Proof of Immortality, the IV league

[jond]

Zoo,
- I should have added that...

You should have also added that the scientific model of OOFLam does not contain a soul, and is therefore rather considerably more likely than your version of OOFLam which does contain one.

 

[John Jones]

- Here's a map of our discussion, as I see it -- from the edge of space.

- I claimed that through the use of Bayesian statistics, I could virtually prove immortality.
- Later, I reduced that to something more direct, logically speaking -- i.e., through the use of Bayesian statistics, I could virtually disprove OOFLam (that we each have only one finite life to live (at most)).
- I think there were (was?) six basic objections to my claim: 1)Bayes isn't applicable, 2)there is no real "self," 3)the likelihood of my current existence -- given OOFLam -- is not infinitesimal 4)the likelihood I prescribe for my existence -- given ~OOFLam -- doesn't make sense, 5)it is not reasonably possible that OOFLam is wrong and 6) the likelihood of MY current existence is not an appropriate entry for P(E|OOFLam).

- Lately, we've been working on #2 and #5.
- Under #5, we got into NDE's and the validity of anecdotes as evidence re scientific hypotheses.
- Re anecdotes, most participants disagreed that anecdotes could constitute evidence re scientific hypotheses, but several posters actually agreed with me -- that anecdotes can constitute such evidence.

- Currently, I'm trying to engage some of those latter participants in exploring the credibility of NDE anecdotes.

You forgot to mention all the fringe resets, the quoting people out of context, and ignoring the vast number of posts that debunk your claims

 

[abaddon]

- Here's a map of our discussion, as I see it -- from the edge of space.

What, now you are claiming to be on the ISS? That you are somehow impartial? That you have no skin in this game you are playing?

Nonsense.

 

[godless dave]

- Currently, I'm trying to engage some of those latter participants in exploring the credibility of NDE anecdotes.

You already engaged people and you failed to convince them. The best you could offer were unverified anecdotes that wouldn't be evidence for immaterial souls even if they were true.

 

[Loss Leader]

- Here's a map of our discussion, as I see it -- from the edge of space.

- I claimed that through the use of Bayesian statistics, I could virtually prove immortality.
- Later, I reduced that to something more direct, logically speaking -- i.e., through the use of Bayesian statistics, I could virtually disprove OOFLam (that we each have only one finite life to live (at most)).
- I think there were (was?) six basic objections to my claim: 1)Bayes isn't applicable, 2)there is no real "self," 3)the likelihood of my current existence -- given OOFLam -- is not infinitesimal 4)the likelihood I prescribe for my existence -- given ~OOFLam -- doesn't make sense, 5)it is not reasonably possible that OOFLam is wrong and 6) the likelihood of MY current existence is not an appropriate entry for P(E|OOFLam).

- Lately, we've been working on #2 and #5.
- Under #5, we got into NDE's and the validity of anecdotes as evidence re scientific hypotheses.
- Re anecdotes, most participants disagreed that anecdotes could constitute evidence re scientific hypotheses, but several posters actually agreed with me -- that anecdotes can constitute such evidence.

- Currently, I'm trying to engage some of those latter participants in exploring the credibility of NDE anecdotes.

"Lately, I've been trying to surround the word 'anecdote' with so much fury and nonsense that people forget the essential failure that dependence on anectotes entail."

 

[Dancing David]

Fudbucker,
- Then, multiply the likelihood of such a universe by the likelihood of your current existence.

That is the point where your argument fails, you do exist.
Period, therefore the odds of your existence are 1.

The fact that organic being seem to exist in a chaotic universe does mean that there is an infinite or whatever the random scramble of 10^70 particles produces.

There is defendant history. Period. You, if you exist come about from the dependent history or your parents procreating and the combination of genetics.

Secondly, none of that has anything to do with immaterialism or souls or whatever.

If you exist there is an organic body, until you demonstrate something beyond an organic body there is nothing but angels dancing on the head of the pin and what the prior probability of them wearing sneakers versus high heels is.

 

[JayUtah]

Later, I reduced that to something more direct, logically speaking --

No, you transformed it into something indirect. Instead of proving X, you conjured up a false dilemma between X and Y and tried to disprove Y. Your disproof involves simply throwing a lot of unevidenced doubt at it. This is what every fringe theorist does. When faced with their inability to prove their point, they try to cast aspersions on what the rest of the world believes in order to make their claim stand by default.

...but several posters actually agreed with me -- that anecdotes can constitute such evidence.

No. You're equivocating between "anecdotes are evidence" and "these anecdotes are evidence of life after death." You tried to get the people who agreed with the former to carry their agreement to the latter and those who answered repudiated you. Don't pretend people agree with you when they say they don't.

Currently, I'm trying to engage some of those latter participants in exploring the credibility of NDE anecdotes.

It doesn't matter. They aren't evidence of life after death. I've even quoted your own sources saying as much.

 

[jt512]

ANY old evidence has a probability of 1 by virtue of the fact it's already known. This is the problem of "old evidence". One of the solutions is to look at the evidence counterfactually.

One of the best examples is to imagine you're going to be executed by being shot by a hundred sharpshooters. The order to fire is given, and you're miraculously not dead. The probability that you're alive is 1, of course, yet the probability that it was just an accident you survived is now practically zero. Because counterfactually, you know that your continued existence would be incredibly improbable on the chance hypothesis (that all 100 shooters missed or misfired or had sudden heart attacks, etc.). So your continued existence constitutes incredibly powerful evidence that the execution was rigged in some way.

The data always exist. Otherwise, there wouldn't be anything to analyze. So, it's not the probability of the data per se that we we care about, but the probability of the data under one or more models about how the data arose. The P(E) term in Bayes' Theorem is the average probability of the data under all the hypotheses in the hypothesis space—actually, a weighted average, whose weights are the prior probabilities of each hypothesis.

 

[Fudbucker]

The data always exist. Otherwise, there wouldn't be anything to analyze. So, it's not the probability of the data per se that we we care about, but the probability of the data under one or more models about how the data arose. The P(E) term in Bayes' Theorem is the average probability of the data under all the hypotheses in the hypothesis space—actually, a weighted average, whose weights are the prior probabilities of each hypothesis.

Data that is collected after a hypothesis is formulated (i.e., data that is predicted by the hypothesis) is much easier to work with than data that exists before the hypothesis. See Mercury's eccentric orbit and relativity theory.

" In other words: ee must have been predicted on the basis of h∧kh∧k. Moreover, again by (i), the confirmatory impact will be stronger the more surprising (unlikely) the evidence was unless hh was conjoined to kk. So, under TE, relevance confirmation turns out to embed a squarely predictivist version of hypothetico-deductivism! As we know, this neutralizes the charge of underdetermination, yet it comes at the usual cost: the old evidence problem. In fact, if TE is in force, then clause (ii) of (SP) implies that no statement that is known to be true (thus assigned probability 1) can ever have confirmatory import.

Interestingly, the Bayesian predictivist has an escape (neatly anticipated, and criticized, by Glymour 1980a, 91–92). Consider Einstein and Mercury once again. As effectively pointed out by Norton (2011a, 7), Einstein was extremely careful to emphasize that the precession phenomenon had been derived “without having to posit any special [auxiliary] hypotheses at all”. Why? Well, presumably because if one had allowed herself to arbitrarily devise ad hoc auxiliaries (within kk, in our notation) then one could have been pretty much certain in advance to find a way to get Mercury’s data right (recall: that’s the lesson of the underdetermination theorem)."

https://plato.stanford.edu/entries/confirmation/#BayConThe

But this is getting extremely wonky and probably nobody besides you or me cares about this that much.

 

[jt512]

Data that is collected after a hypothesis is formulated (i.e., data that is predicted by the hypothesis) is much easier to work with than data that exists before the hypothesis. See Mercury's eccentric orbit and relativity theory.

" In other words: ee must have been predicted on the basis of h∧kh∧k. Moreover, again by (i), the confirmatory impact will be stronger the more surprising (unlikely) the evidence was unless hh was conjoined to kk. So, under TE, relevance confirmation turns out to embed a squarely predictivist version of hypothetico-deductivism! As we know, this neutralizes the charge of underdetermination, yet it comes at the usual cost: the old evidence problem. In fact, if TE is in force, then clause (ii) of (SP) implies that no statement that is known to be true (thus assigned probability 1) can ever have confirmatory import.

Interestingly, the Bayesian predictivist has an escape (neatly anticipated, and criticized, by Glymour 1980a, 91–92). Consider Einstein and Mercury once again. As effectively pointed out by Norton (2011a, 7), Einstein was extremely careful to emphasize that the precession phenomenon had been derived “without having to posit any special [auxiliary] hypotheses at all”. Why? Well, presumably because if one had allowed herself to arbitrarily devise ad hoc auxiliaries (within kk, in our notation) then one could have been pretty much certain in advance to find a way to get Mercury’s data right (recall: that’s the lesson of the underdetermination theorem)."

https://plato.stanford.edu/entries/confirmation/#BayConThe

But this is getting extremely wonky and probably nobody besides you or me cares about this that much.

I don't even understand it.

 

[Fudbucker]

I don't even understand it.

The problem of old evidence? Suppose a theory (Y) predicts X and X occurs. If X is highly improbable, then X is going to confirm Y to a fairly high degree.

But suppose that evidence X (Mercury's eccentric orbit) is already known, before Y (relativity theory) is even formulated. Long before Einstein came along, people were trying to explain Mercury's weird orbit. Einstein then had to be extra careful to make sure his theory wasn't ad hoc in any way (that he didn't already incorporate Mercury's orbit into the framework of this theory). Because it was obvious that relativity theory wasn't ad hoc, its prediction of Mercury's eccentric orbit was very successful, even though it "predicted" something that people had known about for a long time.

To put it another way: it's a lot more impressive if I predict an earthquake before it happens, than after it happens.

 

[jt512]

The problem of old evidence? Suppose a theory (Y) predicts X and X occurs. If X is highly improbable, then X is going to confirm Y to a fairly high degree.

I think you need to say that if X is highly improbable under some other theory, Z, then X confirms Y relative to Z.

But suppose that evidence X (Mercury's eccentric orbit) is already known, before Y (relativity theory) is even formulated. Long before Einstein came along, people were trying to explain Mercury's weird orbit. Einstein then had to be extra careful to make sure his theory wasn't ad hoc in any way (that he didn't already incorporate Mercury's orbit into the framework of this theory). Because it was obvious that relativity theory wasn't ad hoc, its prediction of Mercury's eccentric orbit was very successful, even though it "predicted" something that people had known about for a long time.

The problem occurs when you use the data to formulate the theory, and then claim that the theory predicted the data. This has been called HARKing, hypothesizing after the results are known. A researcher observes an unanticipated correlation between two variables in a data set, finds that it is statistically significant, formulates a theory for it, and then claims that the theory "predicts" the data. It's similar to what Skeptics call the "Texas sharpshooter fallacy."

Another example comes from political scientist Allan Lichtman and his Keys to the White House Model. Writers often claim that Lichtman's model has correctly predicted the popular vote in every US presidential race from 1860 to 2012. But that is ridiculous, since he developed the model in 1981 by using the data from the elections from 1860 to 1980. His model should only get credit for correct predictions on new data, elections after 1981.

 

[Fudbucker]

I think you need to say that if X is highly improbable under some other theory, Z, then X confirms Y relative to Z.





The problem occurs when you use the data to formulate the theory, and then claim that the theory predicted the data. This has been called HARKing, hypothesizing after the results are known. A researcher observes an unanticipated correlation between two variables in a data set, finds that it is statistically significant, formulates a theory for it, and then claims that the theory "predicts" the data. It's similar to what Skeptics call the "Texas sharpshooter fallacy."

Another example comes from political scientist Allan Lichtman and his Keys to the White House Model. Writers often claim that Lichtman's model has correctly predicted the popular vote in every US presidential race from 1860 to 2012. But that is ridiculous, since he developed the model in 1981 by using the data from the elections from 1860 to 1980. His model should only get credit for correct predictions on new data, elections after 1981.

That's the problem of old evidence. The solution I've always thought was best is to counterfactually look at the theory as if it were proposed without the evidence already having been known. I.e., we pretend Einstein knew nothing about Merucry's orbit. Does relativity successfully predict Mercury's orbit, or was there some ad hoc reasoning going on? Even if Einstein knew nothing about Mercury, he would have formulated his theory the same way, so it gets credit for "predicting" something everyone (including Einstein) already knew about.

 

[jt512]

That's the problem of old evidence. The solution I've always thought was best is to counterfactually look at the theory as if it were proposed without the evidence already having been known. I.e., we pretend Einstein knew nothing about Merucry's orbit. Does relativity successfully predict Mercury's orbit, or was there some ad hoc reasoning going on? Even if Einstein knew nothing about Mercury, he would have formulated his theory the same way, so it gets credit for "predicting" something everyone (including Einstein) already knew about.

The solution in more prosaic sciences is to test the hypothesis on an independent data set.

 

[Fudbucker]

The solution in more prosaic sciences is to test the hypothesis on an independent data set.

Sure, but you can't just ignore evidence that predates the hypothesis. Relativity (and I'm going from memory here), received significant confirmation both by explaining Mercury's orbit in a non-ad hoc way and predicting the observation of light bending during a solar eclipse.

 

[jt512]

Sure, but you can't just ignore evidence that predates the hypothesis. Relativity (and I'm going from memory here), received significant confirmation both by explaining Mercury's orbit in a non-ad hoc way and predicting the observation of light bending during a solar eclipse.

It's not whether the data predated the hypothesis (ie, a model) that matters; it is whether the data was used in the formulation of the model. Data that was used to formulate a model biases the model in favor of that data, so it is not valid to use that same data to test the hypothesis. To test the hypothesis, new data are needed. The reason we prefer data that were observed after the model was formulated is that we know that the scientist could not have used the data in the formulation of the model.

 

[abaddon]

It's not whether the data predated the hypothesis (ie, a model) that matters; it is whether the data was used in the formulation of the model. Data that was used to formulate a model biases the model in favor of that data, so it is not valid to use that same data to test the hypothesis. To test the hypothesis, new data are needed. The reason we prefer data that were observed only after the model was formulated is that it we know that the scientist could not have used the data in the formulation of the model.

Which leaves you claiming that immortal souls exist. Sorry. You bought it, you own it.

 

[marplots]

It's not whether the data predated the hypothesis (ie, a model) that matters; it is whether the data was used in the formulation of the model. Data that was used to formulate a model biases the model in favor of that data, so it is not valid to use that same data to test the hypothesis. To test the hypothesis, new data are needed. The reason we prefer data that were observed only after the model was formulated is that it we know that the scientist could not have used the data in the formulation of the model.

What should count as "new data" then? It seems to me that mere repetition of the same type of data (from the same source) shouldn't count, but I'm not sure.

Best would be something new and unexpected. But let's say my model is based on one type of starfish being bigger than another (maybe I'm interested in starfish growth related to global warming). It doesn't seem like I can then bolster my model by finding more examples of one starfish being bigger, point to this "new" data, and claim I've found confirmation.

 

[Mojo]

It's not whether the data predated the hypothesis (ie, a model) that matters; it is whether the data was used in the formulation of the model.

The relationship of the data to he model is also important. In the current case the "new data" Jabba is suggesting using is not something that is predicted by the model. It is not something that the model says is impossible. It is just one of many possible outcomes that is consistent with the model. Because there are many of them the likelihood of this particular outcome is low, but the likelihood of one of them having occurred is 1. Selecting the one that happens to have occurred says nothing about the likelihood of the model being correct.

 

[Jabba]

The relationship of the data to he model is also important. In the current case the "new data" Jabba is suggesting using is not something that is predicted by the model. It is not something that the model says is impossible. It is just one of many possible outcomes that is consistent with the model. Because there are many of them the likelihood of this particular outcome is low, but the likelihood of one of them having occurred is 1. Selecting the one that happens to have occurred says nothing about the likelihood of the model being correct.

Mojo,
- The "new data" I'm using is the scientific supposition(s) of how a "self" (or the process/illusion of a self) comes about. This aspect of the surrounding data has not been taken into account (as far as I can tell) by Scientific consideration re the prior probability of OOFLam.