Originally posted by royalchickenI almost certainly agree. But you have to remember that, given any irregular uncertainty which is in itself unlikely to carry a probability of exactly 0.5 (even chance), the likelyhood she is incorrect stands limited only by, though not with certain negligibility, the positive indifference between whether the fact is likely to be true or, in this case, unlikely to be probably uncertain.
Even if she is right with probability 1, we needn't be certain she's right unless we know about other conditions. For example, if you ignored the argument about whether such a thing is algorithmically possible and asked me to choose a random positive integer, it would, with probability 1, not be a power of 2, but you certainly can't be certain that I w ...[text shortened]... ent to her being certainly right if there were a finite number of spellings of 'Knutborn'.
It remains only possible to conjecture whether variant spellings of "Knutborne" may be assigned any meaningful likelyhood within the bounds of geographical uncertainty assigned a probability of less than 1.