To figure out the odds of the existence of God based on you having a certain type of experience, you can’t just ignore the knowledge about what kind of brain you have.
The figures that I am using to find P(God | experience) include P(experience | no God) and P(experience | God). P(experience | no God) must surely integrate over every possible universe in which there is no God, or else it will be P(experience | a specific universe), which is the wrong figure. Similarly, P(experience | God) must integrate over every universe in which an omnipotent, omniscient being exists; or else, once again, it is P(experience | specific universe), and is the wrong figure.
I’m trying to simplify the equation, for purposes of debate, to an update on exactly one piece of evidence. If we start including the specifics of the brain, then that opens up the question of P(God | human brain), which is an entire, and much bigger, debate on its own. And one that I have no intention of entering into with you; the inferential distance between us is simply too large, and I’ve been having enough trouble trying to communicate with you as it is.
Yes, there would. If people normally disintegrate only when exposed to thousands of degrees of temperature, but God can disintegrate people whenever he wants, every time God stays hidden but disintegrates a person who is not hidden, that’s a seam in the rules and science will not be able to explain that event.
That’s a very good point.
It’s rather unlikely that it could be replicated under laboratory conditions, which would mean it’s not an obvious seam… but yes, I see your point, it could be a bit of a seam. (Or it could be as-yet undiscovered physics of some sort, of course).
I’m trying to simplify the equation, for purposes of debate, to an update on exactly one piece of evidence. If we start including the specifics of the brain, then that opens up the question of P(God | human brain), which is an entire, and much bigger, debate on its own.
You can’t simplify that equation. If you simplify the equation that way it’s only useful when figuring out if there is a god, given a creature of unknown type who has a mystical experience. You are not a creature of unknown type, and throwing out the information that you have a brain that is inordinately prone to malfunctions that produce mystical experiences will distort the answer.
It’s like concluding there is a high probability that someone is insane because he thinks he’s Napoleon, while discarding the information that he’s ruling France in the year 1805. P(insane|claim Napoleon) is high, but P(insane|claim Napoleon && rules France in 1805)” is not high. Discarding this information is wrong.
If you simplify the equation that way it’s only useful when figuring out if there is a god, given a creature of unknown type who has a mystical experience.
That is exactly the point I was trying to make; that the fact that such an experience happened is a piece of evidence in favour of the existence of an omnipotent, omniscient being.
You are not a creature of unknown type, and throwing out the information that you have a brain that is inordinately prone to malfunctions that produce mystical experiences will distort the answer.
The trouble with this, if you start insisting that certain information cannot be left out, is that you are cherry-picking what information that is, and specifically selecting information that you believe leads to the conclusion that you want. I could probably name a dozen other pieces of information which support the idea that an omniscient, omnipotent being exists and insist on including them too (for example; consider P(God | intelligent life exists)).
If we start down this path, then we could both simply throw new pieces of information into the discussion again and again; it will take months, and it would lead absolutely nowhere.
It’s like concluding there is a high probability that someone is insane because he thinks he’s Napoleon, while discarding the information that he’s ruling France in the year 1805. P(insane|claim Napoleon) is high, but P(insane|claim Napoleon && rules France in 1805)” is not high. Discarding this information is wrong.
I would like to refer you to a short story that discusses that point in greater detail.
In short, yes, P(insane | claims Napoleon) is high, and if I meet someone who tells me ‘One of my ancestors claimed he was Napoleon’ then I will consider it a high probability that that ancestor was insane.
That is exactly the point I was trying to make; that the fact that such an experience happened is a piece of evidence in favour of the existence of an omnipotent, omniscient being.
The fact that such an experience happened, given no other information, may be evidence for God. But you have other information and are ignoring it.
Furthermore, even as evidence for God, it’s only evidence in a very weak sense compared to typical cases where one speaks about something being evidence. It makes it more likely that there’s a God, but it also makes it more likely that there is a pagan god, or a sorcerer, or an extraterrestrial with a mind-control ray.
The trouble with this, if you start insisting that certain information cannot be left out, is that you are cherry-picking what information that is
It’s not cherry picking to refuse to discard information which affects the result by orders of magnitude more than the probability you’re hoping to get from it. A mystical experience in a human brain is much better evidence for a brain malfunction than it is for God.
In short, yes, P(insane | claims Napoleon) is high, and if I meet someone who tells me ‘One of my ancestors claimed he was Napoleon’
My scenario not only involves being in 1805, but also ruling France. If you use the information that someone is your ancestor (presumably from 1805), but you omit the information that he was ruling France then, you have still discarded relevant information and as a result will come up with a probability of insanity that is much too high.
The fact that such an experience happened, given no other information, may be evidence for God. But you have other information and are ignoring it.
sigh
Very well. If you insist that I not ignore information, then there is a lot more information that needs to be properly considered. Decades of experiences that need to be sorted and categorised. Hundreds of writings, ancient and modern; thousands of accounts, of one sort or another.
We’ll just have to go through it all, piece by piece. Only the bits that are relevant, of course, but there’s a lot that is relevant.
For a start, let me consider the probability of the existence of God, given the existence of brains (or any form of hardware capable of supporting intelligence in some or other form).
Where, just to be clear, I use the word ‘God’ to refer to any omniscient, omnipotent being.
For that, we need to consider the probability that a randomly chosen universe will contain brains, at some or other point in its history.
In order to do that, let us consider all possible universes.
Here, I start by considering the null universe; a universe that contains no matter. Since it contains no matter, it cannot contain brains.
Then, let us add a piece of matter; a single quark. this one-quark universe also cannot contain brains, as there is insufficient matter to form the brains.
I don’t think it’s sensible to consider different one-quark universes as different; no matter where the quark is, or how fast it is moving, it’s just a coordinate transform to make it identical to any other one-quark universe.
Then, let us add a second quark. Now, it is sensible to differentiate between different two-quark universes, because there is a measurement that can change from one universe to another; and that is the distance between the two quarks. There is also another measurement that can change from one universe to another, and that is their relative velocity. However, it is not always clear whether two diffferent snapshots of two-quark universes are different universes, or the same universe at different times. Still, no brains are possible.
There are even more three-quark than two-quark universes; at a rough estimate, I’d say that the number of three-quark universes is half of the square of the number of two-quark universes. Still no brains.
Four-quark universes; here, I expect the total number of universes to be roughly x^3/6 where x is the number of two-quark universes.
For an N-quark universe, in general, I expect that the total number of possible universes would be roughly x^(N-1)/(N-1)!. This is an interesting function; it grows exponentially for low N, the rate of increase slows as N approaches x, and finally, when N exceeds x, the number of possible N-particle universes actually begins to drop. Of course, x is either infinite or at least very very large, so we can expect that most universes will have quite a lot of matter; more than enough to form brains.
Now, what are the odds of a universe which contains enough matter containing brains? Note that the universes under consideration here contain a number of quarks, in what is essentially a random configuration. All possible configurations are under consideration here, so at least some of them will have brains (in some cases, Boltzman brains; in other cases, brains with bodies and whole civilisations around them). However, since every possible random configuration is permitted, it would seem to me that a highly ordered set of particles, like a brain, must be relatively rare.
So, P(Brain | no God) seems like it should be fairly low.
Now, let us consider P(Brain | God). This time, we do not have to consider every possible universe; only those universes that could exist in the presence of an omniscient, omnipotent being. It seems likely that such a being will adjust any universe to His liking, and possibly create one if one does not exist.
It also seems quite probable, to me, that if a single intelligent being exists, then that being is quite likely to realise that he (or she) would like a conversation with someone else. Being omnipotent, it seems likely that God would create someone to talk with, and thus create brains.
So, it seems likely that P(Brains | God) is high.
Putting a high value for P(Brains | God) and a low value for P(Brains | no God) into the equation from earlier:
...will produce the result that P(God | Brains) > P(God). This seems to be a fairly major effect.
A mystical experience in a human brain is much better evidence for a brain malfunction than it is for God.
Do you have any figures for the likelihood of such a brain malfunction? Or any form of data at all to back up this assertion?
My scenario not only involves being in 1805, but also ruling France.
Yes. That is quite a lot of bits of information; since it eliminates all but one of the people throughout all of history who thought they were Napoleon.
It’s a bit like the difference between asking “what are the odds of getting a six on a fair die roll?” and “what are the odds that I got a six on the fair die I rolled on Thursday 10 April 2014 at 21:10?”—though the difference is a bit more pronounced, as I’m sure there were more than six people throughout history who claimed to be Napoleon.
The figures that I am using to find P(God | experience) include P(experience | no God) and P(experience | God). P(experience | no God) must surely integrate over every possible universe in which there is no God, or else it will be P(experience | a specific universe), which is the wrong figure. Similarly, P(experience | God) must integrate over every universe in which an omnipotent, omniscient being exists; or else, once again, it is P(experience | specific universe), and is the wrong figure.
I’m trying to simplify the equation, for purposes of debate, to an update on exactly one piece of evidence. If we start including the specifics of the brain, then that opens up the question of P(God | human brain), which is an entire, and much bigger, debate on its own. And one that I have no intention of entering into with you; the inferential distance between us is simply too large, and I’ve been having enough trouble trying to communicate with you as it is.
That’s a very good point.
It’s rather unlikely that it could be replicated under laboratory conditions, which would mean it’s not an obvious seam… but yes, I see your point, it could be a bit of a seam. (Or it could be as-yet undiscovered physics of some sort, of course).
You can’t simplify that equation. If you simplify the equation that way it’s only useful when figuring out if there is a god, given a creature of unknown type who has a mystical experience. You are not a creature of unknown type, and throwing out the information that you have a brain that is inordinately prone to malfunctions that produce mystical experiences will distort the answer.
It’s like concluding there is a high probability that someone is insane because he thinks he’s Napoleon, while discarding the information that he’s ruling France in the year 1805. P(insane|claim Napoleon) is high, but P(insane|claim Napoleon && rules France in 1805)” is not high. Discarding this information is wrong.
That is exactly the point I was trying to make; that the fact that such an experience happened is a piece of evidence in favour of the existence of an omnipotent, omniscient being.
The trouble with this, if you start insisting that certain information cannot be left out, is that you are cherry-picking what information that is, and specifically selecting information that you believe leads to the conclusion that you want. I could probably name a dozen other pieces of information which support the idea that an omniscient, omnipotent being exists and insist on including them too (for example; consider P(God | intelligent life exists)).
If we start down this path, then we could both simply throw new pieces of information into the discussion again and again; it will take months, and it would lead absolutely nowhere.
I would like to refer you to a short story that discusses that point in greater detail.
In short, yes, P(insane | claims Napoleon) is high, and if I meet someone who tells me ‘One of my ancestors claimed he was Napoleon’ then I will consider it a high probability that that ancestor was insane.
The fact that such an experience happened, given no other information, may be evidence for God. But you have other information and are ignoring it.
Furthermore, even as evidence for God, it’s only evidence in a very weak sense compared to typical cases where one speaks about something being evidence. It makes it more likely that there’s a God, but it also makes it more likely that there is a pagan god, or a sorcerer, or an extraterrestrial with a mind-control ray.
It’s not cherry picking to refuse to discard information which affects the result by orders of magnitude more than the probability you’re hoping to get from it. A mystical experience in a human brain is much better evidence for a brain malfunction than it is for God.
My scenario not only involves being in 1805, but also ruling France. If you use the information that someone is your ancestor (presumably from 1805), but you omit the information that he was ruling France then, you have still discarded relevant information and as a result will come up with a probability of insanity that is much too high.
sigh
Very well. If you insist that I not ignore information, then there is a lot more information that needs to be properly considered. Decades of experiences that need to be sorted and categorised. Hundreds of writings, ancient and modern; thousands of accounts, of one sort or another.
We’ll just have to go through it all, piece by piece. Only the bits that are relevant, of course, but there’s a lot that is relevant.
For a start, let me consider the probability of the existence of God, given the existence of brains (or any form of hardware capable of supporting intelligence in some or other form).
Where, just to be clear, I use the word ‘God’ to refer to any omniscient, omnipotent being.
So, what is P(God | Brains)?
According to Bayes’ Theorem:
P(God | Brains) = P(Brains | God)*P(God)/P(Brains)
Of course, P(Brains) = P(Brains | God)P(God)+P(Brains | no God)P(no God).
Thus, a simple substitution gives:
P(God | Brains) = P(Brains | God)P(God)/[P(Brains | God)P(God)+P(Brains | no God)*P(no God)]
So. What is P(Brains | no God)?
For that, we need to consider the probability that a randomly chosen universe will contain brains, at some or other point in its history.
In order to do that, let us consider all possible universes.
Here, I start by considering the null universe; a universe that contains no matter. Since it contains no matter, it cannot contain brains.
Then, let us add a piece of matter; a single quark. this one-quark universe also cannot contain brains, as there is insufficient matter to form the brains.
I don’t think it’s sensible to consider different one-quark universes as different; no matter where the quark is, or how fast it is moving, it’s just a coordinate transform to make it identical to any other one-quark universe.
Then, let us add a second quark. Now, it is sensible to differentiate between different two-quark universes, because there is a measurement that can change from one universe to another; and that is the distance between the two quarks. There is also another measurement that can change from one universe to another, and that is their relative velocity. However, it is not always clear whether two diffferent snapshots of two-quark universes are different universes, or the same universe at different times. Still, no brains are possible.
There are even more three-quark than two-quark universes; at a rough estimate, I’d say that the number of three-quark universes is half of the square of the number of two-quark universes. Still no brains.
Four-quark universes; here, I expect the total number of universes to be roughly x^3/6 where x is the number of two-quark universes.
For an N-quark universe, in general, I expect that the total number of possible universes would be roughly x^(N-1)/(N-1)!. This is an interesting function; it grows exponentially for low N, the rate of increase slows as N approaches x, and finally, when N exceeds x, the number of possible N-particle universes actually begins to drop. Of course, x is either infinite or at least very very large, so we can expect that most universes will have quite a lot of matter; more than enough to form brains.
Now, what are the odds of a universe which contains enough matter containing brains? Note that the universes under consideration here contain a number of quarks, in what is essentially a random configuration. All possible configurations are under consideration here, so at least some of them will have brains (in some cases, Boltzman brains; in other cases, brains with bodies and whole civilisations around them). However, since every possible random configuration is permitted, it would seem to me that a highly ordered set of particles, like a brain, must be relatively rare.
So, P(Brain | no God) seems like it should be fairly low.
Now, let us consider P(Brain | God). This time, we do not have to consider every possible universe; only those universes that could exist in the presence of an omniscient, omnipotent being. It seems likely that such a being will adjust any universe to His liking, and possibly create one if one does not exist.
It also seems quite probable, to me, that if a single intelligent being exists, then that being is quite likely to realise that he (or she) would like a conversation with someone else. Being omnipotent, it seems likely that God would create someone to talk with, and thus create brains.
So, it seems likely that P(Brains | God) is high.
Putting a high value for P(Brains | God) and a low value for P(Brains | no God) into the equation from earlier:
P(God | Brains) = P(Brains | God)P(God)/[P(Brains | God)P(God)+P(Brains | no God)*P(no God)]
...will produce the result that P(God | Brains) > P(God). This seems to be a fairly major effect.
Do you have any figures for the likelihood of such a brain malfunction? Or any form of data at all to back up this assertion?
Yes. That is quite a lot of bits of information; since it eliminates all but one of the people throughout all of history who thought they were Napoleon.
It’s a bit like the difference between asking “what are the odds of getting a six on a fair die roll?” and “what are the odds that I got a six on the fair die I rolled on Thursday 10 April 2014 at 21:10?”—though the difference is a bit more pronounced, as I’m sure there were more than six people throughout history who claimed to be Napoleon.