home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math
- Path: sparky!uunet!mcsun!sunic!sics.se!sics.se!torkel
- From: torkel@sics.se (Torkel Franzen)
- Subject: Re: Expansion of set theory
- In-Reply-To: boshuck@triples.math.mcgill.ca's message of Thu, 13 Aug 92 17:44:08 GMT
- Message-ID: <1992Aug13.190645.19211@sics.se>
- Sender: news@sics.se
- Organization: Swedish Institute of Computer Science, Kista
- References: <1992Aug7.002122.24601@access.usask.ca>
- <1992Aug12.164408.10547@thunder.mcrcim.mcgill.edu>
- <1992Aug12.182912.4659@sics.se>
- <1992Aug13.174408.10574@thunder.mcrcim.mcgill.edu>
- Date: Thu, 13 Aug 1992 19:06:45 GMT
- Lines: 30
-
- In article <1992Aug13.174408.10574@thunder.mcrcim.mcgill.edu> boshuck@triples.
- math.mcgill.ca (William Boshuck) writes:
-
- >On the other
- >hand, I don't quite know what theorems you may be talking
- >about. A big theorem in this area, that there is no uniform
- >algorithm to decide whether a given Diophantine equation has
- >solutions (due to Matiesevitz (spelling?) I think) doesn't
- >use any aspects of the theory of large cardinals (an easy
- >presentation of this theorem can be found in Bell and
- >Machover's book on mathematical logic).
-
- I had in mind the following application of the theorem of Matiyasevic
- (which built on work by Davies, Putnam, and Robinson). The theorem states
- that every recursively enumerable set is Diophantine, i.e. is equal to
- the set of k such that the equation P(k,x1,..xn)=0 has a solution, for some
- integer polynomial P. This is very remarkable. The theorem is, as you note,
- elementarily provable. This means that an ordinary consistency statement -
- say "ZF is consistent" - is elementarily equivalent to the statement that
- a certain Diophantine equation has no solution. Since the consistency
- statement is provable e.g. in ZF+"there is an inaccessible cardinal", the
- assumption that there is an inaccessible cardinal has consequences regarding
- the unsolvability of Diophantine equations that are not provable in ZF
- itself.
-
- For a given Diophantine equation - e.g. Fermat's - this of course tells
- us nothing. But it is a striking fact concerning the relation between
- "concrete" and "metaphysical" mathematics that should be more widely known
- than it is.
-
-