Talk:Diaconescu's theorem
Mathematics Start‑class Low‑priority | ||||||||||
|
In the second formula, I changed "{0} to {1} if not P". Assuming U refers to the first set and V to the second, "not P" implies "x = 1" which rules out "V = {0}". MichaelShoemaker (talk) 21:22, 17 July 2009 (UTC)
I changed it back. On rereading, my interpretation was not correct. MichaelShoemaker (talk) 21:25, 17 July 2009 (UTC)
This proof is a strawman argument
IMHO, the correct intepretation of the AC is: given a family of sets there exists a function such that . It is a much more reasonable interpretation: whenever we have choice involved in practice, we don't just have sets, we have sets in certain roles, here represented by the indices.
This way, if we took then would not imply , i.e. we could still make different choices, and the truth of would not be recoverable.
The proof weasels in the otherwise completely unwarranted step by having the choice function act on the values and not on the roles.