1
FINITE PHASE TRANSITIONS
Harvey M. Friedman*
Ohio State University
September 26, 2010
10AM
DRAFT
This topic has been discussed earlier on the FOM email list
in various guises. The common theme is: big numbers and
long sequences associated with mathematical objects. See
http://www.cs.nyu.edu/pipermail/fom/1998July/001921.html
http://www.cs.nyu.edu/pipermail/fom/1998
October/002332.html
http://www.cs.nyu.edu/pipermail/fom/1998
October/002339.html
http://www.cs.nyu.edu/pipermail/fom/1998
October/002356.html
http://www.cs.nyu.edu/pipermail/fom/1998
October/002365.html
http://www.cs.nyu.edu/pipermail/fom/1998
October/002383.html
http://www.cs.nyu.edu/pipermail/fom/1998
October/002395.html
http://www.cs.nyu.edu/pipermail/fom/1998
October/002407.html
http://www.cs.nyu.edu/pipermail/fom/1998
October/002409.html
http://www.cs.nyu.edu/pipermail/fom/1998
October/002410.html
http://www.cs.nyu.edu/pipermail/fom/1998
November/002439.html
http://www.cs.nyu.edu/pipermail/fom/1998
November/002443.html
http://www.cs.nyu.edu/pipermail/fom/1999March/002752.html
http://www.cs.nyu.edu/pipermail/fom/1999May/003134.html
http://www.cs.nyu.edu/pipermail/fom/1999July/003251.html
http://www.cs.nyu.edu/pipermail/fom/1999July/003253.html
http://www.cs.nyu.edu/pipermail/fom/2006March/010292.html
http://www.cs.nyu.edu/pipermail/fom/2006March/010293.html
http://www.cs.nyu.edu/pipermail/fom/2006March/010290.html
http://www.cs.nyu.edu/pipermail/fom/2006March/010279.html
http://www.cs.nyu.edu/pipermail/fom/2006March/010281.html
http://www.cs.nyu.edu/pipermail/fom/2006April/010305.html
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document2
Suppose we have a
Π
0
2
theorem (
∀
k)(
∃
n)(A(k,n)). We then get
a recursive function F(k) = the least n such that A(k,n)
holds. In the intended cases, we have F(0) < F(1) < .
.. .
We want to look at F(0),F(1),.
.. and determine where there
is a "qualitative jump" in size. I.e., a "phase
transition". In the cases we focus on, after the first few
terms  say about 16 or less  we simply get qualitatively
indistinguishable very large integers.
There are a number of forms that such results can take.
Some are more descriptive than quantitative and others are
more quantitative than descriptive.
QUANTITATIVE APPROACH.
In the quantitative approach, we simply provide upper and
lower bounds on some of the terms in F(0),F(1),.
.. using a
notation system for integers. We try to make exact
calculations, if possible.
But what notation system to use for the integers?
In case the numbers involved are less than, say, 10
100
, the
usual base 10 notation is the clear choice.
But if the numbers involved are greater than, say, 10
10^100
,
base 10 notation is generally of no use whatsoever since it
cannot even be presented.
There are many special approaches that may be particularly
illuminating. However, we are led to the following general
approach. We identify a finite list of constants and basic
functions of one or more variables on the nonnegative
integers. We then use closed terms to name particular
nonnegative integers, which are used for lower and upper
bounds. Of course, we should strive to use small terms  or
natural terms  for this purpose.
This is the end of the preview.
Sign up
to
access the rest of the document.
 Fall '08
 JOSHUA
 Math, Recursion, Exponentiation, Natural number, Quantification, proof theoretic integer, theoretic integer

Click to edit the document details