home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!europa.asd.contel.com!darwin.sura.net!jvnc.net!netnews.upenn.edu!sagi.wistar.upenn.edu
- From: weemba@sagi.wistar.upenn.edu (Matthew P Wiener)
- Newsgroups: sci.math
- Subject: Re: When's a proposition provable w/ choice indep. of choice?
- Keywords: Axiom of Choice, logic, set theory, intuitionism, constructionism, nonstandard analysis
- Message-ID: <97732@netnews.upenn.edu>
- Date: 15 Nov 92 18:50:27 GMT
- References: <1e3qvjINN3ul@iskut.ucs.ubc.ca>
- Sender: news@netnews.upenn.edu
- Reply-To: weemba@sagi.wistar.upenn.edu (Matthew P Wiener)
- Organization: The Wistar Institute of Anatomy and Biology
- Lines: 41
- Nntp-Posting-Host: sagi.wistar.upenn.edu
- In-reply-to: pruss@math.ubc.ca (Alexander Pruss)
-
- In article <1e3qvjINN3ul@iskut.ucs.ubc.ca>, pruss@math (Alexander Pruss) writes:
- > Are there any ways of judging whether a proposition is in some way
- >independent of the axiom of choice, without trying to prove/disprove it
- >without Choice? The application would be to take a proposition which is
- >proved with the axiom of choice, and be able to assert that it holds even
- >without the axiom of choice, without going through the proof.
-
- Yes--it's a notion called "absoluteness". The only case I've seen of
- the above was an invocation of model-theoretical results that depended
- on GCH, and the claim that the final result itself did not need it.
-
- >[...] A similar but stricter question is whether there is any guideline
- >for determining if a proposition of a certain form being provable in
- >ordinary set theory is thence constructively provable. [...]
-
- The easy result, which you realized yourself, is that all first order
- statements about the positive integers are absolute. (Meaning that
- all quantifiers are for integers.) A step beyond is the realization
- that an existential second-order quantifier over sets of the integers
- preserves absoluteness upwards. That is, given a purely first-order
- statement, quantified by "there exists a function f:N->N" or something
- like that, if its true in a small model, it will be true in a larger
- model.
-
- The canonical choice of "small model" is Goedel's constructible universe
- L. Since L models AC, anything proven in L that is absolute upward is
- true without assuming AC. This kind of diagram chasing, where truth and
- not elements is what gets kicked around, bugs some people, so it's worth
- realizing that all that's done is the search for the quantifier witness
- is done via the time-honored method of looking for a particular kind of
- witness. In this case, one that's part of a well-ordered universe.
-
- Things get exciting though. Shoenfield has an absoluteness theorem
- which state that the first non-obviosity here also works. If you have
- a statement "there exists f, for all g,h <1o arithmetic>", then it too
- will be absolute between L and V. So any use of choice is redundant
- for such a statement also.
-
- See Jech SET THEORY for the precise statement and proof.
- --
- -Matthew P Wiener (weemba@sagi.wistar.upenn.edu)
-