I think there are conceivable situations where you can’t easily just ask whether or not you should create the child first without looking at each option with the child, because how exactly you create them might matter for their welfare or the welfare of others, e.g. you can imagine choosing between a risky procedure and a safe procedure (for the child’s welfare, not whether or not they will be born) for implantation for an already fertilized egg or in vitro fertilization with an already chosen sperm and egg pair. Maybe the risky one is cheaper, which would be one kind of benefit for the parents.
To run through an example, how would you handle the benign addition argument for the repugnant conclusion (assuming world A’s population is in both world A+ and world Z, and the populations in world A+ and world Z are identical)? You could imagine the above example of in vitro fertilization being structurally similar, just an extra population of size 1 instead of 99, and smaller differences in welfare.
Figure 1 for the benign addition argument from Huemer, In defense of repugnance
You can pick a pairwise comparison to rule out any option on a person-affecting view, since it looks like A < A +, A+ < Z, and Z < A. Maybe all three options should be permissible?
Or maybe something like Dasgupta’s approach? It has 2 steps:
Select the best available option for each possible population. This doesn’t require any stance on extra people or identity.
For one kind of wide person-affecting view, do this instead for each possible population size, rather than each population.
Choose between the best options from step 1. There are multiple ways to do this, and there may be multiple permissible options due to incomparability.
Give only weight to necessary people, who are common to all options, or all of the best options from 1. This seems closest to necessitarianism and what you’re suggesting.
Give more weight to necessary people than extra people. I think this is Dasgupta’s original approach.
Give only weight to necessary people and badly off people (equal or unequal weight). This captures the procreation asymmetry.
For a more natural kind of wide person-affecting view, only the identities of the necessary people should matter, whereas the identities of the extra people do not.
Applying this to the benign addition argument, if worlds A+ and Z have the same populations, then A+ would be ruled out in step 1, A would be chosen, and we’d avoid the repugnant conclusion. If the extra people (compared to A) are completely different between A+ and Z, and identity matters (not using any wide modifications), then no option is ruled out at step 1, and the necessitarian approach (2.1.) would lead to A+.
There are also presumably different ways to handle uncertainty. While many of your decisions may affect who will exist in the future, the probabilities that a given individual who hasn’t been conceived will exist in each outcome might still be positive in each option, and you can still compare the welfare of these probabilistic people, e.g.:
A exists with probability 2%.
A exists with probability 1%, but is expected to be better off than in 1, conditionally on existing in each. (Or expected to be worse off.)
We might also add that 1 would actually resolve with A existing if and only if 2 would actually resolve with A not existing.
I think there are conceivable situations where you can’t easily just ask whether or not you should create the child first without looking at each option with the child, because how exactly you create them might matter for their welfare or the welfare of others, e.g. you can imagine choosing between a risky procedure and a safe procedure (for the child’s welfare, not whether or not they will be born) for implantation for an already fertilized egg or in vitro fertilization with an already chosen sperm and egg pair. Maybe the risky one is cheaper, which would be one kind of benefit for the parents.
To run through an example, how would you handle the benign addition argument for the repugnant conclusion (assuming world A’s population is in both world A+ and world Z, and the populations in world A+ and world Z are identical)? You could imagine the above example of in vitro fertilization being structurally similar, just an extra population of size 1 instead of 99, and smaller differences in welfare.
You can pick a pairwise comparison to rule out any option on a person-affecting view, since it looks like A < A +, A+ < Z, and Z < A. Maybe all three options should be permissible?
Or maybe something like Dasgupta’s approach? It has 2 steps:
Select the best available option for each possible population. This doesn’t require any stance on extra people or identity.
For one kind of wide person-affecting view, do this instead for each possible population size, rather than each population.
Choose between the best options from step 1. There are multiple ways to do this, and there may be multiple permissible options due to incomparability.
Give only weight to necessary people, who are common to all options, or all of the best options from 1. This seems closest to necessitarianism and what you’re suggesting.
Give more weight to necessary people than extra people. I think this is Dasgupta’s original approach.
Give only weight to necessary people and badly off people (equal or unequal weight). This captures the procreation asymmetry.
For a more natural kind of wide person-affecting view, only the identities of the necessary people should matter, whereas the identities of the extra people do not.
Applying this to the benign addition argument, if worlds A+ and Z have the same populations, then A+ would be ruled out in step 1, A would be chosen, and we’d avoid the repugnant conclusion. If the extra people (compared to A) are completely different between A+ and Z, and identity matters (not using any wide modifications), then no option is ruled out at step 1, and the necessitarian approach (2.1.) would lead to A+.
For Dasgupta’s approach, see:
http://users.ox.ac.uk/~sfop0060/pdf/Welfare%20economics%20of%20population.pdf
https://philpapers.org/rec/DASSAF-2
There are also presumably different ways to handle uncertainty. While many of your decisions may affect who will exist in the future, the probabilities that a given individual who hasn’t been conceived will exist in each outcome might still be positive in each option, and you can still compare the welfare of these probabilistic people, e.g.:
A exists with probability 2%.
A exists with probability 1%, but is expected to be better off than in 1, conditionally on existing in each. (Or expected to be worse off.)
We might also add that 1 would actually resolve with A existing if and only if 2 would actually resolve with A not existing.