Proof of Immortality, VII

Status
Not open for further replies.
I can't get past this fundamental point in your claim. You acknowledge that it is the weakest link, but it's a show-stopper.

This is his overall schtick. It's how he keeps the semblance of debate going literally for years as nothing more than a shell game. When Jabba says that the Texas sharpshooter fallacy is the weakest link in his argument, what he wants you to believe by that is that the rest of his argument is otherwise strong. It isn't. The Texas sharpshooter fallacy is only one of about a dozen of specific, largely distinct errors we have identified in his argument. Each one of them is a individual show-stopper. But he will only discuss one at a time, so that at each step in the argument he can say, "Okay, I admit this thing we're talking about now is a weak point, but it's the only one." He couches this limitation as a tenet of his principles of "effective debate," of which he claims to be a guru.

Last summer I tried to bust the shell game. You know how to bust up a shell game. You make the operator turn over all the shells at the same time after picking your shell, to prove there's a pea under one of them even if it wasn't the one you chose. I told Jabba he should sketch out, in one or two sentences, how his argument would handle each of the fatal flaws. And he had to do it all at once, not strung out in his usually evasive way. That's why I limited the responses for that purpose to something he could write briefly without needing to supply a lengthy argument or evidence for any of them. Otherwise experience has shown that he'd just evade differently by saying "I don't have time to do what you ask; it's too big a task."

See, shell-game operators are smart. They know you suspect it's a scam, so they go through a pretense of showing you the pea behaves normally -- but they do so only one shell at a time. "See, here's the pea under this one shell I've turned over." Shuffle, shuffle, shuffle. "See, here's the pea again where you thought it should be, under the one shell I've turned over." Shuffle, shuffle, shuffle. "Now where's the pea?" Of course the guess will be wrong, because sleight of hand ensures the pea will never be under the selected shell when the selection actually counts. But if the player selects a shell, and then all the shells are turned over to reveal no pea under any of them, the jig is up. (Ideally the operator has to open his hands too, but you see where the analogy is going.)

Jabba promised he would address all the fatal flaws, but only if he could be allowed to address them one at a time, in his longstanding fashion. That is, only if he could be allowed to shuffle the shells between examinations as he's been doing hitherto. When we tried to change the game in a way that required him to ensure fairness, he admitted he couldn't do it. Or wouldn't do it; no real difference. I'll let you draw your own conclusions from that data regarding whether Jabba's approach is an effective means of determining whether his proof is valid.
 
I'm saying that if time consists of 14 billion years, and I live to be 100, the likelihood of now corresponding with the time of my existence is 1/140,000,000...

Neither now nor your lifetime is a uniformly distributed random variable, nor are they independent variables. Your math is wrong.
 
- No.
- I'm saying that if time consists of 14 billion years, and I live to be 100, the likelihood of now corresponding with the time of my existence is 1/140,000,000 -- and that likelihood pretty much assures that OOFLam is wrong.

"What are the chances of me shooting this very spot in my foot? I've even drawn a circle around it to show how improbable it was!"
 
- I'm saying that if time consists of 14 billion years, and I live to be 100, the likelihood of now corresponding with the time of my existence is 1/140,000,000 -- and that likelihood pretty much assures that OOFLam is wrong.

Why don't you read the responses to this?

You could not have lived before the stars formed, before humans existed, or before your parents met.
 
- No.
- I'm saying that if time consists of 14 billion years, and I live to be 100, the likelihood of now corresponding with the time of my existence is 1/140,000,000 -- and that likelihood pretty much assures that OOFLam is wrong.

Nope, time is not a random lottery!
 
- No.
- I'm saying that if time consists of 14 billion years, and I live to be 100, the likelihood of now corresponding with the time of my existence is 1/140,000,000 -- and that likelihood pretty much assures that OOFLam is wrong.

Jabba,
For your little foray into Bayes inference, you picked Jabba, an existing person. We could have explored questions about Jabba's hair color, his age, his height, his...so many, many things all without committing any sort of sharpshooter fallacy.

However, having picked Jabba, an existing person, the answer to any question of his existence is already known.

P(E) = 1.
 
- No.
- I'm saying that if time consists of 14 billion years, and I live to be 100, the likelihood of now corresponding with the time of my existence is 1/140,000,000 -- and that likelihood pretty much assures that OOFLam is wrong.

That would only be true if you had an equal chance of existing at each point in that whole 14 billion year period, which you just told me you weren't saying.
- Likelihood of the Bayesian variety has nothing (directly) to do with science. It has only to do with math and the givens.
 
...And, if OOFLam is true, the odds of my current existence is extremely small -- and consequently, my current existence casts doubts upon OOFLam.


No. Something with extremely long odds that occurs does not imply that there is doubt on that thing occurring. Many things happen every day that defy odds. You've been given many examples over the past few years.
 
Likelihood of the Bayesian variety has nothing (directly) to do with science.

Utter hogwash. Any statistical inference must properly model the underlying system in order to produce useful results. And the underlying systems we observe in nature do not invariably line up as discrete, uniform probability distributions the way your thinking seems limited to.

In terms of Bayes (as opposed to frequentist thinking), we may certainly incorporate the effects of informal knowledge. But insofar as that knowledge is not based at least in significant part on observation and empirical analysis, it is not useful for inference. It certainly couldn't be used to prove anything. For example, in constructing a Bayesian search plan, we may consult with the local sheriff's deputies for their opinions about how hard it would be to overlook a missing child in some given part of the search area. Such information would be strictly informal as science reckons rigor, and would be subjective in the way Bayesians use the term. But it would still ostensibly be based on relatively unbiased real-world expertise of those deputies and their familiarity with the land. Those are still factors informed by facts, not flights of fancy.

There is absolutely nothing in Bayesian reasoning that says you get to ignore science when the science rightly applies.

It has only to do with math and the givens.

The proper formulation of the math and the givens is dictated by the behavior of the system being modeled, which in turn is dictated by the natural laws that pertain to it. But since you don't understand how probability density can vary, you just want to pretend it doesn't exist or doesn't apply. As I said before, you're trying to make the problem fit your understanding rather than expand your understanding to properly accommodate the problem.
 
Last edited:
The science is part of the givens.
- You're right (in our current case) -- but that is the extent to which science is involved. If a scientific claim is not part of the givens, it has nothing to do with "likelihood" of the Bayesian variety.
- In this case, all the science we're given is that time exists for 14 billion years, and we're asking for the likelihood that now would be within a particular century (100 years).
 
You're right (in our current case)

Yes, it's valid for the case we're talking about. It's irrelevant whether it might be valid or invalid for some other case. It applies to this case, and here it refutes this case.

...but that is the extent to which science is involved.

It's an extent that pertains to your question and refutes your model of it.

If a scientific claim is not part of the givens, it has nothing to do with "likelihood" of the Bayesian variety.

Straw man. We're not talking about "scientific claims." If you're using statistics to model a real-world system, whatever science that applies to that real-world system also applies to your model. A model that does not accurately capture the behavior of the underlying system is not useful. Science is what we use to describe the behavior of all such underlying systems.

In this case, all the science we're given is that time exists for 14 billion years, and we're asking for the likelihood that now would be within a particular century (100 years).

But your claim is not limited to the existence of time. Your claim also involves the likelihood of human life arising. It is a scientific fact that human life is not equally viable across all 140 million centuries. Human life is far more likely to arise in the past 200,000 years than in the previous eons. That's an example of a non-uniform probability density.

But all that is largely moot because your "particular century" was not chosen at random from all the possible centuries. It was specifically selected by you because it's the century corresponding to your life. No random variable was involved in that event.
 
- You're right (in our current case) -- but that is the extent to which science is involved. If a scientific claim is not part of the givens, it has nothing to do with "likelihood" of the Bayesian variety.
- In this case, all the science we're given is that time exists for 14 billion years, and we're asking for the likelihood that now would be within a particular century (100 years).


But that's not all the science we're given. Among other things, we know things about the state of the universe over the span of that 14 billion years. We know that the planet we live on has only existed for 4.5 billion years, only supported multi-cellular life for less than a billion, and modern humans have only existed for around 200,000 years. For almost all of that 14 billion years, the existence of you or any other human was impossible.

Science also tells us that your specific existence was the result of two specific people reproducing, which further limits the span of time in which you could exist.
 
Last edited:
Utter hogwash. Any statistical inference must properly model the underlying system in order to produce useful results. And the underlying systems we observe in nature do not invariably line up as discrete, uniform probability distributions the way your thinking seems limited to.

In terms of Bayes (as opposed to frequentist thinking), we may certainly incorporate the effects of informal knowledge. But insofar as that knowledge is not based at least in significant part on observation and empirical analysis, it is not useful for inference. It certainly couldn't be used to prove anything. For example, in constructing a Bayesian search plan, we may consult with the local sheriff's deputies for their opinions about how hard it would be to overlook a missing child in some given part of the search area. Such information would be strictly informal as science reckons rigor, and would be subjective in the way Bayesians use the term. But it would still ostensibly be based on relatively unbiased real-world expertise of those deputies and their familiarity with the land. Those are still factors informed by facts, not flights of fancy.

There is absolutely nothing in Bayesian reasoning that says you get to ignore science when the science rightly applies.



The proper formulation of the math and the givens is dictated by the behavior of the system being modeled, which in turn is dictated by the natural laws that pertain to it. But since you don't understand how probability density can vary, you just want to pretend it doesn't exist or doesn't apply. As I said before, you're trying to make the problem fit your understanding rather than expand your understanding to properly accommodate the problem.
- "Likelihood" is only part of the Bayesian formula -- science does get involved, but except for the givens, only in the "prior probabilities."
 
- "Likelihood" is only part of the Bayesian formula -- science does get involved, but except for the givens, only in the "prior probabilities."

No, that doesn't address anything I said. Probability density affects all parts of the Bayesian model. Can you describe probability density in your own words?
 
- You're right (in our current case) -- but that is the extent to which science is involved. If a scientific claim is not part of the givens, it has nothing to do with "likelihood" of the Bayesian variety.
- In this case, all the science we're given is that time exists for 14 billion years, and we're asking for the likelihood that now would be within a particular century (100 years).

You second statement is silly, there is not a random assignment of time.
Here is always now.
 
- You're right (in our current case) -- but that is the extent to which science is involved. If a scientific claim is not part of the givens, it has nothing to do with "likelihood" of the Bayesian variety.
- In this case, all the science we're given is that time exists for 14 billion years, and we're asking for the likelihood that now would be within a particular century (100 years).


But now has to be some time. And the only time we could ask that question is, by definition, ...now. To put another way: What is the likelihood of now falling within a 100-year time span of the time of asking?

Now is relative. Now will always fall within a time span of the moment you define now.
 
Status
Not open for further replies.

Back
Top Bottom