4 ms·

I don't like these kind of arguments (see doomsday argument) because they do indeed have a deep logical flaw. They use definitions and mechanics from probabilit

I don't like these kind of arguments (see doomsday argument) because they do indeed have a deep logical flaw. They use definitions and mechanics from probability theory for phenomena that are not repeatable. For instance, my blood type is not a "likely" or "unlikely" choice out of all existing blood types. It is predetermined to a large degree by my parents. So is my country of birth. Of course, if you pick a human at random, or repeat the test with many different humans, the argument is valid again. But this shows you why the extrapolation to aliens does not work. We don't have any aliens in our sample set. Without postulating a quasi-randon mechanism that made you a human instead of an alien, this argument simply doesn't work. Or rather it doesn't have any meaning.

setr 5y ago
The point he’s making is that if you don’t know what the factors are, then your best bet is to pick from the larger group — you’ll be more likely to be correct. If aliens are sufficiently common, then we are much more likely than not to be a “normal” species. It is possible we are not normal, but that’s what a confidence interval is for.
- adwn 5y ago
  > The point he’s making is that if you don’t know what the factors are, then your best bet is to pick from the larger group — you’ll be more likely to be correct. If your best bet only has a miniscule probability of being right, then your best meta-bet is not to bet. At some point the correct conclusion is to say "We don't know", instead of fooling ourselves with empirically-unsupported navel gazing.
  - croon 5y ago
    "More likely than not" is not a "miniscule probability".
- choeger 5y ago
  You make a mistake in your argument. Some assumption being "more likely to be correct" has no meaning without a distribution that can be described as random. Without repeating the assumption multiple times the words "probability" or "likelihood" simply cannot be applied.
heisenzombie 5y ago
Even without “a quasi random process that made you a human” (which seems reasonable to me, but let’s put it aside for the moment) If every intelligent creature in the universe wrote down “I am on a high-population planet” and then we zipped around with a spaceship to pick them all up, we would find most of the pieces of paper to be correct.
- choeger 5y ago
  Your second paragraph is absolutely correct, yet it leads to a circular argument, because it does not tell you anything about how correct your paper is. In fact, in that experiment, the probability of your paper being among the correct ones is the same quantity as the actual population density. You cannot derive any meaningful information from assuming you are above,about,or below average.
croon 5y ago
Yes, we don't know whether ours specifically is a good sample planet or not, but statistically it is, and out of all the universe's planets' statisticians, most postulating this theory are.
- simiones 5y ago
  Exactly - so, the statistical argument tells us absolutely nothing about other aliens at all. With different distributions of aliens in the universe, we would be more or less likely to be wrong. You could also use this same argument to claim that most aliens speak English, Chinese or Indian; that they are Christian, Hindu or Muslim, that they have whitish skin, brown eyes, brown hair, 2 arms, 2 legs, 10 fingers, etc. In fact, according to this argument, aliens are quite likely to be humans in fact. In other words, the (correct) statistical argument that the article devotes its first half to is completely inconsequential. Only the second, much more argumentative side, the one that uses speculative biological and chemical knowledge, has any predictive power, and it is on very shaky ground given our extremely limited understanding of the possibilities of non-Earth life.

setr 5y ago

The point he’s making is that if you don’t know what the factors are, then your best bet is to pick from the larger group — you’ll be more likely to be correct. If aliens are sufficiently common, then we are much more likely than not to be a “normal” species. It is possible we are not normal, but that’s what a confidence interval is for.

adwn 5y ago
> The point he’s making is that if you don’t know what the factors are, then your best bet is to pick from the larger group — you’ll be more likely to be correct. If your best bet only has a miniscule probability of being right, then your best meta-bet is not to bet. At some point the correct conclusion is to say "We don't know", instead of fooling ourselves with empirically-unsupported navel gazing.
- croon 5y ago
  "More likely than not" is not a "miniscule probability".
choeger 5y ago
You make a mistake in your argument. Some assumption being "more likely to be correct" has no meaning without a distribution that can be described as random. Without repeating the assumption multiple times the words "probability" or "likelihood" simply cannot be applied.

heisenzombie 5y ago

Even without “a quasi random process that made you a human” (which seems reasonable to me, but let’s put it aside for the moment) If every intelligent creature in the universe wrote down “I am on a high-population planet” and then we zipped around with a spaceship to pick them all up, we would find most of the pieces of paper to be correct.

choeger 5y ago
Your second paragraph is absolutely correct, yet it leads to a circular argument, because it does not tell you anything about how correct your paper is. In fact, in that experiment, the probability of your paper being among the correct ones is the same quantity as the actual population density. You cannot derive any meaningful information from assuming you are above,about,or below average.

croon 5y ago

Yes, we don't know whether ours specifically is a good sample planet or not, but statistically it is, and out of all the universe's planets' statisticians, most postulating this theory are.

simiones 5y ago
Exactly - so, the statistical argument tells us absolutely nothing about other aliens at all. With different distributions of aliens in the universe, we would be more or less likely to be wrong. You could also use this same argument to claim that most aliens speak English, Chinese or Indian; that they are Christian, Hindu or Muslim, that they have whitish skin, brown eyes, brown hair, 2 arms, 2 legs, 10 fingers, etc. In fact, according to this argument, aliens are quite likely to be humans in fact. In other words, the (correct) statistical argument that the article devotes its first half to is completely inconsequential. Only the second, much more argumentative side, the one that uses speculative biological and chemical knowledge, has any predictive power, and it is on very shaky ground given our extremely limited understanding of the possibilities of non-Earth life.