Pi01*113005

# Pi01*113005 - 1 This is an MS Word formatted version of my...

This preview shows pages 1–4. Sign up to view the full content.

1 This is an MS Word formatted version of my FOM posting #258 of November 27, 2005, at http://www.cs.nyu.edu/pipermail/fom/2005- November/009395.html Quite a number of typographical errors in the FOM posting have been corrected. P 0 1 INCOMPLETENESS by Harvey M. Friedman November 30, 2005 "Beautiful" is a word used by mathematicians with a semi rigorous meaning. We give "arguably beautiful" explicitly Pi01 sentences independent of ZFC. See Proposition A from section 1 and Proposition B from section 2, and variants. Proposition A is simplest and has a graph theoretic flavor, with algebraic overtones. Proposition B has a fixed point flavor. 1. P 0 1 INDEPENDENT STATEMENTS USING ANTICHAINS. We use [1,n] for the discrete interval {1,. ..,n}. Let A [1,n] k . We write A’ = [1,n] k \A. This treats [1,n] k as the ambient space. Let R [1,n] 2k . We define RA = R[A] = {y [1,n] k : ( \$ x A)(R(x,y))}. We say that R is strictly dominating if and only if for all x,y [1,n] k , if R(x,y) then max(x) < max(y). We start with the basic ‘complementation theorem for RA’. THEOREM 1.1. For all k,n 1 and strictly dominating R [1,n] 2k , there exists A [1,n] k such that RA = A’. Furthermore, A [1,n] k is unique. For A [1,n] k and t 1, we write A\t = {x A: t is not a

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
2 coordinate of x} = "A with t omitted". Here is a modification of Theorem 1.1 which we call the ‘complementation theorem for R[A\(8k)!]’. THEOREM 1.2. For all k,n 1 and strictly dominating R [1,n] 2k , there exists A [1,n] k such that R[A\(8k)!] = A’. Furthermore, A [1,n] k is unique. We now incorporate the antichain concept. Let R [1,n] 2k . We say that A is an antichain for R if and only if A [1,n] k and A,RA are disjoint. Of course, we have the following "maximal antichain" theorem. THEOREM 1.3. For all k,n 1, every R [1,n] 2k has a maximal antichain. Note that Theorem 1.3 is virtually contentless, since we are in a finite context where every nonempty class has a maximal element. Note that Theorem 1.1 provides a much stronger kind of antichain, which we call a ‘complete’ antichain. We refer to the following as a ‘complete antichain’ theorem. THEOREM 1.4. For all k,n 1, every strictly dominating R [1,n] 2k has an antichain A such that RA contains A’. Furthermore, A is unique. However, the analog of Theorem 1.4 for R[A\(8k)!] is false. THEOREM 1.5. The following is false. For all k,n 1, every strictly dominating R [1,n] 2k has an antichain A such that R[A\(8k)!] contains A’. We will weaken the conclusion in a simple way. Our development depends heavily on a very strong regularity condition on R. We say that R [1,n] k is order invariant if and only if for all x,y in [1,n] k of the same order type, R(x) iff R(y). The number of such R is bounded by an exponential expression in k that does not depend on n.
3 The imposition of order invariance is still not sufficient: THEOREM 1.6. The following is false. For all k,n 1, every strictly dominating order invariant R [1,n] 2k has an antichain A such that R[A\(8k)!] contains A’.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

## Pi01*113005 - 1 This is an MS Word formatted version of my...

This preview shows document pages 1 - 4. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online