> Imagine that at the beginning of the experiment, beauty knew that were she to sample another room after she was done sampling rooms, it would be red. SIA holds that this is relevantly like the actual situation—she has, from the outset, evidence of an extra red room.
Yes, and it's a problem for SIA. It treats a situation
"Observing a red room when I couldn't have observed a non-red room" - which is what Beauty observes when she awakens in a red room.
exactly the same as
"Observing a red room when I could have observed a non-red room" - which happens when the beauty checks other rooms.
And this is a contradiction of conservation of expected evidence.
//"Observing a red room when I couldn't have observed a non-red room" - which is what Beauty observes when she awakens in a red room.
exactly the same as
"Observing a red room when I could have observed a non-red room" - which happens when the beauty checks other rooms.//
She had to observe a red room for the room that she was currently in!
//And this is a contradiction of conservation of expected evidence.//
No, it's what happens whenever you have evidence from the outset for some proposition. Your expectation of the evidence doesn't change, your expectation of the sample changes. If you have a currently biased sample, you should expect it to look worse over time.
Imagine that there are some number of red and blue jellybeans in a jar. I pull out two and they're both red. I should expect the sample to get less red over time because I think I have a biased sample.
Oh, I see! You are talking about the fact that being awake in the red room means that this particular room can't be sampled anymore when you go and check the other rooms. Yes, this makes sense. Still I don't think it explains SIA's update.
Let's explore more the jelly beans case as a direct non-anthropic analogue of the problem. Suppose there is a bag of 81 jelly beans, some of them red and some of them blue. Consider three different cases:
1. You pick 8 beans from the bag and two of them are red.
2. You pick 9 beans from the bag and three of them are red
3. One red bean is removed from the bag. Then you pick 8 beans and two of them are red.
All three cases give you different evidence about the ration of the beans in the bag.
According to 1. there are about 81/4 red beans which is rounded to 20.
According to 2. there are about 81/3=27 red beans
According to 3. there are about 1+80/4=21 red beans
Okay, I've read the reddit post and it seems that its author doesn't make a mistake of treating the situation as 1. and correctly interprets it as 3.
So the question is why do you think it should be interpreted as 2.? I don't think we have a disagreement about the jelly bean case? Then what's so different about the anthropic variant?
SIA holds that 2 is the right interpretation because you should treat the being who might be created as randomly selected from possible people. The claim is that you should update on your existence in the way described, for all the reasons to support SIA.
I know that SIA claims that. But whether this claim is true or not is the whole question. And in this particular situation it really doesn't seem that this claim is true.
Here SIA demands that I reasoned about my awakening differently from how I'd reasoned about jelly beans and I don't see any reason why. The situations seem completely isomorphic. I do not see how it makes sense even if we accept the existence of souls. What's so different about my awakening in a red room compared to observing a red jelly bean removed from the bag? What other information I possess?
Or are you sayng that we are just supposed to bite the bullet? That this is a counterintuitive result for which you do not have any justification but we have to accept it anyway due to all the good results that SIA produces?
> Imagine that at the beginning of the experiment, beauty knew that were she to sample another room after she was done sampling rooms, it would be red. SIA holds that this is relevantly like the actual situation—she has, from the outset, evidence of an extra red room.
Yes, and it's a problem for SIA. It treats a situation
"Observing a red room when I couldn't have observed a non-red room" - which is what Beauty observes when she awakens in a red room.
exactly the same as
"Observing a red room when I could have observed a non-red room" - which happens when the beauty checks other rooms.
And this is a contradiction of conservation of expected evidence.
//"Observing a red room when I couldn't have observed a non-red room" - which is what Beauty observes when she awakens in a red room.
exactly the same as
"Observing a red room when I could have observed a non-red room" - which happens when the beauty checks other rooms.//
She had to observe a red room for the room that she was currently in!
//And this is a contradiction of conservation of expected evidence.//
No, it's what happens whenever you have evidence from the outset for some proposition. Your expectation of the evidence doesn't change, your expectation of the sample changes. If you have a currently biased sample, you should expect it to look worse over time.
I don't understand what you mean by it. Could you give some non-anthropic example?
Imagine that there are some number of red and blue jellybeans in a jar. I pull out two and they're both red. I should expect the sample to get less red over time because I think I have a biased sample.
Oh, I see! You are talking about the fact that being awake in the red room means that this particular room can't be sampled anymore when you go and check the other rooms. Yes, this makes sense. Still I don't think it explains SIA's update.
Let's explore more the jelly beans case as a direct non-anthropic analogue of the problem. Suppose there is a bag of 81 jelly beans, some of them red and some of them blue. Consider three different cases:
1. You pick 8 beans from the bag and two of them are red.
2. You pick 9 beans from the bag and three of them are red
3. One red bean is removed from the bag. Then you pick 8 beans and two of them are red.
All three cases give you different evidence about the ration of the beans in the bag.
According to 1. there are about 81/4 red beans which is rounded to 20.
According to 2. there are about 81/3=27 red beans
According to 3. there are about 1+80/4=21 red beans
Okay, I've read the reddit post and it seems that its author doesn't make a mistake of treating the situation as 1. and correctly interprets it as 3.
So the question is why do you think it should be interpreted as 2.? I don't think we have a disagreement about the jelly bean case? Then what's so different about the anthropic variant?
SIA holds that 2 is the right interpretation because you should treat the being who might be created as randomly selected from possible people. The claim is that you should update on your existence in the way described, for all the reasons to support SIA.
I know that SIA claims that. But whether this claim is true or not is the whole question. And in this particular situation it really doesn't seem that this claim is true.
Here SIA demands that I reasoned about my awakening differently from how I'd reasoned about jelly beans and I don't see any reason why. The situations seem completely isomorphic. I do not see how it makes sense even if we accept the existence of souls. What's so different about my awakening in a red room compared to observing a red jelly bean removed from the bag? What other information I possess?
Or are you sayng that we are just supposed to bite the bullet? That this is a counterintuitive result for which you do not have any justification but we have to accept it anyway due to all the good results that SIA produces?