Appendices - Killing Happy Animals: Explorations in Utilitarian Ethics - page 156

(Actualism and Presentism are alternative views on whose welfare to
count in the aggregation of welfare that have not been discussed in this
topic.) My own concern here is only with Necessitarianism. Arrhenius
defines the Person-Affecting Restriction as follows:

(a) If outcome A is better than B, then A is better than B for at least one

individual.
If outcome A is better than B for someone but worse for no one, then

(b)

A is better than B.

This is in line with the narrow interpretation of the Person-Affecting
Restriction. I made the distinction between the Wide and the Narrow
Person-Affecting Restriction in Chapter 7.
Arrhenius brings forward only one argument in order to defend the
claim that the Person-Affecting Restriction implies neither Actualism,
nor Necessitarianism, nor Presentism.
The example he uses focuses on Presentism. We are asked to assume
that outcome A consists of the x-people enjoying 10 units of welfare,
outcome B consists of the x-people enjoying 5 units of welfare.
Arrhenius explains that a strict presentist would count A and B as
equally good, or perhaps as incommensurable, if the x-people where
future people. This is correct. As a strict presentist is only concerned with
presently living people, it is not evident how he would judge outcomes in
which only the welfare of future people differs. Thus, this evaluation by
the Presentist would not be in line with the Person-Affecting Restriction.
After all, according to the Person-Affecting Restriction, if outcome A is
better than B for someone but worse for no one, then A is better than B.
However, does this argument show that the Prior Existence View
can be incompatible with the Person-Affecting Restriction? No, this
argument does not apply to the Prior Existence View. If we assume
two different outcomes, in which the x-people enjoy 10 or 5 units
of welfare respectively, we cannot say that the Prior Existence View
would count the outcomes as equally good or as incommensurable if
the x-people were contingent people . After all, if the x-people exist in all

x-people x-people

Outcome A Outcome B

Figure 4 Two outcomes with the x-people