I've copied and pasted it here for your convenience, with some
reformatting it. Hope you don't mind.
-------------------------------------------------------------------------
Subject: Mathematical Relativity (MR) - The End of Absolute Truths in
The Concept of The "Natural Numbers"
Current Revision: 01
Author: Nam Nguyen
Contact:
namduc...@shaw.ca
Date: Jan. 29, 2014
*************************************************************
* *
* Introduction *
* *
*************************************************************
In this presentation, we'll put forward the notion that certain
arithmetic truth values in the natural numbers are relativistic,
in the sense that there would exist statement-formulas written
in the language of arithmetic whose truth values in the natural
numbers can be assumed either way at will, without contradicting
any other properties of the natural numbers.
The mechanism we'll use to prove the existence of such truth
relativity would be twofold:
- First, we'll use a certain language-structure theoretical
meta language to formalize the intuitive, informal concept
of the natural numbers. In this way our inference in meta
level about the natural numbers would be with clarity and
precision that intuition couldn't adequately offer.
- Secondly, we'll employ a logic principle which would basically state
that if the required information to assert a meta statement S as true,
or false, is more than what can be available in the body of
permissible knowledge of the underlying reasoning framework, then
it's impossible to assert S as true or false within the framework.
As such, the truth value of S can be chosen at will and can be said
to be relative to the choice of true or false one can choose. Thus a
relativity would exist.
The presentation would be organized into sections as per the following:
- Introduction (this section).
- Section I - Assumptions.
- Section II - Definitions.
- Section III - Meta Thesis.
- Section IV - Knowledge Operators K(), nK().
- Section V - The Natural Numbers as a Language Structure.
- Section VI - Meta Theorems.
*************************************************************
* *
* Section I - Assumptions *
* *
*************************************************************
This section contains some general assumptions. There are some
other assumptions specific to certain subsequent sections which
will be listed there.
Assum-01: The underlying logic framework is FOL(=) which is
First Order Logic with equality. Fundamental understanding
of FOL(=) and its definitions, including language structure
theoretical definitions, is assumed in the presentation.
In addition, the FOL concept of finiteness, or being finite,
is taken for granted, with no further dependency or explanation
is required.
Assum-02: The language of arithmetic is the familiar L(0,<,S,+,*).
Assum-03: The presentation is language structure theoretically centric,
no knowledge of any specific formal system (axiomatization)
is required.
*************************************************************
* *
* Section II - Definitions *
* *
*************************************************************
This section contains some common or (here) frequently used definitions.
There might be some other definitions specific to certain subsequent
sections in which case those definitions would be listed there.
=================> First Order Definitions
Def-01: prime(x) <-> (S0 < x) /\ ~Ez[ (z < x) /\ ~(z = S0) /\
(Et[((t=x) \/ (t<x)) /\ (x = z*t)]) ]
Def-02a: even1(x) <-> Ey[x=y+y]
Def-02b: even2(x) <-> Ey[x=2*y]
Def-02c: even(x) <-> (even1(x) \/ even2(x))
Def-03a: odd1(x) <-> Ey[x=(y+y+S0)]
Def-03b: odd2(x) <-> Ey[x=((SS0*y)+S0)]
Def-03c: odd(x) <-> (odd1(x) \/ odd2(x))
Def-04a: GC(x) <-> (even(x) /\ (SS0<x)) -> Ep1p2[prime(p1) /\
prime(p2) /\ (x=p1+p2)]
Def-04b: nGC(x) <-> (even(x) /\ (n<x)) -> Ep1p2[prime(p1) /\ prime(p2)
/\ (x=p1+p2)]
where n is an even numeral greater than SS0.
Def-04c: aGC(x) <-> (even(x) /\ (SS0<x)) -> Ap1p2[prime(p1) /\
prime(p2) /\ (p1+p2<x \/ x<p1+p2)]
Def-04d: n-aGC(x) <-> (even(x) /\ (n<x)) -> Ap1p2[prime(p1) /\
prime(p2) /\ (p1+p2<x \/ x<p1+p2)]
Where n is an even numeral greater than SS0.
Def-05a: GC <-> Ax[GC(x)]
Def-05b: aGC <-> Ax[aGC(x)]
Def-05c: nGC <-> Ax[nGC(x)]
Def-05d: n-aGC <-> Ax[n-aGC(x)]
Def-07a: Assuming a P(x), the statement "There are infinitely many
examples of P" would be symbolized as '(I)P(*)' and defined
as:
(I)P(*) <-> Ex[P(x)] /\ AxEy[P(x) -> (P(y) /\ Ez(y = x + Sz))]
This is called I-form (Inductive) of infinity expression.
Def-07b: Assuming a P(x), the statement "There are infinitely many
examples of P" could also be symbolized as '(aI)P(*)' and
defined as:
(aI)P(*) <-> Ex[P(x)] /\ AxEy[P(x) -> (P(y) /\ (x < y))]
This is called aI-form (anti-Inductive) of infinity expression.
Def-07c: P(*) <-> ((I)P(*) \/ (aI)P(*))
This is the general form of infinity expression about the
naturals.
Def-08: cGC <-> aGC(*)
=================> Meta Level Definitions
Def-M00: A meta statement S is a statement written in natural language
about FOL(=) framework; logically, S must necessarily be
(exclusively) true or false. In addition, given a meta
about S is also a meta statement.
Def-M01: A language structure theoretical set M, for a particular
language L, is a set of 2-tuples defined as:
M = {
<'U', U>,
<Cin, Uin>, ...
<Sn, Rn>, ...,
}
where:
- U is a non-empty set of individuals. U is called the universe of M,
and is denoted by U(M), or just U when the context is understood.
- Each Cin is an individual constant symbol of the language,
corresponding to the particular individual Uin in U.
- Each Sn is a non-individual constant symbol of the language,
symbolizing the set Rn of m-tuples of individuals in U. Each Rn
in turn might or might not be a (language structure theoretical)
function set.
A corollary of the definition is that each language structure M is
a language structure theoretical set, but not necessarily the other
way around.
*************************************************************
* *
* Section III - Meta Thesis *
* *
*************************************************************
These are meta theses or principles we'd accept as being true,
on the basis of their being so self-evidently logical. The meta
theses are abbreviated in the form MTi
MT1 - HP (Humility Principle)
Given a meta statement S, if the information or knowledge required to
assert both S and its negation as true or false is outside the collection
of available knowledge about the underlying logic framework [FOL(=)],
then it is impossible to assert S and its negation as true or false
within the context of the framework.
In this case, the impossibility of knowing the truth value would
refer to both S and its negation. Also, in this case both S and
its negation are called relativistic meta statements.
*************************************************************
* *
* Section IV - Knowledge Operators K(), nK() *
* *
*************************************************************
However philosophical it may sound, in term of its meaning, "knowledge"
is a required and taken for granted terminology in mathematical reasoning.
For instance, to know of a FOL theorem, which people would know so
from time to time, one must necessarily know of a finite proof-sequence
of the theorem, which in turn would require one to know of certain
strings as valid language expressions, etc...
It's just that while the meaning of "knowing" could be taken for granted
in FOL(=), the meaning of its negation - "NOT knowing" - is of the taboo
kind of semantics: one seems to be fearful that one would bring
philosophical arguments or even imprecision if one tries to apply "NOT
knowing" in arguments or reasoning. But given HP, we can no longer
dismiss "NOT knowing" as just being philosophical: it is as meaningful
as its negation counterpart, i.e. "knowing". And this unavoidability of
recognizing the mathematical foundation meaning of both mutually
negation words "knowing" and "not knowing" is the basis for our
defining knowledge operators K(), nK() below. Please note, however,
we're NOT referring to the transient kind of "not knowing".
Given a meta statement S, with its corresponding negation neg(S), we'd
have the following definitions.
Def-M.IV.01: K(S) is defined to be a meta statement which is
true iff S being true, within the FOL(=), is in
the collection of valid knowledge about FOL(=)
reasoning.
Def-M.IV.02: nK(S) is defined to be a meta statement which is
true iff S being true, within the FOL(=), is NOT
in the collection of valid knowledge about FOL(=)
reasoning.
Def-M.IV.03: undecide(S) is defined to be a meta statement as:
undecide(S) <=> nK(S) and nK(neg(S)).
If so desired, we could refer to undecide(S) as
relativity(S), or rel(S).
Note 1: In general K(S) if S being true is due to some valid
meta theorems, based on valid, permissible definitions
or methods of reasoning within FOL(=), while nK(S) if
S being true would lead to some contradiction to the known
knowledge or logical principles, including the HP stated
above.
Note 2: For a given first order formula F which would be true or
false in M (a language structure theoretical set), then
K(S) would refer to it being possible to know F is true
in M, and similarly nK(S) would refer to it NOT being
possible to know F is true in M.
---> Examples
For instance, suppose the formal system T is defined as T = {~0=0}
then K(T |- Ex[~x=x]).
Or, to get ourselves more familiar how the K() and nK() operators
are used, we could express it as the following true meta statement:
K(K(T = {~0=0}) => neg(neg(K(T |- Ex[~x=x])))).
*************************************************************
* *
* Section V - The Natural Numbers as a Language Structure *
* *
*************************************************************
---> Def-M.V.01:
By a meta language for describing the construction of a language
structure theoretical set, which is an unformalized set, we'd mean
the following conventions in meta level:
- 'this.S': A meta symbol denoting the underlying set S in
consideration.
- 'e': A meta symbol denoting the usual set membership.
- '=': A meta symbol denoting the usual set or element equality.
- '{', '}: Meta symbols denoting the usual set demarcation.
- 'U': A meta symbol denoting the usual set union operation.
- 'A': A meta symbol denoting the usual set intersection operation.
- '{ x | expression (x)}':
Would denote an expression defining element x's in the
underlying set, symbolically denoted by '{' and '}'.
Corollary: If in defining a set this.S, the "this.S" operator must
necessarily be present on the right side of an occurrence of
'=', as a definiens, then the defined set this.S is infinite.
Otherwise, this.S is finite.
For example, the set this.S defined below is an infinite set.
this.S = { {} } U { x | (x e this.S) => ({x} e this.S) }
---> Def-M.V.02:
The concept of the natural numbers is that of a language structure
theoretical set, denoted by N, where the universe this.U is defined
inductively as:
this.U = {zero} U { x | (x e this.U) => ( Succ(x) e this.U) }
Where:
- zero is a fixed element of an infinite collection K1 of elements.
- Succ is a meta symbol representing any infinite 1-1 unary-mapping
(successor) operation from elements of K1 to elements of K1.
For instance, if zero symbolizes an infinite set in K1, and Succ(x)
symbolizes the power set of the set denoted by x. Hence in this example
all natural numbers are infinite elements of K1.
The language symbols for N are:
- 0: The language individual constant symbol standing for the element
zero.
- <: The (intended) language binary relation symbol standing for being
strictly less than.
- S: The (intended) language 1-ary successor function symbol.
- +: The (intended) language binary addition function symbol.
- *: The (intended) language binary multiplication function symbol.
As defined, N is strictly a language structure theoretical set.
N is intended to be a language structure but at this stage its being
so has not been verified. However, certain knowledge about N would
be assumed here:
- The set of 2-tuples to be symbolized by '<' be the familiar
binary relation for the total order "strictly less than".
- The set of 2-tuples to be symbolized by 'S' be a the familiar
successor function set.
- etc ...
*************************************************************
* *
* Section VI - Meta Theorems *
* *
*************************************************************
These meta theorems are abbreviated as TMi.
---> TM1
A language structure M is a structure theoretical set, but
not necessarily the other way around.
Proof:
Per definition Def-M01, if Rn of n-ary tuples is indeed a function,
then the symbol Sn is indeed a function symbol. On the other hand,
if Sn is supposed to be a function symbol but Rn is not a corresponding
function (set), then M isn't a structure.
In other words, a language structure M is an already verified
structure theoretical set.
Q.E.D.
---> TM2
Per definition Def-M.V.02, there are uncountably many instances for
the concept of the natural numbers.
Proof:
Given the universe this.U is infinite, and the intended assumptions
for the set of 2-tuples to be symbolized by '<', and for the set of
2-tuples to be symbolized by 'S', it's a trivial knowledge that there
exist - by Choice Principle - an uncountable collection K2 of infinite
sets of 2-tuples, any element of which could be considered for being
symbolized by '<', and an uncountable collection K3 of infinite sets
of 2-tuples, any element of which could be considered for being
symbolized by the intended successor function symbol 'S'.
Consequently, we'd have similarly uncountably many choices of function
sets for addition symbolized by '+', and multiplication symbolized by
'*'.
Q.E.D.
Then by convention, if the element m is 0, S0, SS0, or a non-trivial
multiple number then m is still the very same natural number, invariant
to any choice [we've chosen] of the sets n-tuples for the symbols, '<',
'S', '+', '*'.
On the other hand, there are 2 elements, symbolized by 'P1', 'P2' which
are neither 0, nor S0, nor SS0, and are primes in which:
nK(P1 < P2) and nK(P2 < P1).
---> TM3
(Goldbach Conjecture is true) => nK(Goldbach Conjecture is true).
Proof:
First, without loss of generality, we're assuming the Goldbach
Conjecture would be stated for an even number greater than or
equal to 10, and similarly for any mentioned even number which
is or is not a counter example of the conjecture. [In this way,
all the primes stipulated in the conjecture would not be even,
among other things].
Secondly, by an open even interval [E0, ...] we mean the set of
all even numbers greater than or equal to a fixed even number E0:
for instance, in the comment above, we'd be considering the
open even interval [E0=10, ...].
Without referring to the formal system PA, we'd denote P1
and P2 as 2 constant symbols such that prime(P1) /\ prime(P2)
is true in any structure M of the collection K previously
mentioned (where 0, S0, SS0 and all even [or multiple] numbers
would be invariant in their "successor" function positions,
but not the primes).
Then for any given general even number e in the open even interval
[E0, ...] where E0 = 10, it's impossible to choose two primes p, p'
so to conclude e = p + p', since if we suppose e = p + p', we also
can suppose p = P1 and p' = P2 but since nK(P1 < P2), it is impossible
to assert that the even number P1+P2 is invariant, per TM2, which is
a contradiction.
Hence if 'GC' stands for the conjecture, we'd have nK(GC).
Q.E.D.
---> TM4
((Goldbach Conjecture is false) and (cGC is true)) => nK(cGC).
Proof:
Again, without loss of generality, we're assuming all the relevant
even numbers are in [E0, ...] where E0 = 10.
Now let e1 be a general even number, then there exist 2 primes p1, p2
such that (p1+p2 < e1). Then, there exists another e2 such that (e1 < e2)
and there exist 2 primes p3, p4 at least one of which is greater than
both p1, p2 and (p3+p4 < e2).
What the above means is that - in general - for each increasing greater
counter example e of the conjecture, more increasing prime numbers
(but less than e) have to be considered for validating the underlying
even number e is indeed a counter example.
But given in this case there are infinitely many counter examples of
the conjecture, it is impossible to validate that a general counter
example e would indeed be so, since there exist 2 unknown constant
primes P1, P2 (previously stipulated) that we can't know if both P1, P2
are less than e, since nK(P1 < P2) and nK(P2 < P1).
Therefore, the entire infinite list of counter examples of Goldbach
Conjecture if exists can't be validated so, hence can't be known to exist.
Q.E.D.
---> TM5
undecide(cGC)
Proof:
The first order formula statement cGC is expressed in term
of the Goldbach Conjecture (GC), hence there are four cases when
considering the truth values of cGC and its negation ~cGC.
Case (1): GC is true but cGC is also true.
Case (2): GC is true hence cGC is false.
Case (3): GC is false and cGC is also false.
Case (4): GC is false but cGC is true.
Case (1) is a contradiction and can be disregarded.
Cases (2) and (3) have been proved by way of TM3.
Case (4) has been proved by way of TM4.
Hence by definition Def-M.IV.03, we've proved:
- undecide(cGC)
- relativity(cGC).
Q.E.D.
--
-----------------------------------------------------
There is no remainder in the mathematics of infinity.
NYOGEN SENZAKI