are we getting anywhere here =p?
Bayesianism and Sleeping Beauty - Page 11
Blogs > sam!zdat |
Reason
United Kingdom2770 Posts
are we getting anywhere here =p? | ||
sam!zdat
United States5559 Posts
| ||
Reason
United Kingdom2770 Posts
I'm clearly naturally leaning towards the self-indication assumption though I did so before I even knew what it was lol. Do you feel the same way or would you care to play the advocate for the other side so we can explore this issue? As it is people have just thrown out some responses the vast majority of which are similar to mine... | ||
Boblion
France8043 Posts
On January 06 2013 01:34 Kukaracha wrote: Just like that imaginary monkey will plagiarize Shakespeare some day, If an old specie of primates has evolved into humans i don't really know why it would be impossible for them to write books. It would take a while to get to Shakespeare level tho | ||
sam!zdat
United States5559 Posts
edit: but I'm not satisfied by saying this, either. edit: but like I say this is a big part of the reason why I'm skeptical that probability is anything other than a heuristic. edit: I mean that when you talk about probability, you are talking about epistemology and not ontology. Propositions about probability don't tell you anything about the world, they just tell you about your concept of the world. If sleeping beauty always answers 1/2, she'll always be right about the probability even though if she guesses tails she'll be wrong 2/3 of the time. Which goes to show that there's a difference between saying: "X is 1/2 probable" and "If I guess X I will be right 1/2 of the time" Because the fact of the matter changes the conditions under which the utterance is made, and so you get the self-sampling problem. The question is, how many other things are like this? | ||
Reason
United Kingdom2770 Posts
"The question is, how many other things are like this?" I don't know, but I don't think many. As for propositions about probability not telling you anything about the world... well I think it's sort of an issue of time. Probability can't and doesn't claim to be able to tell me what the next flip of the coin will be, it merely assures me and quite correctly so that there's an even chance of it being heads or a tails. Probability can tell you an awful lot about the world, it just can't tell you about the state of the world at a specific point in time. _________________________________________________________ Heads is 1/2 probable. If I flip a coin and guess heads 1/2 the time, I will be right 1/2 the time Your situation has been specifically engineered to show there's a difference between those two statemens and my ^ above situation has been engineered to show that they're the same. She's right about the probability prior to the answering scenario by saying 1/2 but she would be incorrect, I believe, once actually involved in the experiment. So I suppose you could end up with "X is 1/2 probable is not *always* equal to 'If I guess X I will be right 1/2 of the time'" but I don't think that's really what's been established here and thus why there's no accepted solution to this problem. I know there's another similar and much simpler example : There's three cups and one contains the ball. You have to get the ball. You make two guesses at first. Your first guess is wrong, so there's two cups left and one contains the ball. You get offered the choice to swap or stick with your original guess. You always swap, because it improves your chance of getting the right cup from 1/3 to 1/2, even though this seems counter intuitive to say the least. I still don't "understand" it, really. As far as I'm aware there's absolutely no debate about this and it can be proven and tested in a laboratory or in your own home if you're a bit of a nerd, which leaves me confused because it seems identical to this problem yet it seems to be troubling the educated world a bit more than the old cup 'n' ball scenario. Your thoughts? edit: I was aware of the cup 'n' ball scenario prior to reading this thread, so it has obviously heavily influenced my responses and is probably why I'm in favour of the self-indication assumption. | ||
sam!zdat
United States5559 Posts
On January 06 2013 04:40 Reason wrote: Can you explain indexical propositions in general and in the context of the sleeping beauty puzzle, and make an effort to dumb your language down more An indexical is something that points. Words like "now" "here" "I" "you" and so on are indexicals. Normal propositions don't use them. There is no way to formalize the proposition "it is now Monday" in first order logic. "The question is, how many other things are like this?" I don't know, but I don't think many. Maybe not, but what is "like this" exactly? I think it's worth thinking about. As for propositions about probability not telling you anything about the world... well I think it's sort of an issue of time. Probability can't and doesn't claim to be able to tell me what the next flip of the coin will be, it merely assures me and quite correctly so that there's an even chance of it being heads or a tails. Probability can tell you an awful lot about the world, it just can't tell you about the state of the world at a specific point in time. Right, so probability tells you what you don't know about the world. You don't know whether this coin will be heads or tails, and probability lets you say more exactly what it is that you don't know. You use the word "chance" but I don't think you actually know what it means! Can you say what you mean? "Heads is 1/2 probable. if I flip a coin and guess heads 1/2 the time, I will be right 1/2 the time." Your situation has been specifically engineered to show there's a difference between those two statemens and my above situation has been engineered to show that they're the same. So mine wins. In philosophy you engineer thought experiment to show exceptions. Then you have to reconcile your theory with the exception. She's right about the probability prior to the answering scenario by saying 1/2 but she would be incorrect, I believe, once actually involved in the experiment. But she doesn't gain any information, so the probability can't change. So I suppose you could end up with "X is 1/2 probable is not *always* equal to 'If I guess X I will be right 1/2 of the time'" but I don't think that's really what's been established here and thus why there's no accepted solution to this problem. Well, I argue that that is precisely what has been established here, and it's very troubling. My guess is that the second one is not well formed, and we don't really know what we are saying when we say "if I guess X I will be right 1/2 of the time." I know there's another similar and much simpler example : There's three cups and one contains the ball. You have to get the ball. You make two guesses at first. Your first guess is wrong, so there's two cups left and one contains the ball. You get offered the choice to swap or stick with your original guess. You always swap, because it improves your chance of getting the right cup from 1/3 to 1/2, even though this seems counter intuitive. As far as I'm aware there's absolutely no debate about this and it can be proven and tested in a laboratory or in your own home if you're a bit of a nerd, which leaves me confused because it seems identical to this problem yet it seems to be troubling the educated world a bit more than the old cup 'n' ball scenario. Your thoughts? Yeah this is the Monty Hall problem, it's solved. | ||
Reason
United Kingdom2770 Posts
I think once we establish the key difference(s) we can approach this in a more constructive fashion and also perhaps find similar scenarios. When I say chance I'm referring to likelihood or.. probability lol. Probability doesn't tell me if the coin will be heads or tails next, it tells me it's equally likely to be either. That's kind of useful if say you're going for an operation and the doctor says "there's a 2% chance of serious complications and you could die" you don't just say " LIAR, YOU DON'T KNOW WHAT'S GOING TO HAPPEN" you accept the information he has given you, because it is indeed useful. | ||
sam!zdat
United States5559 Posts
On January 06 2013 04:52 Reason wrote: What makes the Sleeping Beauty problem different from the Monty Hall problem aside from rhetorical embellishments? Memory loss. I think once we establish the key difference(s) we can approach this in a more constructive fashion and attempt to find similar scenarios. Good thinking! When I say chance I'm referring to likelihood or.. probability lol. Philosophy 1, Reason 0 That's kind of useful if say you're going for an operation and the doctor says "there's a 2% chance of serious complications and you could die" you don't just say " LIAR, YOU DON'T KNOW WHAT'S GOING TO HAPPEN" you accept the information he has given you, because it is indeed useful. Yeah, it's definitely useful. All models are lies, but many of them are useful. | ||
Reason
United Kingdom2770 Posts
Why are models lies? I respectfully disagree, please elaborate. | ||
sam!zdat
United States5559 Posts
On January 06 2013 04:58 Reason wrote: How does her own memory loss effect the probability of the outcome or the credence that can be attributed to the statement? That's what we've been talking about for all these pages. Why are models lies? Because if there were a model which weren't a lie, it would just be the thing you were trying to model, and it would be pointless. Because you can't define what you mean by 'chance' and I count that as a victory for Philosophy. | ||
Reason
United Kingdom2770 Posts
On January 06 2013 05:04 sam!zdat wrote: That's what we've been talking about for all these pages. Because if there were a model which weren't a lie, it would just be the thing you were trying to model, and it would be pointless. Because you can't define what you mean by 'chance' and I count that as a victory for Philosophy. This reads just like the Monty Hall problem to me =/ I don't understand. I explained what I meant by chance. Likelihood or probability. edit: From what I've read everyone has just treated it like the Monty Hall problem and given you a bunch of mathematical proofs, if that's what you were hoping for then fine. I don't get it and I don't think others do either from their responses. | ||
imallinson
United Kingdom3482 Posts
On January 06 2013 05:04 sam!zdat wrote: Because if there were a model which weren't a lie, it would just be the thing you were trying to model, and it would be pointless. Just because a model is not 100% accurate, which is always the case or it wouldn't be a model, doesn't mean it's a lie, it means it's an approximation. The only case where it would be a lie is if you tried to claim it was 100% accurate or even more accurate than it actually is. | ||
sam!zdat
United States5559 Posts
On January 06 2013 05:30 Reason wrote: I explained what I meant by chance. Likelihood or probability. Those are just synonyms. edit: From what I've read everyone has just treated it like the Monty Hall problem and given you a bunch of mathematical proofs, if that's what you were hoping for then fine. I don't get it and I don't think others do either from their responses. It's like the Monty Hall problem, but much harder. On January 06 2013 05:31 imallinson wrote: Just because a model is not 100% accurate, which is always the case or it wouldn't be a model, doesn't mean it's a lie, it means it's an approximation. The only case where it would be a lie is if you tried to claim it was 100% accurate or even more accurate than it actually is. An approximation is one kind of lie. Another kind of lie is claiming that your approximation is not a lie. | ||
Reason
United Kingdom2770 Posts
On January 06 2013 05:43 sam!zdat wrote: Those are just synonyms. It's like the Monty Hall problem, but much harder. An approximation is one kind of lie. Another kind of lie is claiming that your approximation is not a lie. So by giving you synonyms don't you think I know what I mean by something? I don't know why you'd randomly ask me to define a word anyway... I'm going to read this entire thread again and then comment tomorrow but from what I've seen I don't think we're really getting at the root of the problem here. I'm still of the opinion that it's just the Monty Hall problem re-written and I know that's not true, but unfortunately "memory loss" isn't a sufficient enough explanation for me. I'll try to figure this out myself I guess... might be useful to quote the Monty Hall problem in the OP for new readers and explain specifically what's different about this and why it's not been solved, quote the opposing positions and give a bit of background etc... meh. | ||
hypercube
Hungary2735 Posts
On January 06 2013 04:58 Reason wrote: How does her own memory loss effect the probability of the outcome or the credence that can be attributed to the statement? It's not the memory loss that changes the probability. It's the fact that the mad philosopher changes the distribution between the two questions. It's equivalent to putting 2 red and 2 blue balls into a box and asking what the probability of picking red is. Then taking out a red one and asking for the probability of picking out a red one again. The whole discussion about memory loss is a red herring and samizdat already admitted as much. Even on Sunday Sleeping Beauty knows that her answer will change by time she wakes up. | ||
sam!zdat
United States5559 Posts
On January 06 2013 05:49 Reason wrote: So by giving you synonyms don't you think I know what I mean by something? No... absolutely not... In fact giving synonyms is a sure proof that you DON'T know what you mean... I thought by the way you said it that you understood this. I don't know why you'd randomly ask me to define a word anyway... It's not random, it's the very crux of the issue!! (edit: lololol) I'm going to read this entire thread again and then comment tomorrow but from what I've seen I don't think we're really getting at the root of the problem here. Probably not, I admire your perseverance <3 I'm still of the opinion that it's just the Monty Hall problem re-written My appeal to authority says no. Nobody thinks this is identical to Monty Hall problem, afaik. and I know that's not true, but unfortunately "memory loss" isn't a sufficient enough explanation for me. Yeah it's not, it's just pointing. @hypercube it's definitely not a red herring, the problem is that she KNOWS she will lose her memory. If she didn't know it would be totally unproblematic. I'll try to figure this out myself I guess... might be useful to quote the Monty Hall problem in the OP for new readers and explain specifically what's different about this and why it's not been solved, quote the opposing positions and give a bit of background etc... meh. If you want to write up the monty hall problem, I will edit it into the OP. On January 06 2013 05:50 hypercube wrote: It's the fact that the mad philosopher changes the distribution between the two questions. No, he asks one question which is equivocal. | ||
Reason
United Kingdom2770 Posts
When I said the "chances" of something happening, what I meant was this: x = the number of times a particular event with multiple outcomes is repeated y = the number of times that a particular outcome will be the end result of x repetitions of the event, on average c = "the chances" (of the particular outcome in question happening in a single repetition of the event) y/x = c Was that really necessary lol? | ||
Boblion
France8043 Posts
c = x/y >= 1 doesnt really make sense. More like y/x = c ? | ||
Reason
United Kingdom2770 Posts
| ||
| ||