Proof of Immortality, the IV league

[JayUtah]

I still don't understand why he thinks it's a gotcha. If you could have one soul looking through two sets of eyes, that would be evidence that selves aren't material and aren't tied to a specific physical brain.

Because he thinks a new perspective from the newly-minted clone amounts to having a new consciousness rather than having two beings, each of which has the property of consciousness. To him it's tantamount to proving consciousness must be a thing. He can't differentiate between the notions of possessing a soul and exhibiting a property.

 

[Sideroxylon]

It's also be both highly confusing and useful. Imagine the black ops you could achieve if the team members didn't need to actually communicate!

Also, kinky stuff.

You could literally go **** yourself.

 

[Belz...]

You could literally go **** yourself.

Yeah but you don't need to share experiences to do that.

 

[jsfisher]

Irrelevant.

I disagree.

Let H be "my parents have met".
Let ~H be "my parents have not met".
Let E be "I exist".

Ok, then P(E) = 1, or are you trying to claim there is a possibility you do not exist? And, if you didn't exist (!), who would be your parents?

...
If we're following this reasoning of "you're conditioning on your own existence therefor it should be P(E | E, H) = P(E | E, ~H) = 1" then my existence would not be evidence for my parents having met.

Correct. It would not be. P(H | E) = P(H) and P(~H | E) = P(~H)

 

[Hokulele]

- Each new consciousness has its own self.
- If I were physically replicated, I would not be looking out of two sets of eyes. That "I" is the "self" I'm talkin about.

If you were replicated, would you know whether or not you were the original you?

 

[caveman1917]

I disagree.



Ok, then P(E) = 1, or are you trying to claim there is a possibility you do not exist? And, if you didn't exist (!), who would be your parents?



Correct. It would not be. P(H | E) = P(H) and P(~H | E) = P(~H)

Wut? What does it even matter what P(E) is?

Suppose there are 2 persons, Alice and Bob. We want to know whether these 2 persons have met, and our priors are equal P(H) = P(~H). Suppose Charlie comes forwards and says "I am their child", then according to you this is not evidence for Alice and Bob to have met? Or is it only Charlie himself who is not "allowed" to reach such conclusion?

 

[marplots]

You are assuming that the hypothesis you derive from the first sample is true for the population, hence why you don't see new information from the second sample. Suppose your first sample showed that about half of the numbers are evenly divisible by 3. The second sample refutes that, showing that it was just a statistical fluke in your first sample. You got new information and have updated your beliefs about "half of them are divisible by 3".

Yes, but that's a reflection on the quality/representativeness of the original sample, not the hypothesis. Our hypothesis isn't: "our sample is good" it's something else which assumes the sample is good and duplicates the sample.

The whole point of the thought experiment was to avoid hypotheses which are "proven" by revisiting the data used to construct them. I see repeated samples of the same quality as doing just that.

What you might argue is that repeated sampling has the potential to refute an hypothesis, even if it cannot add evidence to support it. I agree with that. If further sampling of the same data set shows a different result, then there's a fly in the ointment. Which leads me to: evidence for ~A may exist without evidence for A - by way of ~~A. At least when A is constructed from the same pool used to test it.

 

[caveman1917]

Or try this:

Let H be "humans exist"
Let ~H be "humans do not exist"
Let E be "I exist"

Let P(E | H) >> P(E | ~H) = 0. Then my existence, as a human, is not evidence for the existence of humans?

 

[caveman1917]

The whole point of the thought experiment was to avoid hypotheses which are "proven" by revisiting the data used to construct them.

The data used to construct the hypothesis isn't the population but the first sample, it is exactly by taking a second sample that you are avoiding revisiting the data.

Strictly speaking this only holds true exactly in infinite sample spaces, but it holds approximately in finite ones as well.

What you might argue is that repeated sampling has the potential to refute an hypothesis, even if it cannot add evidence to support it.

It can do both, it's just easier to argue the refutation case.

 

[marplots]

The data used to construct the hypothesis isn't the population but the first sample, it is exactly by taking a second sample that you are avoiding revisiting the data.

Can you give an example? I find specifics helpful.

 

[caveman1917]

Can you give an example? I find specifics helpful.

I did. A sample of integers in which half of them are divisible by 3, then "refuted" by a new sample in which only 1/3 are. Although this is an "extreme" case it is not a special case, there will always be a change strictly greater than zero if the sample size is strictly smaller than the population size.

The positive case can be argued if you accept that a larger sample is better (more evidence) than a smaller sample, then taking a second sample enlarges your total sample and hence increases your certainty of the hypothesis derived from the first sample.

 

[marplots]

I did. A sample of integers in which half of them are divisible by 3, then "refuted" by a new sample in which only 1/3 are. Although this is an "extreme" case it is not a special case, there will always be a change strictly greater than zero if the sample size is strictly smaller than the population size.

The positive case can be argued if you accept that a larger sample is better (more evidence) than a smaller sample, then taking a second sample enlarges your total sample and hence increases your certainty of the hypothesis derived from the first sample.

I think you are missing the thrust of the argument.

You seem to be equating a sample which accurately reflects the characteristics of the population as being equivalent to a "good" hypothesis - after all, that's where you are getting your validation.

However, the hypothesis is constructed based on the sample, specifically to match it. The hypothesis in such circumstances must match the population to the extent the sample does. So long as my sample does match the population, any further sampling with the same fidelity will also match the population. We haven't done anything new.

"I looked in my driveway and my car was there. I hypothesize that pixies put it there. To support this hypothesis, I looked again to see that not only was the car there, but so were the tires, the engine, and the windshield."

"Well, OK. I believe you. The car is there. Any further sampling you do will give us the same result. But what about those pixies?"

"The pixies put the car there. The car being there supports my hypothesis."

"But you only made that up after you knew the car was there!"

"Yes, but I re-sampled and the car was still there."

"But you didn't add any new information."

"Oh, but my wife looked from the other side and saw the car too. She also spotted the little door where you put gas in."

"Did she see any pixies?"

"No. The pixies put the car there, they don't stick around."

 

[caveman1917]

However, the hypothesis is constructed based on the sample, specifically to match it. The hypothesis in such circumstances must match the population to the extent the sample does. So long as my sample does match the population, any further sampling with the same fidelity will also match the population. We haven't done anything new.

Yes we have, we've become more certain that the sample does indeed match the population regarding some property, or equivalently, that the hypothesis derived from the sample holds true for the population.

 

[marplots]

Yes we have, we've become more certain that the sample does indeed match the population regarding some property, or equivalently, that the hypothesis derived from the sample holds true for the population.

Exactly. The hypothesis is constructed that way. I know the answer I want and then construct the question.

But there's more to the story. Suppose I get another sample and find my original model doesn't match. I now alter my model to fit the combined sample. Is my new model better? Or have I merely repeated the same error?

Of course, we can wash-rinse-repeat for any new samples that come in. Until we get tired or our money runs out or we get close enough to the characteristics of the actual population that we can quit.

 

[caveman1917]

Exactly.

So you concede the following?

Now, you seem to be saying the second sampling confirms my hypothesis in some way. I'm saying that it adds no new information, since one of the characteristics of my first sampling was that it represents the entire population fairly.

 

[jsfisher]

Wut? What does it even matter what P(E) is?

In your hypothetical, it matters a great deal.

Suppose there are 2 persons, Alice and Bob. We want to know whether these 2 persons have met, and our priors are equal P(H) = P(~H). Suppose Charlie comes forwards and says "I am their child", then...

...then you have an entirely different hypothetical. Here you started with two arbitrary people then added new information about them having a common offspring.

For the previous hypothetical, you started with the existing offspring and used that to identify the parents. That the offspring is an offspring is not new information at this point.

 

[aleCcowaN]

...Who knows? I'm going with another fringe reset. It's the only constant in this thread.

Fringe reset.... he always tries to get around the problem that "emergent property" in the scientific model is entirely dissimilar from the "immortal soul" of his model.

I'll never understand why all these epistemological hedonists, you know, the ones with sky daddies and mummies, nirvanas, blessings, immortal souls, houries, jebus pals, an the whole incense-myrrh-and-candle paraphernalia wouldn't stay at home believing more and even more and have to come up the mountain here to try and sink us in their ocean of doubts. Doubts they have because in their hearts they suspect a truth awful to them, hence unacceptable, I'd add.

If I were credulous I would find something diabolic in their pursues.

 

[abaddon]

I'll never understand why all these epistemological hedonists, you know, the ones with sky daddies and mummies, nirvanas, blessings, immortal souls, houries, jebus pals, an the whole incense-myrrh-and-candle paraphernalia wouldn't stay at home believing more and even more and have to come up the mountain here to try and sink us in their ocean of doubts. Doubts they have because in their hearts they suspect a truth awful to them, hence unacceptable, I'd add.

If I were credulous I would find something diabolic in their pursues.

It's because they don't believe it themselves. If they did, they simply would not visit venues such as this. No further argument would be desired or required.

They are not here to convince you, or me, or anyone. They are here to attempt to convince themselves. One can observe this in multiple threads on multiple topics by multiple posters.

So what of the skeptics? Well, largely the skeptics start that if you want to convince me of your claim, best you present some pretty good evidence for your claim. If, however, your objective is to convince you of your own claim, why are you here?

 

[caveman1917]

In your hypothetical, it matters a great deal.

No it doesn't, as long as it's not zero it completely cancels out of the equation.

...then you have an entirely different hypothetical. Here you started with two arbitrary people then added new information about them having a common offspring.

For the previous hypothetical, you started with the existing offspring and used that to identify the parents. That the offspring is an offspring is not new information at this point.

Can Charlie say "my existence is evidence for my parents, Alice and Bob, having met"? Yes or no?

 

[marplots]

So you concede the following?

It's not a concession, it's essential to my argument.