation. It is quite possible that the birth of a child, even if it looks fairly certain to be happy, could inadvertently cause many indirect cases of suffering. The world’s population gets a little bigger, and the various socio-economic strains and imbalances connected to this growth cause large populations of people to suffer just a little more than they might have otherwise. If this suffering looks likely to tip the LOD balance, then the recommended action would be to not have the child.
There is a flip side to this scenario of course, and that is underpopulation. A decreasing population may have subtle negative knock-on effects for many people. In the most extreme hypothetical example of underpopulation, we can think of the last procreating couple on earth. If their children would be assured a good chance at a reasonably happy life, should they go ahead and have those children? Contentment utilitarianism would suggest that yes, they should, because the absence of such children (indeed, the absence of any humans at all) would yield the average LOD of a non-existent person, which is, of course, worse than the existence of a person (or persons) whose integrated LOD is less than average.
In the real world today, overpopulation and underpopulation issues are both at work, depending on where you look. The problem of whether to have a child or not will therefore have different solutions depending on where the parents live.
3. The “My Sister’s Keeper” problem. Jodi Picoult’s riveting novel, My Sister’s Keeper, raises a deeply thought-provoking moral conundrum which I discussed very briefly in a previous post. Imagine that you have a child, and that she develops a very serious health problem that can only be resolved by an organ transplant. Suppose also that the organ – let’s say it’s a kidney – must come from a donor whose genetic make-up is only likely to be found in a close family member. However, no existing family member has such a genetic make-up. The parents are therefore presented with the following option: have another child with the hope that this child will donate her organ to her sibling. The only alternative to this option is to let the first child die. Let us further assume that the parents had no desire, before this problem arose, to have a second child.
The relevant actions here are to have a second child, or to do nothing. The relevant people are the parents, the sick child (let’s call her Sarah), and the future child (Fiona). The relevant desires include the desire, held by both Sarah and her parents, that Sarah live a long and contended life. Fiona, if she is born, will have all the usual desires of the average person (we will not assume that she is likely to be significantly more or less content than average). In addition to these desires, it is possible that she will desire to keep her kidneys, i.e. that she will strongly rebel against her family’s request. There is also a strong possibility that Fiona will develop serious self-confidence issues when she realizes that the sole purpose of her existence is to be an organ donor for her sister. This can be expressed as a strong desire to have been born under different circumstances.
Let’s look at the consequences of each action, starting with the decision to do nothing. If nothing is done, Sarah’s quality of life will slowly deteriorate until she finally passes away, still at a young age. All relevant people’s LODs suffer strong, long-term increases under this scenario (with the obvious exception of Fiona’s who, because she is non-existent, is assumed to have an average LOD).
If the parents decide instead to have Fiona, their LODs will decrease, as will the LOD of Sarah, who will live a long and healthy life. Looking at just the parents and Sarah, then, we can conclude that having Fiona is by far the better choice.
However, Fiona’s LOD in the latter choice may be considerably higher than average, given her possible unwillingness to donate her kidney, and potential issues of self-worth. The question, then, is whether this increase in LOD would balance the substantial decreases experienced by the rest of her family. Does Fiona’s suffering balance the happiness, and indeed life, that she would provide to her family? Such a question is obviously difficult to answer, but it seems to me that having Fiona is more likely to be the more beneficial action for everyone.
There is another problem, though, namely that Fiona has no say in the decision, since she doesn’t exist at the time it is made. In my essay, I specified that no action should be considered if it would raise objections among any of the relevant people. We therefore have to ask the rather odd question: would Fiona object to being born for the sole purpose of providing an organ for her sibling? Without knowing in advance what kind of person Fiona is going to be, it is not possible to answer this question with any certainty.
The best we can do, then, is look at what the majority of people would think in this situation: would most people find repulsive the idea that they were born to save the life of a sibling? If so, then contentment utilitarianism would recommend that Fiona should not be born, since it cannot advocate an action that would raise objections among any relevant people.
* Note that I make no change to the definition of relevant actions.