Lots of people are pro-choice—and not in the abortion context. They think that it’s just intrinsically good—as an end in and of itself—for one to have more options. However, Gustafsson presents quite a difficult challenge for the entire enterprise—demonstrating a paradox that arises if we take free choice to have intrinsic value, as long as we accept various other plausible axioms. Here, I’ll explain the axioms.
First, if rational people may differ as to which option is the most preferred in an option set, the offered freedom of choice has some intrinsic value.
This is called “The Value of Rational Diversity.”
The basic idea is that if there are lots of things that are roughly equally good, then giving people extra options is better. If all places to live are equally good for people, it’s valuable for people to have more choice over where to live. This seems very obvious if we accept that freedom of choice has intrinsic value.
Second, if an option set is expanded with an option that must be less preferred than the already available options by any rational person, the intrinsic value of the offered freedom of choice does not increase.
This is called “The Insignificance of Dominated Options”
This says that if you give people extra options that are much worse than their existing options, then that doesn’t make them better off. If a person is deciding between two types of jam that are equally good, and you give them the option to instead spread bat feces over their jam, this extra option doesn’t make them better off.
Third, if an option set is expanded, the intrinsic value of the offered freedom of choice does not decrease.
This is called “The Harmlessness of Expansions.” This is straightforward enough—if you expand the options that one has, this doesn’t decrease the intrinsic value of their freedom of choice.
Fourth, if an option set has only one option, it does not offer any intrinsically good freedom of choice.
This is called “The Parity of No-Choice Situations.” This is straightforward too. Freedom of choice is about the value of being able to choose between multiple options—if you only have one choice, then that won’t provide valuable freedom of choice.
fifth, the relation ‘at least as good freedom of choice as’ is transitive.
This is called “The Transitivity of Weakly Better Freedom of Choice.” This one says that if one thing has at least as good freedom of choice as another which has at least as good freedom of choice as a third thing, the first thing has at least as good freedom of choice as the third thing.
To show the paradox, Gustafsson starts by imagining the following three options; buy insurance, skip insurance, get free insurance. One can rationally prefer either buy insurance or skip insurance to the other—they’re on a par—but get free insurance is better than either of them.
(1) The option set {buy insurance, skip insurance, get free insurance} offers intrinsically at least as good freedom of choice as the option set {get free insurance}.
This follows from the harmlessness of expansions.
(2) The option set {buy insurance, skip insurance, get free insurance} does not offer intrinsically better freedom of choice than the option set {get free insurance}.
This follows from the insignificance of dominated options.
(3) The option set {get free insurance} offers intrinsically equally good freedom of choice as the option set {buy insurance, skip insurance, get free insurance}.
This follows from (1) and (2).
(4) The option set {buy insurance, skip insurance, get free insurance} offers intrinsically at least as good freedom of choice as the option set {buy insurance, skip insurance}.
This follows from the harmlessness of expansions.
(5) The option set {buy insurance, skip insurance} offers intrinsically better freedom of choice than the option set {buy insurance}.
This follows from the value of rational diversity.
(6) The option set {get free insurance} offers intrinsically better freedom of choice than the option set {buy insurance}.
This follows from (3), (4), (5), and the Transitivity of Weakly Better Freedom of Choice.
(7) The option set {get free insurance} does not offer intrinsically better freedom of choice than the option set {buy insurance}.
This follows from the Parity of No-Choice Situations.
But 6 and 7 contradict.
The basic insight is this. If you have one great option and two good options that seems to be just as good, freedom of choice wise, as having just the one great option—remember, adding the option to spread rat feces rather than jam on your bread doesn’t benefit you. But it seems that if you have one great option and two good options, that is as valuable in terms of freedom of choice as having just the two good options—after all, adding a third option doesn’t reduce the freedom of choice value. But the two good options is more valuable from the standpoint of freedom of choice as having just one good option, but that good option is just as valuable from the standpoint of freedom of choice as the one great option because all cases where you only have one choice have the same freedom of choice. Thus, if we use > to mean better than from the standpoint of freedom of choice.
one great option = one great option and two good options ≥ two good options > one good option = one great option. Thus, we get the result that one great option is better than itself.
This is a pretty big problem if you think choice has intrinsic value. I’m convinced by it!
This seemed pretty bizarre to me at first with the example of "get free insurance", since I automatically interpret "get free x" as "have the wide choice of paying $y >= 0 for x".
The problem with The Parity of No-Choice Situations becomes much clearer to me when I consider the possibility that I have the choice to say "blip" or "blop" whenever anything consequentially relevant happens to me. That would defuse the paradox, but in a way that seems completely divorced from ethics. I end up unable to understand what "No-Choice" could even mean, other than the fatalist frame in which everything is No-Choice.
Your argument presupposes that choice does not have intrinsic value rather than demonstrate it. Specifically, the "The Insignificance of Dominated Options” remains unsubstantiated. Indeed, the whole point of thinking there is intrinsic value to choice means that something other than the consequence of your option matters.
If I get the option to send my kid to two good schools in addition to having the option of sending them to a great school, that would be better than having the latter option alone because I have more free choice.
I'm not saying this proves that there is intrinsic value to choice, I'm saying you haven't proven that there isn't.