Alice: Let’s take population “A+” again. Now imagine that instead of having a population of people with lives barely worth living, the second continent is inhabited by a smaller population with the same very high percentage of resources and utility per person as the population of the first continent. Call it “A++. ” Would you say “A++” was better than “A+?”
Bob: Sure, definitely.
I don’t find this obvious. I also don’t find it obvious that A+ is better than A, or even that some people existing is better than no people existing. My ethical intuitions just don’t seem to give answers for this kind of thing, even on the personal level of trying to answer the question “On a purely selfish basis, is it better for me, personally, to exist or not to exist?” My usual approach of asking myself “Do I anticipate experiencing pleasure or misery from this situation?” doesn’t return an answer that makes any sense, because I can’t experience either pleasure or misery if I don’t exist.
Suppose I define my utility as pleasure / (misery^2). (Misery is worse than pleasure is good.) If I don’t exist, misery is zero, which is wonderful. But my pleasure is also zero, which is terrible. 0⁄0 is undefined, so when I try to calculate the utility of not existing, all I get is an error. That’s the kind of situation I feel like I’m in.
What’s the difference between “On a purely selfish basis, is it better for me, personally, to exist or not to exist?” and “Would I commit suicide, all other things being equal?”?
“Would I commit suicide, all other things being equal?”?
My suicide affects other people. I have both selfish and altruistic desires; “not wanting other people to grieve for me” is a good enough reason not to kill myself.
I read “not existing” as “not ever existing”, so the difference is everything that happened between when you would have started existing and when you would have committed suicide.
(English badly needs separate words for ‘physically exist at a particular time’ and ‘exist, in an abstract timeless sense’. Lots of philosophical discussion such as A-theory vs B-theory would then be shown to be meaningless: does the past exist? Taboo “exist”: the past no longer exists_1, but it exists_2 nevertheless.)
You can tell whether a timeless decision agent would prefer to have been born by giving it opportunities to make decisions that acausally increase its probability of being born.
EDIT: For example, you can convince the agent that it was created because its creator believed that the agent would probably make paperclips. If the TDT agent values its existence, it will make paperclips.
I don’t think a causal decision agent has anything that can be called a “preference to have been born”.
I don’t think my actual utility in real life follows that equation, but it’s an example that has properties that make the example work. (Another analogy would be that the utility of being dead comes out to the square root of minus one, which can’t be directly compared with real numbers.)
I don’t find this obvious. I also don’t find it obvious that A+ is better than A, or even that some people existing is better than no people existing. My ethical intuitions just don’t seem to give answers for this kind of thing, even on the personal level of trying to answer the question “On a purely selfish basis, is it better for me, personally, to exist or not to exist?” My usual approach of asking myself “Do I anticipate experiencing pleasure or misery from this situation?” doesn’t return an answer that makes any sense, because I can’t experience either pleasure or misery if I don’t exist.
Suppose I define my utility as pleasure / (misery^2). (Misery is worse than pleasure is good.) If I don’t exist, misery is zero, which is wonderful. But my pleasure is also zero, which is terrible. 0⁄0 is undefined, so when I try to calculate the utility of not existing, all I get is an error. That’s the kind of situation I feel like I’m in.
What’s the difference between “On a purely selfish basis, is it better for me, personally, to exist or not to exist?” and “Would I commit suicide, all other things being equal?”?
“Would I commit suicide, all other things being equal?”?
My suicide affects other people. I have both selfish and altruistic desires; “not wanting other people to grieve for me” is a good enough reason not to kill myself.
I read “not existing” as “not ever existing”, so the difference is everything that happened between when you would have started existing and when you would have committed suicide.
(English badly needs separate words for ‘physically exist at a particular time’ and ‘exist, in an abstract timeless sense’. Lots of philosophical discussion such as A-theory vs B-theory would then be shown to be meaningless: does the past exist? Taboo “exist”: the past no longer exists_1, but it exists_2 nevertheless.)
You can tell whether a timeless decision agent would prefer to have been born by giving it opportunities to make decisions that acausally increase its probability of being born.
EDIT: For example, you can convince the agent that it was created because its creator believed that the agent would probably make paperclips. If the TDT agent values its existence, it will make paperclips.
I don’t think a causal decision agent has anything that can be called a “preference to have been born”.
So once your misery goes below one unit, you get insane gains in utility for small reductions in misery?
I don’t think my actual utility in real life follows that equation, but it’s an example that has properties that make the example work. (Another analogy would be that the utility of being dead comes out to the square root of minus one, which can’t be directly compared with real numbers.)