Immortality & Bayesian Statistics

[Dason]

- I'll quit bothering you guys.

Just FYI - the thing we really want is for you to learn from your mistakes. You still have misconceptions and we've tried to point them out. If you take a step back and actually admit to yourself that maybe you are doing the math wrong and learn from the things we try to tell you then the discussion will be much better. You can't just ask us to look over your material and then ignore us repeatedly when we tell you where there are major issues.

 

[noetsi]

I know next to nothing about this type of statistics. But a generally held view of statistics, repeated endlessly in methods books, is that one can not prove causality (and thus reality) with any type of statistics. Hume's black swan commentary comes into play, in addition structural breaks (that is changes in reality) and unusual events have to be considered.

 

[Jabba]

- I'm back.
- I think I've learned a lot since I last spoke with you guys, and I'd like to present my case again...
- The formula I'll be using is P(H|E) = P(E|H)P(H)/(P(E|H)P(H) + P(E|~H)P(~H)). Any objections so far?

 

[CB]

I have an objection: I don't appreciate having our time wasted with this Socratic-questioning nonsense. If you have a question to post, just post it (in full).

 

[ondansetron]

For someone who says he isn't much of a typist, you've done an awful lot of typing...

Also, I believe I saw a prior argument:
"17. For dummies:
a. The likelihood of a "red state" to elect Candidate X is 10%.
b. State A elects Candidate X.
c. State A is probably not a "red state.""

I thought this kind of probabilistic reasoning is fallacious since it falls into an inverse probability trap. P(X|red)=.1 but then you try to say P(Red|x) which isn't necessarily equal to P(X|red)=.1. This is how people screw up p-values all the time.

 

[Jabba]

ondansetron,
- Is this any better?
If an event is unlikely – given a particular hypothesis (H) – but the event occurs, the occurrence will tend to have a negative effect upon the probability of H — but, it need not.
For instance, it could be that given the complementary hypothesis – the event would be even more unlikely.
Or, it could be that all possible events – given H – are equally unlikely (e.g. a fair lottery) — if so, the particular event needs to be “set apart” in a way that is relevant to the hypothesis in order to impact the hypothesis.
If – given H – an event is impossible, but does occur, H must be wrong.

 

[Jabba]

I have an objection: I don't appreciate having our time wasted with this Socratic-questioning nonsense. If you have a question to post, just post it (in full).

Cowboy,
- I disagree. I think it's the best way to go. If you think it's nonsense, ignore it.

 

[ondansetron]

ondansetron,
- Is this any better?
If an event is unlikely – given a particular hypothesis (H) – but the event occurs, the occurrence will tend to have a negative effect upon the probability of H — but, it need not.
For instance, it could be that given the complementary hypothesis – the event would be even more unlikely.
Or, it could be that all possible events – given H – are equally unlikely (e.g. a fair lottery) — if so, the particular event needs to be “set apart” in a way that is relevant to the hypothesis in order to impact the hypothesis.
If – given H – an event is impossible, but does occur, H must be wrong.

This is different than improbable. Now, you're saying H=>~X (if Hypothesis is true, then we won't see X). The contrapositive is true: X => ~H (If we see X, then H is not true). However, the converse isn't necessarily true. That is, you can't say ~X => H (if we don't see X, H is true), necessarily (would need an if and only if clause in the original). This reasoning is different from the probabilistic versions of the argument (i.e. different from if H then PROBABLY not X, does not mean X then probably not H).

For your earlier argument, I guess that makes sense in that Bayesian updating will often alter the probability of H based on the observed X.

None if this is my expertise, and I may be a little rusty on the formal logic, but I think that the basic issue is that probabilistic logic is different than deterministic in that way.

 

[Jabba]

This is different than improbable. Now, you're saying H=>~X (if Hypothesis is true, then we won't see X). The contrapositive is true: X => ~H (If we see X, then H is not true). However, the converse isn't necessarily true. That is, you can't say ~X => H (if we don't see X, H is true), necessarily (would need an if and only if clause in the original). This reasoning is different from the probabilistic versions of the argument (i.e. different from if H then PROBABLY not X, does not mean X then probably not H).

For your earlier argument, I guess that makes sense in that Bayesian updating will often alter the probability of H based on the observed X.

None if this is my expertise, and I may be a little rusty on the formal logic, but I think that the basic issue is that probabilistic logic is different than deterministic in that way.

ondansetron,
- I don't really understand your last paragraph -- but, I think that I agree with the rest.
- I'm really just saying that the likelihood -- given H -- of a bit of new info is one of the variables determining the posterior probability of H. If the new info is unlikely, it will have a negative effect on the posterior probability of H -- though, the posterior probability of H (based upon other variables as well) may not be smaller than the previous probability of H, and could even be greater.
- I think that makes sense...

 

[Jabba]

1. New information may affect the probability of an existing hypothesis (H). For our purposes here, the current probability is called the “prior probability.”
2. An old event may be new info if it hasn’t already been considered in the current/prior probability of H.
   
   
3. If an event is unlikely – given a particular hypothesis (H) – but the event occurs, the occurrence will tend to have a negative effect upon the probability of H — but, it need not.
4. For instance, it could be that given the complementary hypothesis – the event would be even more unlikely.
5. Or, it could be that all possible events – given H – are equally unlikely (e.g. a fair lottery) -- if so, the particular event needs to be "set apart" in a way that is relevant to the hypothesis in order to impact the hypothesis.
6. If – given H – an event is impossible, but does occur, H must be wrong.
7. Otherwise, what we call Bayesian statistics is used to evaluate the effect of new and relevant information upon the probability of H. This result is called the “posterior probability.”
8. I claim that by using my own current existence as the new info, Bayesian Statistics, virtually proves that I’m not mortal.

 

[Dason]

Does your reasoning apply to any arbitrary person? If there is literally nothing special about you in your reasoning then the fact that the reasoning could apply to people that have died proves a mistake in your reasoning.

 

[Jabba]

Dason,
- It applies to everyone.
- So far, I think this issue is my weakest link -- but, I still think it works...
- From #90 above.
Or, it could be that all possible events – given H – are equally unlikely (e.g. a fair lottery) -- if so, the particular event needs to be "set apart" in a way that is relevant to the hypothesis in order to impact the hypothesis.
- We're all equally unlikely, and in my opinion, equally immortal. Somehow, in order for me to be correct, we must all be set apart from each other...

- First, I'm not a solipsist -- I do assume that the rest of you guys are conscious entities. But then,
- My "self" is the only thing (or process) that I know exists -- the rest could be my imagination.
- If I didn't ever exist, it would be as if nothing ever existed. And, the likelihood of it ever existing -- given modern science -- must be virtually zero.
- If I didn't currently exist, it would be as if nothing currently existed, and the likelihood of me currently existing is even (much) less than the likelihood of me ever existing...
- That gives enormous significance to my current existence...
- And, the thing is, every current human (conscious being?) has the same reason to believe that the opinion that she or he can have only one finite life, at most (OOFLam), is wrong -- and that she or he is not mortal.
- But we all take our current existence for granted – when, scientifically speaking, that’s the very last thing (process) we should take for granted…
- My best guess is that there are "objective" reasons for setting events apart from other similar events, but there are also "subjective" reasons. And the fact that my "self" is the only thing (or process) that I know exists -- the rest could be my imagination -- and that if I didn't ever exist, it would be as if nothing ever existed. And, the likelihood of it ever existing -- given modern science -- must be virtually zero is sufficient subjective reason.

 

[Jabba]

- Couple of problems:
- I need to show superscript -- is there a way to do that here?
- Guess everybody is going to ignore me if I try to stick to my "Socratic" method.
- I don't know if the following will help, but here's my overall claim...

- E is my current existence. H is the hypothesis that I can only have one finite life (at most) and ~H is the hypothesis that H is not true.

- P(H|E) = P(E|H)P(H)/(P(E|H)P(H) + P(E|~H)P(~H))
- P(H|E) = (virtually zero)*.99/((virtually zero)*.99 + .62 * .01)
- P(H|E) = (virtually zero)/(virtually zero) + .0062)
- P(H|E) = (virtually zero)/.0062)
- P(H|E) = virtually zero
- Whereas,
- P(~H|E) = P(E|~H)P(~H)/(P(E|~H)P(~H) + P(E|H)P(H))
- P(~H|E) = .62 * .01/(.62 * .01 + (Virtually zero) * .99)
- P(~H|E) = .0062/(.0062 + (Virtually zero))
- P(~H|E) = virtually 1.
- I need to work on P(E|~H), but it is a real number. I'll get back to that.
- Also, I think that the prior probabilities are more in favor of ~H than I'm allowing, but that even if they are much less in favor of ~H, they are still real numbers.

 

[Jabba]

Using Bayes' theorem is not an instance of deductivereasoning. And if you aren't using deductive reasoning, you can't "prove" things. Statistical and probabalistic arguments can be used to demonstrate evidence, not proof.



This argument is known as the probabalistic modus tollens, and it is a logical fallacy (see Cohen, 1994, p. 998). c does not follow from a and b. E.g., what if the likelihood that a blue state elects Candidate X is 5%? What if there are many more red states than blue states (i.e., what if the prior probability of a state being red is very high)?

The value of Bayes theorem is that it allows us to avoid this line of faulty reasoning, and instead actually calculate values like c (the posterior probability).

Your argument summarised seems to be:

1. Your personal existence would be very improbable if reincarnation does not exist
2. You do personally exist
3. Therefore reincarnation does probably exist

The essential problem I see with your argument is that you are not specifying either (4) the prior probability of reincarnation, before taking into account your personal existence, nor (5) the conditional probability of your personal existence if reincarnation does exist. 4 in particular would surely be vanishingly small, considering that there is no plausible physical mechanism via which reincarnation might occur. You need (4) and (5) to apply Bayes theorem here - your argument as it stands is not a Bayesian argument.

CowboyBear,

- I didn't mean that Bayesian Inference was deduction, I meant that the following was.
- The likelihood of drawing a particular sample from a particular
population has mathematical implications re the probability that a
particular sample was, in fact, drawn from that population...
- In this case, I (the sample) was very likely not drawn from a
hypothetical population of people having just one, finite, life to live.

- I did mean that P(~H) was .01 -- that before taking into account my personal existence, there was a real possibility ofprobability of rei ncarnation

 

[Buckeye]

Where did the .01 probability come from? I think the idea is to assign a probability based on past experiments/research. But there is not much detail on this subject. So, assigning the probability is difficult. I believe this is what Cowboy is saying.

 

[Jabba]

Where did the .01 probability come from? I think the idea is to assign a probability based on past experiments/research. But there is not much detail on this subject. So, assigning the probability is difficult. I believe this is what Cowboy is saying.

Buckeye,
- Thanks for contributing.
- Looking back over my earlier arguments on this site, I can see (at least in part) why I didn't get vry far -- I got caught up in defending a claim that I shouldn't have. If you go back to #57 you'll see that I was finally suspecting I had been wrong... It's the posterior probabilities (like the prior probabilities that must add to 1.00.

- Anyway, I still think that the likelihood of my existence -- given WEBR -- is scientifically and logically supported. From #31 "above":
a.I would never be here if my parents never met.
b.The same is true if either set of my grandparents had never met.
c.Etc., etc., etc.
d.And then, not only does my current, personal, existence depend upon each of these specific +meetings -- back to the beginning of sexual reproduction on this planet -- but each meeting had to ultimately involve fornication, as well as the meeting of the appropriate sperm cell and appropriate ovum, and also, everything that led up to sexual reproduction in life forms in the first place, not to mention everything that led up to life.
e.Like, for instance, the big bang (apparently).
f.And what’s more, my dad probably created a sextillion sperm cells in his lifetime – and, I am the combination of one, specific sperm cell from my dad, and one specific ovum from my mom.
g.Between just the two of them, my Mom and Dad had hundreds of sextillions of different potential children (selves) that were never born!
h.(And, my Dad’s Dad – and, my Mom’s dad -- probably had just as many sperm cells as did my dad…)
i.Anyway, any other combination from my Mom and Dad would be my brother or sister.
j.And then, what about all those potential kids my Dad and Cleopatra could have had, had they met?
k.And, what about all the potential kids and grandkids (etc.) of the potential kids of my Dad and Cleopatra???
l.And, if we can count the potential kids of potential kids, of potential grandkids, etc., our iceberg reaches down to infinity.
m.And scientifically speaking, this is just the tip -- of the tip -- of the iceberg in regard to the number of specific events that had to occur in order for me to currently exist…
n.My current existence is pretty ****ed unlikely if the bio-chemical explanation is correct.
o.And, what this really suggests, of course, is that mainstream science has it all wrong – and, that my self is sort of an inherent, and eternal, resident of reality…

- Please let me know where you disagree.

 

[CB]

This bit:

a.I would never be here if my parents never met.
b.The same is true if either set of my grandparents had never met.
c.Etc., etc., etc.
d.And then, not only does my current, personal, existence depend upon each of these specific +meetings -- back to the beginning of sexual reproduction on this planet -- but each meeting had to ultimately involve fornication, as well as the meeting of the appropriate sperm cell and appropriate ovum, and also, everything that led up to sexual reproduction in life forms in the first place, not to mention everything that led up to life.
e.Like, for instance, the big bang (apparently).
f.And what’s more, my dad probably created a sextillion sperm cells in his lifetime – and, I am the combination of one, specific sperm cell from my dad, and one specific ovum from my mom.
g.Between just the two of them, my Mom and Dad had hundreds of sextillions of different potential children (selves) that were never born!
h.(And, my Dad’s Dad – and, my Mom’s dad -- probably had just as many sperm cells as did my dad…)
i.Anyway, any other combination from my Mom and Dad would be my brother or sister.
j.And then, what about all those potential kids my Dad and Cleopatra could have had, had they met?
k.And, what about all the potential kids and grandkids (etc.) of the potential kids of my Dad and Cleopatra???
l.And, if we can count the potential kids of potential kids, of potential grandkids, etc., our iceberg reaches down to infinity.
m.And scientifically speaking, this is just the tip -- of the tip -- of the iceberg in regard to the number of specific events that had to occur in order for me to currently exist…
n.My current existence is pretty ****ed unlikely if the bio-chemical explanation is correct.

Has absolutely zero connection with this bit:

o.And, what this really suggests, of course, is that mainstream science has it all wrong – and, that my self is sort of an inherent, and eternal, resident of reality…

It's sort of like saying "All swans I have observed thus far have been white, therefore baked beans are made of aliens". Nothing about your argument makes any sense, you haven't developed any meaningful understanding of Bayes theorem, and you're so fixed on trying to demonstrate the validity of your argument (rather than actually learning anything) that this discussion is pointless.

 

[Jabba]

This bit:



Has absolutely zero connection with this bit:



It's sort of like saying "All swans I have observed thus far have been white, therefore baked beans are made of aliens". Nothing about your argument makes any sense, you haven't developed any meaningful understanding of Bayes theorem, and you're so fixed on trying to demonstrate the validity of your argument (rather than actually learning anything) that this discussion is pointless.

- I sort of overstated, but all I'm trying to say is that the current existence of my self -- given H (that my self can have only one finite existence (at most)) -- is so unlikely that we have to wonder about the validity of H.

- I'm still learning how to navigate this forum -- please keep that in mind...

- But anyway, I'm just trying to say that my unlikelihood -- given H -- has mathematical implications (albeit indefinite) re the probability of H.
- The implications are indefinite because P(E|H) is only one of the variables in the formula.

 

[Jabba]

Firstly: I don't mean to be rude, but part of the reason why you may not be holding everyone's attention is that you have taken pages and pages of posts to convey what is really a very simple argument, that could be summarised in a paragraph. You're asking a lot of the reader to do that summarising for you.

Cowboybear,
- Try this. There are 4 variables involved when trying to determine the posterior probability of a hypothesis (H) given new info (E). The formula being P(H|E) = P(E|H)P(H)/(P(E|H)P(H) + P(E|~H)P(~H)).
- In regard to the hypothesis that my "self" can have only one finite life to live (at most), I propose that P(E|H) is virtually zero, P(H) is .99, P( E|~H) is about .60 and P(H) is .01.
- I also suggest that the P(E|H) is so small that the P(~H) could be any real number (say .000001) and P(H|E) would still be virtually Zero.
- I also propose that science keeps coming to (and will keep coming to) new, and important, conclusions. Quantum Mechanics seems to be taking us where no scientist has gone before -- including the importance of the observer and consciousness.

 

[Jabba]

Let's take a simple counterexample to show why the likelihoods don't have to add to 1.

Imagine you are walking behind someone, and can't make out their gender very confidently. You have observed, however, that they are wearing pants. Let's assume the following probabilities:

50% of people are male, and 50% female; that is, P(Male) = 0.50
A large percentage, say 99%, of men wear pants; that is, L(Pants|Male) = 0.99
But quite a good number of women do too, say 55%; that is, L(Pants|Female) = 0.55

Now answer the following questions:

1. What is the posterior probability that the person is male?
2. Do the two likelihoods add to 1?
3. If not, is there any problem with that? Do the proportion of women who wear pants and the proportion of men who wear pants really have to sum to 1?[/quote]CowboyBear,
- Re:
1. L(M|P)=0.99*0.5/(0.99*0.5+0.55*0.5)=.642857
2. No.
3. No.

If that still doesn't make sense to you, I want you to forget about your immortality argument for a while and go do some actual reading to get a grasp of Bayes theorem. At the moment we are not being held up by a disagreement, but just by you not grasping the framework of Bayes theorem. Please have the humility to accept that.

Once you've done that, if you really want to apply Bayes theorem to your problem, you need to forget about this approach of trying to calculate L(me|~WEBR) from the other probability terms and instead try to find some other way of estimating this probability:

- Agreed.

In a world with immortality or reincarnation, what would be the probability of your current existence?

Note: Even without a good grasp of Bayes theorem, it should be intuitively obvious to you that this should not be a very large probability! By your current assumption, L(me|~WEBR) is close to 1, but that's obviously not the case - for you to exist, even in a world with reincarnation or immortality, all kinds of unlikely things had to go "right".

- Under ~WEBR (I'll plan on calling it "~H" from now on.), no body is required, let alone this particular body. If my particular body were required, P(M|~H) would be just as small as P(M|H).
- Otherwise, P(M|~H) should be determined by trying to determine the different possible specific hypotheses under ~H, their different probabilities of being true, the likelihood of my current existence given each, and then adding them all up.
- For now, I'll make H, and therefore ~H, simpler by addressing only the issue of my mortality rather than of human mortality. If you're with me so far, perhaps we could address the specifics together...

Sorry to be harsh, but I'm not going to engage further in this discussion unless you show some actual effort at trying to grasp Bayes theorem - I have work to do.

- I accept that until recently, my explanations and answers haven't been very good or forthcoming... Hopefully, that's improving.