Most moral problems begin with a ready set of available actions from which to choose. Should I lie or shouldn’t I? Should I give money to charity or spend it on my children? And so on. Let us call the set of available actions A. Importantly, one particular “action” is often present in A, namely no action at all. Indeed, many moral problems involve just one “proactive” action, and the question is whether this action should be performed or not.
This approach to relevant actions means that consensual utilitarianism is somewhat of a minimalist approach to morality. There might, in fact, be actions that produce far greater utility than any of the relevant actions, but which are not available to the people involved. There may even exist high-utility actions which the people involved have simply failed to consider (this possibility is considered further in the chapter on consent). Whatever the case may be, we need a way of ranking available actions according to some measure of moral fitness.
Our next step is to establish which beings are relevant to the moral problem at hand. Given that we have established the contentment index (CI) as our moral metric, we will select beings on the basis of this metric. I use the word “beings” deliberately. If CI is to be our metric, we must acknowledge that animals belonging to certain other species also have the capacity for happiness or suffering. There is no a priori reason, within the framework of this discussion, for excluding these animals. Human suffering is not unique simply because we ourselves are human.
Peter Singer, in Practical Ethics, argues that there are two types of being relevant to moral problems: Those that are not aware of their own existence through time and therefore have no goals for the future, and those that are aware of their own existence and do have goals. One particular brand of utilitarianism, namely preference utilitarianism, only regards the latter beings as morally relevant, since only they hold preferences. However, the basic experience of happiness or pleasure seems to me like a more fundamental quality than the ability to hold preferences. After all, both types of being are able to experience pleasure. And while it is certainly true that a fully aware being might experience suffering purely as a result of having its preferences thwarted, this component of suffering would anyway have to be incorporated into the hedonist utilitarian calculation along with other sources of suffering. I therefore prefer the more general, hedonist approach. (I should emphasize that “hedonist” is not meant to evoke the popular notion of debauched pleasure-seeking, but the view that happiness is of paramount importance in ethical matters.)
Our first step in defining relevant beings is to posit that a being is not relevant to the moral problem if her action utility is the same under every action in A, assuming she is fully informed. This is another way of saying that a being is irrelevant if she has no interest in which action is performed – if her state of happiness will not be affected one way or another. So, for instance, if I am trying to decide whether or not to lie to my wife, I need not consider the happiness of the hot dog vendor I pass on the way home, since his happiness will be negligibly affected by my decision, even if he is fully informed of it.
However, if I am trying to decide whether or not to lie to my wife, my wife is a relevant person because, if she were fully informed about my choice, her happiness would be contingent upon it. If she knew I had chosen to lie to her, she would be significantly unhappy, but if she knew I had chosen not to lie to her, she would not. Crucially, this depends on her being fully informed, a process that has requirements of its own. For a start, you cannot be informed about something unless you have the cognitive abilities required for that purpose. Second, you cannot be informed of something if you do not exist. This may seem obvious, but it has important implications that will be explored later in this book.
Can we justify the requirement of being fully informed? Should it be morally permissible for me to lie to my wife on the condition that she never finds out about it? If she never finds out about it, then her happiness is not contingent on my choice to lie or not to lie, and she therefore does not qualify as a relevant person. What about something usually considered more egregious than lying, like adultery? Should it be morally permissible for me to have an affair as long as my wife never finds out? Once again, her happiness would not be contingent on my choice to have an affair or not, because she would never know about my decision one way or another.
These questions all point to a deeper question: What, exactly, do we find morally reprehensible about deceptive behavior? My feeling is that it has to do with autonomy. It is not possible for a person to practice proper autonomy with regards to her own interests if she is not fully informed about those interests, or worse, if she is deliberately misled. Moreover, deception leads to an inability to communicate properly, and is therefore detrimental to almost any social enterprise. Finally, deception leads to instability. It produces conditions which often take great effort to maintain and which, if they collapse, can cause considerable harm.
So, if we can accept that being fully informed is important then we are ready, based on the above discussion, to offer a formal definition of a relevant being.
A relevant being is one who, if fully informed, would have an action utility contingent on the choice of action from the set A of available actions.
We could, technically, consider something inanimate (like a rock) as a relevant being, especially since we’ve adopted the view that CI has a well-defined value of zero for unconscious or non-existent beings. However, inanimate objects will always have a zero CI regardless of the actions being proposed. Their CIs will never be contingent on the choice of action. Furthermore, it is not possible to fully inform inanimate objects. For these reasons, they will never qualify as relevant beings.
Note also that by using action utility, rather than the CI itself, we are making a statement about a being’s overall happiness under each action. A given action utility might, for instance, come in the form of a small increase in happiness over a long time, or a high increase in happiness over a short time.
It is also worth emphasizing that a being may be relevant even if there are no immediate differences to its happiness under any available action. Consider the rather fantastical example provided by commenter Aristarchus at the excellent blog No Forbidden Questions. He asks whether it would be immoral to launch, today, a deadly missile that would orbit the earth inert for several centuries before finally plummeting downward and wreaking destruction on a populous city. Our intuition tells us that this is immoral. Although no one alive today is in danger of being harmed by the missile, members of some future generation are, and it would be as morally reprehensible to harm them as it would be to harm ourselves.
By the above definition of a relevant being, all who will be affected by the missile strike several centuries from now are relevant, even though they do not yet exist. They are relevant because their happiness is contingent on whether the missile is launched or not.
Relevant Time Period
As suggested in the previous chapter, we do not usually need to integrate the CIs of relevant beings over their entire life times (or more) in order to reveal the effects of a given action. We need consider only those time periods over which each being’s CIs are different under different actions. Put differently, we can ignore any period of time over which a being’s CIs are identical under all actions in A. Hence the following formal definition:
Relevant time period
The relevant time period for a given being consists of every time interval over which that being’s CI takes on different values for different actions in A.
This definition is simply another way of saying that the relevant time period for a given being is any interval over which that being is, itself, relevant to the moral problem, by the definition of relevant beings.
If there are n relevant beings in a given moral problem, and m actions, then the number of relevant time periods will be n (one time period per being), while the number of action utilities will be n times m (one for every person-action combination).
In the next chapter, I will apply the above definitions to the development of a calculus for solving moral problems under consensual utilitarianism.
Return to the previous chapter.
Return to the table of contents.