Who is better off if you live and I die? Is one morally obliged to go around impregnating women? Is the repugnant conclusion repugnant? Is secret genocide OK? Does it matter if humanity goes extinct? Why shouldn’t we kill people? Is pity for the dead warranted?
All these discussions come down to the same question often: whether to care about the interests of people who don’t exist but could.
I shan’t directly argue either way; care about whatever you like. I want to show that most of the arguments against caring about the non-existent which repeatedly come up in casual discussion rely on two errors.
Here are common arguments (paraphrased from real discussions):
- There are infinitely many potential people, so caring about them is utterly impractical.
- The utility that a non-existent person experiences is undefined, not zero. You are calculating some amount of utility and attributing it to zero people. This means utility per person is x/0 = undefined.
- Causing a person to not exist is a victimless crime. Stop pretending these people are real just because you imagine them!
- If someone doesn’t exist, they don’t have preferences, so you can’t fulfil them. This includes not caring if they exist or not. The dead do not suffer, only their friends and relatives do that.
- Life alone isn’t worth anything – what matters is what happens in it, so creating a new life is a neutral act.
- You can’t be dead. It’s not something you can be. So you can’t say whether life is better.
- Potential happiness is immeasurable; the person could have been happy, they could have been sad. Their life doesn’t exist, so it doesn’t have characteristics.
- How can you calculate loss of future life? Maybe they’d live another hundred years, if you’re going to imagine they don’t die now.
All of these arguments spring from two misunderstandings:
Thinking of value as being a property of particular circumstances rather than of the comparison between choices of circumstances.
We need never be concerned with the infinite people who don’t exist. All those who won’t exist under any choice we might make are irrelevant. The question is whether those who do exist under one choice we can make and don’t exist under another would be better off existing.
2, 3 and 4 make this mistake too. The utility we are talking about accrues in the possible worlds where the person does exist, and has preferences. Saying someone is worse off not existing is saying that in the worlds where they do exist they have more utility. It is not saying that where they don’t exist they experience suffering, or that they can want to exist when they do not.
Assuming there is nothing to be known about something that isn’t the case.
If someone doesn’t exist, you don’t just not know about their preferences. They actually don’t have any. So how can you say anything about them? If a person died now, how can you say anything about how long they would have lived? How good it could have been? It’s all imaginary. This line of thought underlies arguments 4-8.
But in no case are we discussing characteristics of something that doesn’t exist. We are discussing which characteristics are likely in the case where it does exist. This is very different.
If I haven’t made you a cake, the cake doesn’t have characteristics. To ask whether it is chocolate flavoured is silly. You can still guess that conditional on my making it it is more likely chocolate flavoured than fish flavoured. Whether I’ve made it already is irrelevant. Similarly you can guess that if a child were born it would be more likely to find life positive (as most people seem to) and to like music and food and sex and other things it’s likely to be able to get, and not to have an enourmous unsatisfiable desire for six to be prime. You can guess that conditional on someone’s life continuing, it would probably continue until old age. These are the sorts of things we uncontroversially guess all the time about our own futures, which are of course also conditional on choices we make, so I can’t see why they would become a problem when other potential people are involved.
Are there any good arguments that don’t rely on these errors for wanting to ignore those who don’t currently exist in consequentialist calculations?