home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!dtix!darwin.sura.net!mips!think.com!news!columbus
- From: columbus@strident.think.com (Michael Weiss)
- Newsgroups: sci.math
- Subject: Re: You know, the integers (Or: how do we know anything?)
- Message-ID: <COLUMBUS.92Jul21155511@strident.think.com>
- Date: 21 Jul 92 19:55:11 GMT
- References: <1992Jul21.132554.152734@ns1.cc.lehigh.edu>
- <1992Jul21.163533.15492@galois.mit.edu>
- Organization: Thinking Machines Corporation, Cambridge MA, USA
- Lines: 47
- NNTP-Posting-Host: strident.think.com
- In-reply-to: jbaez@zermelo.mit.edu's message of 21 Jul 92 16:35:33 GMT
-
- In article <1992Jul21.163533.15492@galois.mit.edu> jbaez@zermelo.mit.edu
- (John C. Baez) writes:
-
- [discussion of impossibility of proving the consistency of ZF omitted]
-
- I used to crave certainty and this sort of thing bugged me. Now I'm
- fairly used to it -- as well as the fact (which once seemed disturbing)
- that no r.e. set of axioms about the integers will have a unique model
- up to elementary equivalence, and NO set of axioms about the integers
- will have a unique model up to isomorphism. It irritates Torkel when I
- say that this means we're not quite sure what the "real" integers ARE.
- I think the sense in which I mean this was most clearly indicated by my
- game show in which one tried to figure out which mathematician was using
- the real integers and which two were fakes.
-
- Well... you presumably believe that there are nonstandard models of Peano
- arithmetic (say) because you accept the sort of set-theoretic reasoning
- used to establish the Compactness theorem of first order logic.
-
- You may say you interpret this reasoning in a purely formal
- sense, as another game played on the checkerboard of ZFC. Or you may
- prefer a more hedonistic interpretation, where the symbols simply
- transcribe a symphony we hear directly with our mathematical ears.
-
- Either way, I assume you would accept the following two theorems
- of ZFC on the same philosophical level:
-
- (1) For any consistent extension T of Peano arithmetic, there are
- non-isomorphic models of T-- i.e., nonstandard models exist.
-
- (2) There are models of Peano arithmetic that can be imbedded in any
- model of Peano arithmetic, and any two such models are
- isomorphic.
-
- If I wanted to dot all the i's, I would define a category of models of PA
- and state the imbedding and isomorphism properties "canonically", but since
- we're among friends, I won't bother.
-
- While the proof of (2) is trivial, I consider it quite adequate
- justification for talking about the Real Integers. We can single out a
- minimal model of Peano arithmetic, and this model is unique up to
- isomorphism in a natural way. What more could one ask for?
-
- Of course, stepping "outside the system", one can see that the model isn't
- *really* unique. But if you can do enough set theory to construct
- nonstandard models, you can tell which are the honest-to-Kronecker
- integers, and which are the fakes.
-