home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky sci.math:17215 sci.philosophy.tech:4594
- Path: sparky!uunet!cis.ohio-state.edu!magnus.acs.ohio-state.edu!usenet.ins.cwru.edu!ukma!hsdndev!husc-news.harvard.edu!husc10.harvard.edu!zeleny
- From: zeleny@husc10.harvard.edu (Michael Zeleny)
- Newsgroups: sci.math,sci.philosophy.tech
- Subject: Re: Numbers and sets
- Message-ID: <1992Dec19.140927.18700@husc3.harvard.edu>
- Date: 19 Dec 92 19:09:25 GMT
- References: <1992Dec15.135030.18526@husc3.harvard.edu> <1992Dec15.214124.7444@guinness.idbsu.edu> <1992Dec17.235906.13828@guinness.idbsu.edu>
- Organization: The Phallogocentric Cabal
- Lines: 118
- Nntp-Posting-Host: husc10.harvard.edu
-
- In article <1992Dec17.235906.13828@guinness.idbsu.edu>
- holmes@opal.idbsu.edu (Randall Holmes) writes:
-
- >In article <1992Dec15.214124.7444@guinness.idbsu.edu>
- >holmes@opal.idbsu.edu (Randall Holmes) writes:
-
- Talking to yourself is a bad sign, Randall.
-
- >>I repent! I repent! No >>>>>>>>>>>>>>>>>!
- >>
- >>My final answer to Zeleny is that I accept his point in a modified
- >>form (which for him may not count as acceptance at all.) I think that
- >>the sets of Zermelo set theory and its extensions (with the Axiom of
- >>Foundation) are best understood as being related to isomorphism types
- >>of well-founded extensional relations. Insofar as ordinals (best
- >>understood as being related to isomorphism types of well-orderings)
- >>are a prominent and structurally significant special case of this kind
- >>of relation, I agree that the ordinals are essential to understanding
- >>ZF. This is sharpened when ZF is extended to ZFC by the addition of
- >>the axiom of choice. The particular approach of ZF makes it
- >>impossible to choose canonical objects to serve as cardinal numbers
- >>without using foundation or choice. The concept of cardinal number
- >>(although not the relations of having less or greater cardinality)
- >>depends on foundation or choice for its expression.
-
- This is the better part of my point, and I could not have said it
- more succinctly myself. Thank you for conceding this much of my claim.
-
- >>But the axioms which mediate this relation with the ordinals (or more
- >>generally with well-founded extensional relations) are strictly
- >>foundation and choice. ZF- is still a powerful set theory (it is no
- >>weaker than ZFC in terms of consistency strength). Its axioms of
- >>separation and replacement embody a concept of _set_ (avoiding paradox
- >>by application of the "limitation of size" doctrine) which is not
- >>dependent on the ordinals. ZF- is not studied much that I know of,
- >>but the situation is the same in ZFA, in which extensionality is
- >>weakened to allow atoms (I believe that Zermelo's original theory
- >>allowed atoms, which argues against Zeleny's claim about the history
- >>of the set idea!) without choice, and this theory is studied; it was
- >>the first theory for which models without choice were constructed.
- >>The iterative hierarchy is a powerful working hypothesis, but it is
- >>not a prerequisite for the modern concept of set, but a refinement
- >>thereof. Notice that it _does_ appear in ZFA or ZF- as the structure
- >>of _part_ of the universe.
-
- Zermelo referred to urelements in the 1906 correspondence with
- Poincar\'e; likewise, his 1908 and 1930 axiomatizations explicitly
- admitted their existence. Also, Fraenkel's 1922 independence proof of
- AC (reprinted in van Heijenoort), as well as subsequent independence
- results of Lindenbaum and Mostowski, used a technique which was to be
- superseded by forcing, and required the existence of a denumerably and
- non-denumerably many urelements, respectively. Notably, the relative
- consistency of NBG with urelements with respect to NBG was shown by
- Mostowski in 1939. However Fraenkel's own contemporary introduction of
- his Axiom of Restriction (which was construed as "inversely analogous"
- to Hilbert's model-maximality geometric Axiom of Completeness, and so
- intended to yield categoricity by stipulating that the only sets to
- exist were those implied by Zermelo's axioms and his own Axiom of
- Replacement), implicitly excluded the existence of urelements (see
- Moore's book on Zermelo, and especially Fraenkel _et alii_ on
- foundations). But of course this strong version of the limitation of
- size doctrine is utterly antithetical to the profligate spirit of ZFC,
- as embodied in Levy's reflexion principles, and thus would be only
- satisfactory to those misguided souls of the Quinean ilk, who would
- "round off smoothly" the set-theoretic ontology.
-
- Having been provoked to undertake this tedious disquisition by your
- parenthetical remark, I still cannot recall ever having claimed either
- that ZFC was originaly construed as a pure set theory, or that such
- historical considerations have any bearing whatsoever on the concept of
- set.
-
- >>By the way, Aczel's AFA is just as good as foundation; ZF- + AFA does
- >>have well-defined canonical cardinal numbers.
-
- True, but irrelevant: as Herb Enderton once observed in this venue, it
- is unclear whether Aczel's entities are actually *sets*, as this term is
- rightly and commonly understood. I would add that the same is true of
- the beasts populating the ontology of type-free theories of the
- Quine-Church ilk. The Axioms of Foundation and Choice are analytically
- true of sets; this is the crux of our philosophical disagreement, and I
- do not expect us to find a way to come to terms on its subject.
-
- >>A definition of "stage of the iterative hierarchy": let a "universe"
- >>be a set which contains all elements of its elements and all subsets
- >>of its elements. Then a "stage of the iterative hierarchy" is a
- >>universe which contains all its proper subsets which are universes.
- >>The correctness of this definition is left as an exercise, but note
- >>that it does not depend on the ordinals!
- >
- >Alas, this definition is incorrect. All sets which have this property
- >are stages, but not all stages have this property. I was working by
- >analogy with the definition of "ordinal" as "transitive set which
- >contains all its transitive proper subsets"; I'm sure there is a
- >refinement of this approach which does work, and I have another
- >definition on tap which is correct.
-
- Make up your mind, Randall.
-
- >>--
- >>The opinions expressed | --Sincerely,
- >>above are not the "official" | M. Randall Holmes
- >>opinions of any person | Math. Dept., Boise State Univ.
- >>or institution. | holmes@opal.idbsu.edu
- >
- >
- >--
- >The opinions expressed | --Sincerely,
- >above are not the "official" | M. Randall Holmes
- >opinions of any person | Math. Dept., Boise State Univ.
- >or institution. | holmes@opal.idbsu.edu
-
- What possesses the tab characters in your .sig file, that they do this
- strange dance on the occasions of follow-ups?
-
- cordially,
- mikhail zeleny@husc.harvard.edu
- "Le cul des femmes est monotone comme l'esprit des hommes."
-