3 ms·

This is addressed in the paper. They use Bayes to update their estimates based on the lack of evidence so far (which is, itself, evidence for lack of advanced

This is addressed in the paper. They use Bayes to update their estimates based on the lack of evidence so far (which is, itself, evidence for lack of advanced stellar civilizations). It has been dissolved in that they've pointed out a fundamental error in the logic that leads to the paradox, and have used statistics and a review of available literature to illustrate this error.

hannasanarion 8y ago
How could you possibly use bayesian updates to calculate the true priors based on only one sample? That's like saying "I have calculated that the sun will never rise again" after seeing it set for the first time in your life.
- thaumaturgy 8y ago
  Your analogy would more accurately be, "it's been dark all my life, so the sun's never going to rise again." The paper goes into this starting on page 12. After establishing estimates based on a variety of factors, they then adjust the results by pointing out that "there were no Type III Kardashev civilizations using more than 85% of the starlight in 10^5 surveyed galaxies" and that "the lack of observable settlement provides a strong bound on extraterrestrial intelligence". They proceed to offer approximations for Bayes conditional probabilities with different formulae for no stars having been observed to have civilizations, no civilizations within a detectable distance, and no settlements within nearby spacetime. After putting all of this together, they conclude in section 4.0 that, "The Fermi observation thus raises our 52% credence for being alone in the galaxy to somewhere between 53% and 99.6%, depending on the type of evidence considered. It likewise raises our 38% credence for being alone in the observable universe to somewhere between 39% and 85%." Essentially, they're saying that if the probability of detection is some value if there are intelligent civilizations in the observable universe, then how does it change the model that they haven't been observed? I am not a knowledgeable enough mathematician to judge whether this is an appropriate use of Bayes or not, but I would expect that someone would have already pointed it out if it weren't. They then conclude section 4.1 with, "...unless one assigns a very strong prior to [a systematic failure to detect observable civilizations], our qualitative conclusions still hold."
TheOtherHobbes 8y ago
Absence of evidence isn't evidence of absence - especially not when dealing with entities and civilisations about which we not only know nothing, but about whose properties we can barely speculate. The only thing Drake can possibly estimate is the number of civilisations similar enough to our own to be recognisable. There's an assumption that every civilisation will pass through a stage where can recognise it - but of course there's no good reason to be confident this is true. So in fact the L term is unknowable. There's also the fact that not all civs will transmit, but some would be visible to better spectroscopic tools. Generally, as our technology evolves you would expect our detection skill to increase too. This isn't included in the original equation.
- Jach 8y ago
  > Absence of evidence isn't evidence of absence Yes it is. You may be confusing 'evidence' for 'proof', but here's a proof that absence of evidence is indeed evidence of absence. (How much? That depends.) http://kim.oyhus.no/absence/index.html http://kim.oyhus.no/absence/index.html
- jeremyjh 8y ago
  Honey, did you leave the lights on downstairs? Well, I don't see any evidence of light, but that isn't evidence for the absence of light. So I guess I just can't know.
- ravenstine 8y ago
  > There's an assumption that every civilisation will pass through a stage where can recognise it - but of course there's no good reason to be confident this is true. That's the inherent problem with only having one example of something. For all we know, our current way of being is fundamentally closing off our perception that would allow us the knowledge that intelligence exists beyond our own. Even our own thought patterns could be fooling us. For example, we could very well visit a planet in another star system and be like... "Wow, this whole planet is covered in rust. What a wasteland! Let's get out of here." And we might not even think that the rust is actually an intelligence. We might figure out that it's living, after years of stomping on it or trying to mine base elements out of it, but we might never know that it's actually much more intelligent than we are because it doesn't have the innate human drive to spread out, conquer, explore, etc., all of which are not necessarily indications of intelligence.

hannasanarion 8y ago

How could you possibly use bayesian updates to calculate the true priors based on only one sample? That's like saying "I have calculated that the sun will never rise again" after seeing it set for the first time in your life.

thaumaturgy 8y ago
Your analogy would more accurately be, "it's been dark all my life, so the sun's never going to rise again." The paper goes into this starting on page 12. After establishing estimates based on a variety of factors, they then adjust the results by pointing out that "there were no Type III Kardashev civilizations using more than 85% of the starlight in 10^5 surveyed galaxies" and that "the lack of observable settlement provides a strong bound on extraterrestrial intelligence". They proceed to offer approximations for Bayes conditional probabilities with different formulae for no stars having been observed to have civilizations, no civilizations within a detectable distance, and no settlements within nearby spacetime. After putting all of this together, they conclude in section 4.0 that, "The Fermi observation thus raises our 52% credence for being alone in the galaxy to somewhere between 53% and 99.6%, depending on the type of evidence considered. It likewise raises our 38% credence for being alone in the observable universe to somewhere between 39% and 85%." Essentially, they're saying that if the probability of detection is some value if there are intelligent civilizations in the observable universe, then how does it change the model that they haven't been observed? I am not a knowledgeable enough mathematician to judge whether this is an appropriate use of Bayes or not, but I would expect that someone would have already pointed it out if it weren't. They then conclude section 4.1 with, "...unless one assigns a very strong prior to [a systematic failure to detect observable civilizations], our qualitative conclusions still hold."

TheOtherHobbes 8y ago

Absence of evidence isn't evidence of absence - especially not when dealing with entities and civilisations about which we not only know nothing, but about whose properties we can barely speculate. The only thing Drake can possibly estimate is the number of civilisations similar enough to our own to be recognisable. There's an assumption that every civilisation will pass through a stage where can recognise it - but of course there's no good reason to be confident this is true. So in fact the L term is unknowable. There's also the fact that not all civs will transmit, but some would be visible to better spectroscopic tools. Generally, as our technology evolves you would expect our detection skill to increase too. This isn't included in the original equation.

Jach 8y ago
> Absence of evidence isn't evidence of absence Yes it is. You may be confusing 'evidence' for 'proof', but here's a proof that absence of evidence is indeed evidence of absence. (How much? That depends.) http://kim.oyhus.no/absence/index.html http://kim.oyhus.no/absence/index.html
jeremyjh 8y ago
Honey, did you leave the lights on downstairs? Well, I don't see any evidence of light, but that isn't evidence for the absence of light. So I guess I just can't know.
ravenstine 8y ago
> There's an assumption that every civilisation will pass through a stage where can recognise it - but of course there's no good reason to be confident this is true. That's the inherent problem with only having one example of something. For all we know, our current way of being is fundamentally closing off our perception that would allow us the knowledge that intelligence exists beyond our own. Even our own thought patterns could be fooling us. For example, we could very well visit a planet in another star system and be like... "Wow, this whole planet is covered in rust. What a wasteland! Let's get out of here." And we might not even think that the rust is actually an intelligence. We might figure out that it's living, after years of stomping on it or trying to mine base elements out of it, but we might never know that it's actually much more intelligent than we are because it doesn't have the innate human drive to spread out, conquer, explore, etc., all of which are not necessarily indications of intelligence.