RE: LOGICAL FALLACIES [John Derbyshire]
They missed the "numerators without denominators" fallacy, which you meet a lot when you argue science points.
It usually proceeds by demonstrating that the de novo probability of event X happening is one in ten to the power 68 (or some such number—this is the numerator). Therefore it can't possibly have happened.
The fallacy is that there were ten to the power 68 possible events that might have happened (that's the denominator), and ONE OF THEM HAD TO.
If you think of shuffling a deck of cards, looking at the order of cards you end up with, and contemplating the extremely tiny "advance" probability that you would have ended up with that particular order, you'll see the point. Yet the Ns without Ds fallacy is surprisingly popular.
It usually proceeds by demonstrating that the de novo probability of event X happening is one in ten to the power 68 (or some such number—this is the numerator). Therefore it can't possibly have happened.
The fallacy is that there were ten to the power 68 possible events that might have happened (that's the denominator), and ONE OF THEM HAD TO.
If you think of shuffling a deck of cards, looking at the order of cards you end up with, and contemplating the extremely tiny "advance" probability that you would have ended up with that particular order, you'll see the point. Yet the Ns without Ds fallacy is surprisingly popular.
No comments:
Post a Comment