Talk:Bourbaki–Witt theorem

If someone could elaborate on the applications of Bourbaki-Witt to computer science, along with any others it might have, it would be much appreciated. I don't know all that much about it other than the basic set theoretic stuff.


 * There is a little about this at Knaster-Tarski theorem. Usually things like that go on in domain theory. Charles Matthews 19:41, 23 Sep 2004 (UTC)

The analysis of the countable special case doesn't seem quite right to me. In particular, surely "If X is countable then a similar argument, along with the completeness of X, shows that the sequence {xn} has a least upper bound x∞, which is a fixed point of f. It is helpful to think of x∞ as the limit of the sequence." doesn't always hold? e.g. if X is the ordinal ω+ω+1, and f is the successor function on ω+ω and the identity on the last 1, then the initial sequence {xn} is just {1,2,3,…}, with limit x∞ = ω, which isn't yet a fixed point. So in the countable special case, we may still need to go to transfinite iterates of f (and indeed we may need any countable order-type).

If this isn't disputed, I'll remove that line, and change the section to be just the finite special case. Pit-trout (talk) 14:10, 26 August 2009 (UTC)

The "proof" given on this page is **not** the one in the two papers that it discusses. I would suggest **reading** those two papers, because they contain a much neater proof. (In fact, it's not even a "proof", because it doesn't cite the Recursion Theorem due to John von Neumann that justifies transfinite recursion, or Hartogs' Lemma that tells you when to stop.) Also, there's now an intuitionistic proof of the result, due to Dito Pataraia. See http://mathoverflow.net/questions/362038 for more details of these comments.

For my own work related to this topic, see http://paultaylor.eu/ordinals

I have forgotten my Wikipedia login.

46.33.143.125 (talk) 08:59, 25 July 2021 (UTC)Paul Taylor