How many times would a coin have to come up heads (if there were some way for it to test this) before there would be a chance you wouldn’t defy the data in a case like this? If you saw 20 heads in a row, would you expect more of them? Or 40?
I’ve seen heads come up about ten times in a row… with a fair coin and with full confidence that it’d continue to come up heads for as long as the coin-tosser wanted it to.
He’d learned how to time the number of flips in the air and catch it at just the right time.
Therefore, seeing heads come up any number of times would be absolutely zero evidence of magic for me—though it would count for loud decibels of “coolness factor”.
Therefore, seeing heads come up any number of times would be absolutely zero evidence of magic for me—though it would count for loud decibels of “coolness factor”. [emphasis added]
Almost nothing gives “absolutely zero evidence” and indefinite feats of extreme dexterity definitely aren’t exceptions.
Evidence smaller than a human brain can process—where counting it as evidence would cause a brain to overestimate the evidence’s value—seems like it would be a case where it’s practical to consider it absolutely zero evidence.
Once you have redefined “absolutely zero evidence” to mean “too small for it to be worthwhile for humans to consider it” precisely what language can you use to describe things that, you know, aren’t evidence?
No, throwing in “absolutely zero” here changes it from “not technically true” to simply muddled thinking about how evidence works.
Incidentally even for human purposes this counts as evidence. Not sufficient evidence to even consider the magic hypothesis outright. But if a consistent trend occurs with respect to coin tossing then that hypothesis must eventually be considered. For example if anyone who says ‘abracadabra’ is suddenly able to rig coin tosses reliably even if they could barely even toss a coin at all before then that is clear evidence that something weird is going on. If you investigate the phenonemon 10,000 times with randomly selected 7 year olds (or even adults) and it happens each time then is strong evidence that something that can be described as magic is occurring. And every single instance is obviously weak evidence. “Absolute zero” is just way off.
Once you have redefined “absolutely zero evidence” to mean “too small for it to be worthwhile for humans to consider it” precisely what language can you use to describe things that, you know, aren’t evidence?
True. The balancing concern is that, in the case of “too small to be worthwhile”, we want people to not consider it, and there are strong biases that make people consider it a labut there’s still a chance!If anything less than “absolutely zero evidence” lets people consider evidence—and overestimating the evidence is worse than underestimating it—then it is preferred language. In that case, it would share the same language as things that aren’t evidence at all.
The balancing concern is that, in the case of “too small to be worthwhile”, we want people to not consider it, and there are strong biases that make people consider it a la but there’s still a chance! If anything less than “absolutely zero evidence” lets people consider evidence—and overestimating the evidence is worse than underestimating it—then it is preferred language.
Controlling behavior by removing the ability to express concepts. Lets call this new language ‘newspeak’. :)
If newspeak increases my utility in dealing with certain agents, I desire to speak in newspeak to them. If newspeak decreases my utility in dealing with certain agents, I desire to speak normally to them. Let me not become attached to the perception of newspeak as intrinsically evil.
I too do all sorts of things with other agents that do not necessarily involve accurate communication. Violence and dirty talk spring to mind as a couple of examples.
Communication transfers ideas from one person to another. If technically correct communication transfers false ideas, it is deception. Accurate communication transfers correct ideas with high fidelity, which isn’t necessarily equivalent to technically correct communication.
How about “less evidence than X”, where X is something ridiculously silly? It seems like it might have the same psychological effects, in which case it would be preferable.
I feel strangely certain that if we started countering “my coin flips coming up mostly heads is evidence of my magic!” with “less evidence for magic than your capacity to continue breathing is”, the kind of person in question would simply write books about how breathing is obviously magical in nature.
Eliezer, exactly how many decibels of evidence would it require to persuade you that there is magic in the universe?
For example, see this claim of magic: http://www.clairval.com/lettres/en/2006/12/08/2061206.htm
How many times would a coin have to come up heads (if there were some way for it to test this) before there would be a chance you wouldn’t defy the data in a case like this? If you saw 20 heads in a row, would you expect more of them? Or 40?
I’ve seen heads come up about ten times in a row… with a fair coin and with full confidence that it’d continue to come up heads for as long as the coin-tosser wanted it to.
He’d learned how to time the number of flips in the air and catch it at just the right time.
Therefore, seeing heads come up any number of times would be absolutely zero evidence of magic for me—though it would count for loud decibels of “coolness factor”.
Almost nothing gives “absolutely zero evidence” and indefinite feats of extreme dexterity definitely aren’t exceptions.
Evidence smaller than a human brain can process—where counting it as evidence would cause a brain to overestimate the evidence’s value—seems like it would be a case where it’s practical to consider it absolutely zero evidence.
Once you have redefined “absolutely zero evidence” to mean “too small for it to be worthwhile for humans to consider it” precisely what language can you use to describe things that, you know, aren’t evidence?
No, throwing in “absolutely zero” here changes it from “not technically true” to simply muddled thinking about how evidence works.
Incidentally even for human purposes this counts as evidence. Not sufficient evidence to even consider the magic hypothesis outright. But if a consistent trend occurs with respect to coin tossing then that hypothesis must eventually be considered. For example if anyone who says ‘abracadabra’ is suddenly able to rig coin tosses reliably even if they could barely even toss a coin at all before then that is clear evidence that something weird is going on. If you investigate the phenonemon 10,000 times with randomly selected 7 year olds (or even adults) and it happens each time then is strong evidence that something that can be described as magic is occurring. And every single instance is obviously weak evidence. “Absolute zero” is just way off.
True. The balancing concern is that, in the case of “too small to be worthwhile”, we want people to not consider it, and there are strong biases that make people consider it a la but there’s still a chance! If anything less than “absolutely zero evidence” lets people consider evidence—and overestimating the evidence is worse than underestimating it—then it is preferred language. In that case, it would share the same language as things that aren’t evidence at all.
Controlling behavior by removing the ability to express concepts. Lets call this new language ‘newspeak’. :)
If newspeak increases my utility in dealing with certain agents, I desire to speak in newspeak to them. If newspeak decreases my utility in dealing with certain agents, I desire to speak normally to them. Let me not become attached to the perception of newspeak as intrinsically evil.
I too do all sorts of things with other agents that do not necessarily involve accurate communication. Violence and dirty talk spring to mind as a couple of examples.
Communication transfers ideas from one person to another. If technically correct communication transfers false ideas, it is deception. Accurate communication transfers correct ideas with high fidelity, which isn’t necessarily equivalent to technically correct communication.
How about “less evidence than X”, where X is something ridiculously silly? It seems like it might have the same psychological effects, in which case it would be preferable.
I feel strangely certain that if we started countering “my coin flips coming up mostly heads is evidence of my magic!” with “less evidence for magic than your capacity to continue breathing is”, the kind of person in question would simply write books about how breathing is obviously magical in nature.
Yes, I agree with this. I’d count it as “so negligible as to be beneath bothering to waste a neuron on it” :)
The evidence here isn’t actually negligible, I don’t think. It’s just that the posterior probability is negligible.
That’s true too—eventually he’d tire of it. Though I’m guessing I’d tire of watching long before then.