[Merged] Immortality & Bayesian Statistics

[Jabba]

No, because you didn't do anything.

But, if that's your hypothesis, then there is no data at all to support the alternative.

Jay

Jay,
- I agree that I didn't really do anything. I was hoping that there had been some confusion in how I had described my hypotheses, so I tried to re-describe them...
- I would contend that there is data to support the alternative (non-A) -- just little or none that anyone on this forum, besides myself, would accept as valid.

- But then, my estimated prior probability for non-A was only .01 -- and again, it's my estimated numbers (probabilities) that members here would disagree with. I'm thinking that if you guys could accept my numbers as justified, you'd have to accept my conclusion as justified...
- Superficially at least, I think that I am using the theorem appropriately -- it's just that my probability estimates (estimates in general being appropriate for Bayesian inference) are questionable -- and, each person judging my conclusion would have to judge for themselves as to the extent to which my estimated probabilities are justified.

--- Jabba

 

[Agatha]

First, you need to define your terms in unambiguous language. You state that the NR hypothesis is "we each live only one, short, life at most".

The R hypothesis, to which you ascribe a probability of one on a hundred, is... what? Some of us live two lives? Everyone lives a long life? Something else?

Then you need to justify your numbers. Your probability of 0.01 for NR implies that you consider that one in every hundred humans living at this moment will have whatever you mean by NR - whether that's immortality, or two lives, or something else which you've yet to explain. What data have you used to determine the 0.01? Whether you think we will accept it as data at all is moot, unless you actually tell us what it is.

 

[Humots]

Jay,
- I agree that I didn't really do anything. I was hoping that there had been some confusion in how I had described my hypotheses, so I tried to re-describe them...
- I would contend that there is data to support the alternative (non-A) -- just little or none that anyone on this forum, besides myself, would accept as valid.

- But then, my estimated prior probability for non-A was only .01 -- and again, it's my estimated numbers (probabilities) that members here would disagree with. I'm thinking that if you guys could accept my numbers as justified, you'd have to accept my conclusion as justified...
- Superficially at least, I think that I am using the theorem appropriately -- it's just that my probability estimates (estimates in general being appropriate for Bayesian inference) are questionable -- and, each person judging my conclusion would have to judge for themselves as to the extent to which my estimated probabilities are justified.

--- Jabba

If A = "we each live only one, short, life at most", then what is non-"we each live only one, short, life at most"?

Judging from your website, you think non-A involves some form of religious immortality. Is it the Christian Heaven? Is it Reincarnation? Is it the Norse Valhalla? Is it the Greek Hades? Is it Uploading and the Rapture of the Geeks?

Superficially you think you are using the theorem appropriately. What does that mean?

And no, you are not using the theorem appropriately, because the theorem requires precisely defined hypotheses and data as entries, not non-"we each live only one, short, life at most" or "all knowledge" or "me".

Today is Christmas Eve. I am putting this aside for a while. Happy Holidays to all.

 

[Brian-M]

Since I'm new to this thread, rather than waste time reading all 403 posts so far I've just skimmed through it to read Jabba's posts. So I have to ask, did I miss the post where he gives his claimed proof of immortality, or has he just been jerking everyone around for over a month?

Just a couple of observations (although I expect many others have beaten me to it).

First, the bit about suspecting there's something wrong with the deck if the first four cards are all aces.

It's commonplace for people to sort out cards into a non-random order. For example, if someone's been playing solitaire before putting the cards back into the box the first four cards you draw will be Ace, two, three and four. If the deck has been used for a magic act, you might find that the entire pack consists of 52 copies of the same card. An unused deck will also have the cards in a nonrandom order, all neatly sorted by suite. Someone might also sort out the deck in order to check that all the cards are there. Four aces, check. Four kings, check... ect.

A pack of cards with an unknown history has a reasonable chance of being in some kind of intentionally ordered sequence (which why people randomize the deck by shuffling it before they start playing). So if you draw a sequence of cards from an unshuffled deck that resembles an intentionally ordered sequence, it's highly probable that the sequence is intentionally ordered, and therefor perfectly natural so suspect that this is the case.

But if the cards are properly shuffled before playing, and the first four cards drawn happen to be all aces, most people wouldn't assume that there's something wrong with the deck. (At least not until a fifth ace showed up.)

As for how unlikely it was for your specific consciousness to come into existence, so what? It's no more unlikely than any other consciousness that could have come into existence in your stead.

For example, if you use a random number generator to set each bit on a terrabyte hard-drive to a random value, and treat the entire content of the hard drive as a single incredibly huge 8 trillion digit binary number, the chances that the number would have come out to specifically that number are mind-bogglingly tiny. But it's no more unlikely than any other number that could have popped up in it's stead.

But I don't see how this has anything to do with proving immortality.

ETA: And that binary number would contain many, many sections that would appear extremely nonrandom to a human who encounters them on their own.

 

[wollery]

Since I'm new to this thread, rather than waste time reading all 403 posts so far I've just skimmed through it to read Jabba's posts. So I have to ask, did I miss the post where he gives his claimed proof of immortality, or has he just been jerking everyone around for over a month?

No, you haven't missed anything.

 

[Brian-M]

No, you haven't missed anything.

I was afraid of the. But I suppose that nobody can find a flaw in his reasoning if he never presents his reasoning for public scrutiny.

 

[Jabba]

1) First, you need to define your terms in unambiguous language. You state that the NR hypothesis is "we each live only one, short, life at most".
2) The R hypothesis, to which you ascribe a probability of one on a hundred, is... what? Some of us live two lives? Everyone lives a long life? Something else?
3) Then you need to justify your numbers. Your probability of 0.01 for NR implies that you consider that one in every hundred humans living at this moment will have whatever you mean by NR - whether that's immortality, or two lives, or something else which you've yet to explain.
4) What data have you used to determine the 0.01? Whether you think we will accept it as data at all is moot, unless you actually tell us what it is.

Agatha,

- I’ve numbered your questions and comments for easier reference.
- Re #1: I don’t think that either of my hypotheses is ambiguous. I’ll get to “R” next," but can you describe how “NR” is ambiguous?
- Re #2: R is simply the “complement” of NR. In other words, if NR is not true, R HAS to be true. R includes ALL OTHER POSSIBILITIES. If you look up “complement,” be sure that the middle vowel is “e,” and not “i.” I’m using the term in its mathematical or logical sense.
- Re #3: First, the probability I propose for NR is .99 – not .01. So, I’m guessing that “NR” was a typo, and you really meant to address the probability of R (R being what I proposed to have a probability of .01). And, as pointed out in #2, R includes ANY of the other possible hypotheses.
- Re #4: Keep in mind that I’m allowing the “Religious” hypotheses only a 1% prior probability -- whereas, I’m allowing the “Non-Religious” hypothesis a 99% prior probability. I’m accepting that the evidence considered so far greatly favors NR. While I’m not allowing much prior probability of R being true, there are actually numerous TYPES of evidence so far considered that marginally support R: tales of reincarnation, “near death experiences”, “ghosts,” etc. There are other, not so obvious, bits of potential evidence for R that I discuss on one of my websites -- but I won’t bother you with them for now.

- As an overview of my attempt with Bayesian statistics:
- I’m allowing that the real evidence -- so far considered -- regarding human mortality greatly supports the NR hypothesis, and gives it a 99% probability of being true.
- In Bayesian terms, the evidence considered so far is what yields NR’s “prior” probability.
- In my Bayesian analysis, I introduce what I think is a piece of relevant evidence that has not yet been considered (that I know of) that I think greatly effects the NR probability.
- In Bayesian terminology, this resulting probability is NR’s “posterior” probability.
- By my reasoning, NR's posterior probability approaches zero.

--- Jabba

 

[Slowvehicle]

Agatha,

- I’ve numbered your questions and comments for easier reference.
- Re #1: I don’t think that either of my hypotheses is ambiguous. I’ll get to “R” next," but can you describe how “NR” is ambiguous?
- Re #2: R is simply the “complement” of NR. In other words, if NR is not true, R HAS to be true. R includes ALL OTHER POSSIBILITIES. If you look up “complement,” be sure that the middle vowel is “e,” and not “i.” I’m using the term in its mathematical or logical sense.
- Re #3: First, the probability I propose for NR is .99 – not .01. So, I’m guessing that “NR” was a typo, and you really meant to address the probability of R (R being what I proposed to have a probability of .01). And, as pointed out in #2, R includes ANY of the other possible hypotheses.
- Re #4: Keep in mind that I’m allowing the “Religious” hypotheses only a 1% prior probability -- whereas, I’m allowing the “Non-Religious” hypothesis a 99% prior probability. I’m accepting that the evidence considered so far greatly favors NR. While I’m not allowing much prior probability of R being true, there are actually numerous TYPES of evidence so far considered that marginally support R: tales of reincarnation, “near death experiences”, “ghosts,” etc. There are other, not so obvious, bits of potential evidence for R that I discuss on one of my websites -- but I won’t bother you with them for now.

- As an overview of my attempt with Bayesian statistics:
- I’m allowing that the real evidence -- so far considered -- regarding human mortality greatly supports the NR hypothesis, and gives it a 99% probability of being true.
- In Bayesian terms, the evidence considered so far is what yields NR’s “prior” probability.
- In my Bayesian analysis, I introduce what I think is a piece of relevant evidence that has not yet been considered (that I know of) that I think greatly effects the NR probability.
- In Bayesian terminology, this resulting probability is NR’s “posterior” probability.
- By my reasoning, NR's posterior probability approaches zero.

--- Jabba

Rich:

Whether you celebrated today as a holiday or a holy day, I wish you joy and peace.

 

[Jabba]

Slowvehicle,
- Thanks very much. You too.
--- Rich

 

[Agatha]

That was jaw-droppingly rude, Jabba; I admit to making typos (as indeed I did with NR) but I am fully aware of the difference between the words 'complement' and 'compliment'. Kindly do not assume ignorance on the part of those to whom you are are addressing your posts.

What is the piece of evidence that you introduced, which you claim affects the probability of the non-religious hypothesis?

 

[Mashuna]

That was jaw-droppingly rude, Jabba; I admit to making typos (as indeed I did with NR) but I am fully aware of the difference between the words 'complement' and 'compliment'. Kindly do not assume ignorance on the part of those to whom you are are addressing your posts.

What is the piece of evidence that you introduced, which you claim affects the probability of the non-religious hypothesis?

I think that his 'four aces' story is what he considers to be the extra evidence.

 

[Pixel42]

I think that his 'four aces' story is what he considers to be the extra evidence.

Despite several people, including I believe a professor of physics, explaining to him in the simplest possible terms that his argument is a well known fallacy. Explanations which he has rudely ignored.

 

[Mashuna]

Despite several people, including I believe a professor of physics, explaining to him in the simplest possible terms that his argument is a well known fallacy. Explanations which he has rudely ignored.

Quite. Also the problem that in order to correctly use Bayesian statistics, one must be very careful with your premises. Jabba has consistently used sweeping and ill-defined generalisations.

 

[Jabba]

1) That was jaw-droppingly rude, Jabba; I admit to making typos (as indeed I did with NR) but I am fully aware of the difference between the words 'complement' and 'compliment'. Kindly do not assume ignorance on the part of those to whom you are are addressing your posts.

2) What is the piece of evidence that you introduced, which you claim affects the probability of the non-religious hypothesis?

Agatha,

- Re #1: I didn't mean to be rude, but I had explained before that R and NR were complements of each other, and when you asked, "The R hypothesis, to which you ascribe a probability of one on a hundred, is... what? Some of us live two lives? Everyone lives a long life? Something else?", I (reasonably, it seems to me) concluded that you didn't know what I meant by "complement."
- Re #2: My "new" piece of evidence is the set of what I perceive as relevant implications of one's own current existence that I've never seen, or heard, addressed in debates about life after death.

--- Jabba

 

[Slowvehicle]

Agatha,

- Re #1: I didn't mean to be rude, but I had explained before that R and NR were complements of each other, and when you asked, "The R hypothesis, to which you ascribe a probability of one on a hundred, is... what? Some of us live two lives? Everyone lives a long life? Something else?", I (reasonably, it seems to me) concluded that you didn't know what I meant by "complement."
- Re #2: My "new" piece of evidence is the set of what I perceive as relevant implications of one's own current existence that I've never seen, or heard, addressed in debates about life after death.

--- Jabba

Rich:

Agatha understands the meaning of "complimentary". If you do, answer her question. Please clearly define your "R hypothesis", and your "NR hypothesis".

 

[shuttlt]

- Re #2: My "new" piece of evidence is the set of what I perceive as relevant implications of one's own current existence that I've never seen, or heard, addressed in debates about life after death.

Could you possibly expand on this without getting into Bayes? Why is your own current existence important? Surely the "scientific" view is that, had you not existed, somebody/thing else either would or would not have existed in your place and it wouldn't have much mattered to science. Why is your existence and all the other possibilities not happening important? Given the "scientific" viewpoint the probability is 1 that you are in a situation where one thing is there case (i.e. you) and a whole bunch of alternatives therefore do not happen.

 

[Jabba]

1) If A = "we each live only one, short, life at most", then what is non-"we each live only one, short, life at most"?

2) Judging from your website, you think non-A involves some form of religious immortality. Is it the Christian Heaven? Is it Reincarnation? Is it the Norse Valhalla? Is it the Greek Hades? Is it Uploading and the Rapture of the Geeks?

3) Superficially you think you are using the theorem appropriately. What does that mean?

4) And no, you are not using the theorem appropriately, because the theorem requires precisely defined hypotheses and data as entries, not non-"we each live only one, short, life at most" or "all knowledge" or "me".

5) Today is Christmas Eve. I am putting this aside for a while. Happy Holidays to all.

Humots,

- Re #1: Please see my last two responses to Agatha.
- Re #2: I tend to believe in what we humans call "transcendence." It seems to me that an allusion to transcendence is what makes a philosophy a "religion." So, in that sense, I do expect immortality to have a religious "explanation." But then, I never should have used "NR," and "R," to represent the complementary hypotheses, as the complement of us living but one, short, life at most, doesn't have to be religious.
- As a point for further discussion, I don't consider "transcendence" to be "supernatural."
- Re #3: By "superficially," I meant something like "without doing any deep thinking about it."
- Re #4: I think that my hypotheses are quite precisely defined. In regard to "all knowledge," I'm just using the Bayesian interpretation -- and,
in regard to "me," I mean "my own current existence."

- Re #5: Happy Holidays to you also... We're expecting the appropriate snow storm tomorrow.

--- Jabba

 

[Agatha]

Agatha,

- Re #1: I didn't mean to be rude, but I had explained before that R and NR were complements of each other, and when you asked, "The R hypothesis, to which you ascribe a probability of one on a hundred, is... what? Some of us live two lives? Everyone lives a long life? Something else?", I (reasonably, it seems to me) concluded that you didn't know what I meant by "complement."

What I was asking you is not "what does complement mean" but "define your hypotheses".

Your NR hypothesis is not clearly defined, and therefore the complement - the R hypothesis - is not clearly defined either. Humots picked up on this in the post following mine when he asked you to explain your R hypothesis and asked if you were meaning Christian resurrection/heaven, Norse Valhalla, etc, because Bayes' theorem requires tightly defined hypotheses and data, not woolly ideas and guesses. If you are going to use Bayes' Theorem, you need to use it properly.

Your 0.01 probability is one in a hundred - that means you estimate that one in every hundred people on this planet will NOT live "one short life, at most". So do they live long lives? Three lives? Eternal lives? Or no life at all? Define your terms.

- Re #2: My "new" piece of evidence is the set of what I perceive as relevant implications of one's own current existence that I've never seen, or heard, addressed in debates about life after death.

Do you mean the bit about everyone requiring lots of chance happenings for their 'self' to be born? In what way does this relate to anything about life after death?

 

[Humots]

Humots,

- Re #1: Please see my last two responses to Agatha.
- Re #2: I tend to believe in what we humans call "transcendence." It seems to me that an allusion to transcendence is what makes a philosophy a "religion." So, in that sense, I do expect immortality to have a religious "explanation." But then, I never should have used "NR," and "R," to represent the complementary hypotheses, as the complement of us living but one, short, life at most, doesn't have to be religious.
- As a point for further discussion, I don't consider "transcendence" to be "supernatural."
- Re #3: By "superficially," I meant something like "without doing any deep thinking about it."
- Re #4: I think that my hypotheses are quite precisely defined. In regard to "all knowledge," I'm just using the Bayesian interpretation -- and,
in regard to "me," I mean "my own current existence."

- Re #5: Happy Holidays to you also... We're expecting the appropriate snow storm tomorrow.

--- Jabba

- Re #1: I've looked at those responses, and you still haven't addressed what the complement of your NR hypothesis is. I ask again, what is non-A?
- Re #2: So non-A involves a form of religious immortality, but not one of the current world religions? What form of immortality are we talking about: an afterlife, reincarnation, something else?
- Re #3: You are applying Bayes inference without doing any deep thinking about it? Why should we take this superficial approach seriously?
- Re #4: No, your hypotheses are not precisely defined. Again, what is non-A? Also, please point me to some reference defining the Bayesian interpretation of "all knowledge". Finally, "my own current existence" is not a precisely defined hypothesis.
- Re #5: Thank you.

 

[Jabba]

First, you need to define your terms in unambiguous language. You state that the NR hypothesis is "we each live only one, short, life at most".

The R hypothesis, to which you ascribe a probability of one on a hundred, is... what? Some of us live two lives? Everyone lives a long life? Something else?

Then you need to justify your numbers. Your probability of 0.01 for NR implies that you consider that one in every hundred humans living at this moment will have whatever you mean by NR - whether that's immortality, or two lives, or something else which you've yet to explain. What data have you used to determine the 0.01? Whether you think we will accept it as data at all is moot, unless you actually tell us what it is.

- Re #1: I've looked at those responses, and you still haven't addressed what the complement of your NR hypothesis is. I ask again, what is non-A?
- Re #2: So non-A involves a form of religious immortality, but not one of the current world religions? What form of immortality are we talking about: an afterlife, reincarnation, something else?
- Re #3: You are applying Bayes inference without doing any deep thinking about it? Why should we take this superficial approach seriously?
- Re #4: No, your hypotheses are not precisely defined. Again, what is non-A? Also, please point me to some reference defining the Bayesian interpretation of "all knowledge". Finally, "my own current existence" is not a precisely defined hypothesis.
- Re #5: Thank you.

jay,
- my 15 month old grandson is on my lap, so i'm typing with one hand.
- agatha/humots and i are having a hard time communicating. i think that i've answered their questions, but they disagree. Can you help us out.
--- jabba