Re: Another Reason Why Collatz is Unprovable


Craig Feinstein wrote:
T.H. Ray wrote:


Hilbert's ideas have had a great influence on
modern
mathematics, even
though they are wrong.<<

Wrong? I can only assume you mean the formalist
program that Hilbert (and Frege, Russell and
Whitehead
among others) meant to pursue before Godel
published
his proof.

Yes, that is what I was talking about. Hilbert's
formalist program only
produces facts which are "trivial", i.e., follow
logically from
well-accepted axioms. Because of this, modern
mathematics research has
become dull, like some kind of sophisticated logic
puzzle. It's never
going to produce anything surprising, as nothing
surprising ever comes
from logic. When one takes mathematics to a higher
level, trying to
understand numbers as they are from computational
experiments, one
finds that mathematics is interesting and one sees
things which are
surprising. (In other words, math is interesting;
logic is
uninteresting.) Wolfram's way of doing it in his book
NKS is ultimately
how mathematics should be done in the computer age,
in my humble
opinion. Another person who also has some interesting
writing on this
subject is Rudy Rucker.<<

Do you think logicians find their research uninteresting?

Let's put it this way. Doing logic puzzles (which is basically what
logicians do with their time, although their puzzles are much more
sophisticated than the recreational variety) is a challenging and
enjoyable activity for many people. But it's never going to produce
anything surprising, since you are basically processing and rehashing
information that you already have.

Computational experimental mathematics *does* produce things which are
surprising, for instance fractals. For "most" computational
experiments, you can't predict where the computations will lead without
actually performing the computations themselves. (This is Rudy Rucker's
reformulation of Wolfram's principle of computational equivalence.)

Collatz is an example of this with a twist: In this case, the
computations always seem to lead to one, yet it is still impossible to
predict this with absolute certainty without performing the
computations themselves.

Absolutely wrong. In my paper

Blueprint for Failure:
How to Construct a Counterexample to the Collatz Conjecture

which has just been published at

<http://home.versatel.nl/galien8> see Volume 5 (2006)

I demonstrate that the list of counterexamples based on
Factor Congruence is empty for any 3n+C system where
C is a power of 3 (including C=1).

Although this does not prove the conjecture, it is an example
of predicting with absolute certainty that does NOT involve
performing the computations themselves. It is an example of
a shortcut which you claim does not exist.

Your whole paper is just a house of cards and I've just
pulled the tablecloth out from under it.

Now, that is surprising! A mathematics
discovery for which there is no logical explanation!


Godel, of course, was a logician. Tarski, as well.

If one thinks of mathematical logic as simply a set
of formalist rules, one might as well think of
arithmetic as accounting.

Computing doesn't hold much fascination for me, though
I appreciate it very much, because I hate to
calculate. Similarly, logical advances shorten some
paths and create new ones. In the end, I think we can
agree that we're all on our way to the same destination.

Tom


"Wrong" is not a particularly good choice of
words, I think. One can be wrong in the proof
strategy
one chooses, to the extent that the strategy is
shown
ineffective. Hilbert is influential, however,
because
he was right, in principle, about what mathematics
is
and what it can do. He saw far past the obvious --
I
find the Hilbert space much more of a "heaven" than
Hilbert claimed that Cantor created.


My own path leads to the independence of
language
and meaning. As Popper said, "I do not believe
in belief." I think, as did he, that what will
be
left when belief has been rendered entirely
irrelevant to a mathematical result, is
something
like Tarski's correspondence theory of truth.
Something
very naturally graspable ...

Can you give any references to this idea?<<


The Stanford Encyclopedia does a good job of
putting it
in negative terms:


http://www.seop.leeds.ac.uk/archives/spr2004/entries/t
ruth-correspondence/#1

When I speak of the correspondence theory of truth,
however, I mean as it was rehabilitated by Karl
Popper
(in Objective Knowledge: an evolutionary approach)
for
scientific application, sans philosophical baggage.
This
view goes by the broad name of metaphysical
realism, in
which verisimilitude ("truth-likeness") takes the
place
of the Platonic idea of truth.

Tom

Thank you Tom, I'll check out that link.

Craig


.

Relevant Pages

  • Re: Another Reason Why Collatz is Unprovable
    ... going to produce anything surprising, ... When one takes mathematics to a higher ... predict counterexamples without running the algorithm itself. ... Proving that there can't exist ...
    (sci.math)
  • Re: Another Reason Why Collatz is Unprovable
    ... going to produce anything surprising, ... When one takes mathematics to a higher ... Hilbert claimed that Cantor created. ... My own path leads to the independence of ...
    (sci.math)
  • Re: Another Reason Why Collatz is Unprovable
    ... going to produce anything surprising, ... When one takes mathematics to a higher ... Do you think logicians find their research uninteresting? ... Hilbert claimed that Cantor created. ...
    (sci.math)
  • Re: meta-proof that p is not np
    ... > given statement cannot be very surprising unless the statement itself ... > Theorem 1, mathematics is not surprising. ... is suprising, because after all the universe has causation so life, ... Incidentally, I thought it was a joke too, because it's pretty funny. ...
    (sci.math)
  • meta-proof that p is not np
    ... given statement cannot be very surprising unless the statement itself ... Theorem 1 leads to many other surprising corollaries. ... Anything that is interesting (not boring) has surprises. ... Theorem 1, mathematics is not surprising. ...
    (sci.math)