[Merged] Immortality & Bayesian Statistics

[Giordano]

Jabba has convinced me that he cannot possibly exist. This is heightened by his lack of response to corrections and questions. I feel him fading away even now...

 

[Giordano]

- You left out "at most."

Do you think that this would fix your problem or Agatha's objection?

Really?

Do you understand the value of 1/infinity? If so, what probability is lower as you see it?

 

[Jabba]

Firstly Dear wall,

But ONCE CREATED , the possibilities collapse to a *1*.

It is the lotto problem. Each possibilities are mind boggling low. But once the lotto has been
drawn that possibility is ONE.

- P(me|A) is the "likelihood" of my current existence, given that I will exist for one finite time at most -- but, not given that I do currently exist. Sounds strange, but that's how the Bayesian statistics work.

 

[zooterkin]

- You left out "at most."

See if you can work out why.

 

[haibut]

that's how the Bayesian statistics work.

No

 

[Giordano]

- P(me|A) is the "likelihood" of my current existence, given that I will exist for one finite time at most -- but, not given that I do currently exist. Sounds strange, but that's how the Bayesian statistics work.

No, not at all. If you wish to understand HOW they do work, see other people's posts here, or read a statistics book.

 

[Agatha]

- You left out "at most."

I did, because it's redundant. If you contend it is not redundant, explain the functional difference between A=we live one finite life and A=we live one finite life at most, bearing in mind anything other than one finite life is accounted for in ~A.

One life is in A, two lives is in ~A, no lives is in ~A, infinite life is in ~A. Numbers of lives can only be expressed in whole positive numbers or zero.

The probability of 'you' existing before you existed does not depend on whether 'selves' live one finite life or one infinite life or several discrete finite lives or whether we are all dreams of my cat.

 

[Aepervius]

Given Jabba has had his error explained multiple ways without him correcting it, I vote to have the thread closed. Is this possible?

i don't think so. I would vote against it anyway. At best you can advise everybody to leave the thread, but to be honest I am getting fun out of it. Fun i had not felt since the crayon picture of sweet yeti or the flying blimp from ufology.

 

[Aepervius]

- P(me|A) is the "likelihood" of my current existence, given that I will exist for one finite time at most -- but, not given that I do currently exist. Sounds strange, but that's how the Bayesian statistics work.

Very funnily left you convenientely the part where I tell you that since me and A are independent P(Me,A)=P(Me) which is very basic conditional probability.

http://en.wikipedia.org/wiki/Conditional_probability

http://en.wikipedia.org/wiki/Conditional_probability#Statistical_independence

http://en.wikipedia.org/wiki/Independence_(probability_theory)

Again your condition "A" live only ocne, has no impact on the probability of a future human to exists, because its death has no impact on its birth. Causality you know.

Whether we are immortal or not has no impact on whether a given individual in future for a specific birth has a probability of being born.


But hey you compeltly left that out because it makes your wall edifice of card fall out. Instead you quote mined.

 

[godless dave]

- Anyway, here's why I think that P(me|A) is 1/∞.

1) I think that scientifically speaking, when a bit of matter reaches a certain state, it produces life as an emergent property.
2) It could be that all life involves a bit of consciousness, but certainly a certain "level" of life does.
3) Consciousness intrinsically involves a "self."
4) Though, there are probably also "levels" of "selfhood." For instance, dogs are at a different level than are humans.

5) There is no pool of potential selves.
6) Each self is a brand new creation, and there is no theoretical limit to the number of different selves possible.
7) If we're ever able to clone humans, we'll find out that each new production does not reproduce the same self.
8) We can infer such from the animals we have cloned.

Why are you ignoring time and causality?

 

[pakeha]

...
3) Consciousness intrinsically involves a "self."
....

No, Jabba.
Consciousness intrinsically involves a brain.
And neural activity.

 

[jsfisher]

Let X be the event Jabba asks about the prior probability of his own existence.
Let Y be the event Jabba exists.

According to Baye's formula:

P(Y|X) = P(X|Y)*P(Y) / P(X)​

We can expand the demoninator:

P(Y|X) = P(X|Y)*P(Y) / [ P(X|Y)*P(Y) + P(X|~Y)*P(~Y) ]​

Clearly, if Jabba does not exist, then there would be no possibility for Jabba to ask about much of anything, let alone probabilities. So, we can safely assign P(X|~Y) a value of exactly 0.

Substituting:

P(Y|X) = P(X|Y)*P(Y) / [ P(X|Y)*P(Y) + 0*P(~Y) ]

P(Y|X) = P(X|Y)*P(Y) / P(X|Y)*P(Y)

P(Y|X) = 1​

Therefore, the likelihood Jabba exists given he has posed a question about probabilities is exactly 1.

In Jabba's notation, this would be P(me) = 1. Ain't statistics grand?

Continuing with Jabba's notation, no matter what he'd like to provide for wild-ass guesses for probabilities, it must be true that:

P(me|A)*P(A) + P(me|~A)*P(~A) = P(me)​

...and P(me) is exactly 1.

The conclusion is, of course, that Jabba's approach is futile, but I'll leave the final step as an exercise for the interested reader.

 

[Loss Leader]

- Anyway, here's why I think that P(me|A) is 1/∞.

1) I think that scientifically speaking, when a bit of matter reaches a certain state, it produces life as an emergent property.

Nowhere here do you give any sort of numbers that could be used to compute a fractional probability.

2) It could be that all life involves a bit of consciousness, but certainly a certain "level" of life does.

Still, no numbers.

3) Consciousness intrinsically involves a "self."

You've equated two words for purposes of definition, but neither word has a probability attached to it.

4) Though, there are probably also "levels" of "selfhood." For instance, dogs are at a different level than are humans.

Perhaps, but this has nothing to do with the odds of you existing.

5) There is no pool of potential selves.

Because if there were, the chance of you existing would be greater than zero? Because you previously DID say that there was an infinitely divisible consciousness of which we are all a part and now you want to disagree with yourself? Because it's a Sunday in February? What evidence do you have for this?

6) Each self is a brand new creation, and there is no theoretical limit to the number of different selves possible.

Who besides you agrees with this? You need it to be true because you need the total number of possible individuals to be infinite.

7) If we're ever able to clone humans, we'll find out that each new production does not reproduce the same self.

Clearly. Although it might be interesting to point out that earlier in the thread you questioned this. You talked about two individuals sharing the same consciousness and something about a computer controlling two robots.

8) We can infer such from the animals we have cloned.

We can know it from glancing at any pair of twins. This still gives us no numbers from which we can compute a probability.

 

[Slowvehicle]

- You left out "at most."

Good afternoon, Mr. Savage.

Please explain what, in your mind, is the substantive difference between

"...we all live one finite life..."

and

"...we all live one finite life at most...".

Seriously: please do not ignore this post, but, instead, explain the difference as you see it.

For that matter, have you ever heard, or read, any scientist (or, for that matter, any other person at all) echoing your construction?

 

[Slowvehicle]

- P(me|A) is the "likelihood" of my current existence, given that I will exist for one finite time at most -- but, not given that I do currently exist. Sounds strange, but that's how the Bayesian statistics work.

No, that is how Savage statistics work. Bayesian statistics do not use, or consider, so imprecise a term as "likelihood".

Bayesian statistics is not just a show suit for your Texas Sharpshooter act.

 

[John Jones]

- OK. That takes care of two of my numerical entries -- not just P(~A), but also P(A). No one here thinks that it's legitimate to assign a real number to P(~A); and consequently, P(A) has to be 1.00.
- That leaves two to go; and, I think that only one of those is a real issue: P(me|A).

- Remember, I'm here to see what intelligent, well-educated skeptics have to offer by way of evidence and logic against my conclusion. So far, I think that given your arguments against assigning a real number to P(~A), most scientists would agree with me -- i.e., you guys are much to sure of yourselves.
- Consequently, I'm ready to move on to my numerical entry for P(me|A).

- I'll be back.

Name some scientists who agree with you. You don't know what you're talking about.

- Consequently, you haven't reached first base.

 

[Aepervius]

- P(me|A) is the "likelihood" of my current existence, given that I will exist for one finite time at most -- but, not given that I do currently exist. Sounds strange, but that's how the Bayesian statistics work.

Jabba explain in very simple term, how the likelyhood of what happens to a person after it is born, has any influence on the likelyhood of that person being born.

P(person_existence,mortality_or_non_mortality)=P(Person_existence) period.

Ever heard of that things called causality ? It does not matter how we die, or if we do not die, or if we reincarnate. The probability of a specific person (DNA lotto+nurture) being born is the same.

They are independent event (ETA well in so far as a person need to be born to be able to die).

Let me make it clear for you. You have a machine which churn paper. On that paper we mark randomly a single number between 1 and 100. Then we do not know what happens with that piece of paper afterward. Some says they are simply all given to the trash and burnt. Others says they are all put in a museum and kept there conserved for eternity and assign a 0.99% chance to that (changed to remove confusion).

Now we have a singhle piece of paper produced by that machine, with a printed number between 1 and 100. We do not know the number. We only the paper is in our hand and we can open it to look at the number.

Now please tell us what is the probability of the number 55 to be on a paper, given that
all paper are burnt in the trash can afterward. P(55_printed,trash_burnt)

Now please tell us what is the probability of the number 55 to be on a paper, given that all paper are put in a museum afterward for eternity. P(55_printed,museum)

It will be always be 1% the probability of 55 being printed, the probability of any number between 1 and 100 to be pritned, independentely of what happens to that paper afterward.

The reason for that is very simple. conditional probability P(A,B) is the probability of A happening if B is a given. But B in our case and your case happens *after* A.
The paper is always burnt or put in the museum after being printed. Therefore the probability of a specific number being printed is not dependent on its ultimate fate !

This is the same in your case. The probability of being mortal(immortal/reincarnated after death has *nothing to do* with the probability of a specific person beign born.

You are trying to reverse causality and cannot obtain any meaningful result.

 

[Giordano]

- So far, I think that given your arguments against assigning a real number to P(~A), most scientists would agree with me -- i.e., you guys are much to sure of yourselves.

So you will have no trouble finding these "most scientists" who agree with you and providing a reference? Thanks. By the way, 1 is not equal to 1 - a large number that is probably 1. 1 is even not equal to 1 - I haven't a clue as to the exact number.

 

[Jabba]

No, that is how Savage statistics work. Bayesian statistics do not use, or consider, so imprecise a term as "likelihood".

Bayesian statistics is not just a show suit for your Texas Sharpshooter act.

Slowvehicle,
- Actually, the first time that "likelihood" was mentioned in this thread, I reacted the same way -- if I ever knew of a technical usage, I didn't remember it. However, it does have a technical useage. For some reason, I'm not able to paste the URL, but if you look up "likelihood function," you should see that its usage is appropriate.

 

[Giordano]

I begin to really like the lottery illustration. I believe winning the NY Powerball is 1/50,000,000 per drawing yet many specific people have. Is this due to reincarnation of the winners?