Gdels Proof

Henrik Jeldtoft Jensen Dept. of Mathematics Imperial College

Kurt Gdel 24.4.06-14.1.78

1

H.J. Jensen, Dept. of Math., IC

ON FORMALLY UNDECIDABLE PROPOSITIONS OF PRINCIPIA MATHEMATICA AND RELATED SYSTEMS 11 by Kurt Gdel, Vienna 1 The development of mathematics in the direction of greater exactness hasas is well knownled to large tracts of it becoming formalized, so that proofs can be carried out according to a few mechanical rules. The most comprehensive formal systems yet set up are, on the one hand, the system of Principia Mathematica (PM)2 and, on the other, the axiom system for set theory of Zermelo-Fraenkel (later extended by J. v. Neumann).3 These two systems are so extensive that all methods of proof used in mathematics today have been formalized in them, i.e. reduced to a few axioms and 27. rules of inference. It may therefore be [($u,v)u,v & x = u * R(b Gl rules & Su x(n|y) ez {z [Pr(l(x)+l(y))]x+y & surmised thatxthese axioms and x) * v of inference are & n = l(u)+1]} to decide all mathematical questions which can in any z = u * y * v also sufficient way at all be expressed formally in the systems concerned. It is shown below that this is x(n|y) derivesand thaton both the systems mentioned n-th term of fact relatively Su not the case, from x in substituting y in place of the there are in x (it simple problems in l(x)). being assumed that 0 < n the theory of ordinary whole numbers4 which [174] 28. cannot be decided from v Fr n,x & not ($p)[n < p is not due in p,x]} 0 St v,x en {n l(x) & the axioms. This situation l(x) & v Fr some way to the specialSt v,x of the systems set up, but holds for a < p <extensive& v Fr of formal (k+1) nature en {n < k St v,x & v Fr n,x & ($p)[n very k St v,x class p,x]}2 systems, including, in particular, all those arising from

H.J. Jensen, Dept. of Math., IC

The essence
First theorem of undecidability: If axiomatic theory is consistent, there exist theorems which can neither be proved or disproved

H.J. Jensen, Dept. of Math., IC

The essence
Second theorem of undecidability: There is no constructive procedure which will prove axiomatic theory to be consistent.
4

Euclids Elements
23 definitions 5 postulates

lived circa 300 BC

465 propositions

5
H.J. Jensen, Dept. of Math., IC

Euclids Elements
The axioms

It is possible to draw a straight line from any point to any point. to produce a finite straight line continuously in a straight line. to describe a circle with any centre and radius. That all right angles equal one another. Parallel lines dont cross
6
H.J. Jensen, Dept. of Math., IC

Euclids Elements

Consistency
Can mutually inconsistent statements be derive from a set of axioms. Say in Euclids geometry

7
H.J. Jensen, Dept. of Math., IC

Euclids Elements

In other words
Can we be sure no one some day derives a proposition which contradicts another proposition.

8
H.J. Jensen, Dept. of Math., IC

PM

9
H.J. Jensen, Dept. of Math., IC

Undecidable
Russells paradox:

Two types of sets: Normal those who dont contain themselves:

&

Non-normal those who do contain themselves

B
10

H.J. Jensen, Dept. of Math., IC

Undecidable

Examples:
Normal set:

A = the set of MPhys 2 student

A A

Non-normal: B = the set of all thinkable things

B
H.J. Jensen, Dept. of Math., IC

B
11

Undecidable

Define:

N = Set of all Normal sets

Question: Is N normal?
Assume N is Normal then N is member of itself, since N contains all Normal Sets per its definition i.e., N N. But if N N then N is non-Normal

So N being Normal implies N being non-Normal !

H.J. Jensen, Dept. of Math., IC

12

Undecidable

Define:

N = Set of all Normal sets

Assume N is non-Normal then N is member of itself per definition of non-Normal. But if N is non-Normal it is a member of itself, and N contains per definition Normal sets, i.e., N is Normal .

So N being non-Normal implies N being Normal !

13
H.J. Jensen, Dept. of Math., IC

The problem is

Self-reference
14
H.J. Jensen, Dept. of Math., IC

How to determine the truth of:

I am a liar!
Epimenides paradox
15
H.J. Jensen, Dept. of Math., IC

The strategy of Gdels proof

Distinguish between: mathematics &

x = x, x2 = 9

meta-mathematics

x=4 is not a solution of x+2=3, PM is consistent

16

H.J. Jensen, Dept. of Math., IC

The strategy of Gdels proof

Enumeration of formalised system: Signs:

mb del nu G er

1 2 3 4

not or If then there is an equals zero immediate successor

17
H.J. Jensen, Dept. of Math., IC

= 0 s

5 6 7

The strategy of Gdels proof

Enumeration of formalised system: Math formulas:
um del n G ber

( x )( x = sy )

28 34 L

There is a number x following right after y

18
H.J. Jensen, Dept. of Math., IC

The strategy of Gdels proof

Enumeration of formalised system: Meta-maths:

The formula G is not demonstrable using the rules of PM

ber l num de G

19
H.J. Jensen, Dept. of Math., IC

The crunch
Gdel constructed a formula G for which he showed that:

G is demonstrable

non G is demonstrable

20
H.J. Jensen, Dept. of Math., IC

More crunch
The meta-mathematical content of G is:

G= The formula G is not

demonstrable Or formally within PM:

( x ) Dem( x, Sub(n,17, n))

21
H.J. Jensen, Dept. of Math., IC

And more crunch

G= The formula G is not
demonstrable So since G cannot be demonstrated, it is, per definition, TRUE, though its TRUTH cannot be proved within PM
22
H.J. Jensen, Dept. of Math., IC

The Conclusion
All axiomatic systems will contain true propositions which cannot be proved within the system And contain propositions which cannot be determined whether true of false
23
H.J. Jensen, Dept. of Math., IC

Some consequences:
The continuum hypothesis:
Cantor

No set can have a number of elements between the cardinality of the natural numbers and the cardinality of the real numbers
24
H.J. Jensen, Dept. of Math., IC

Cardinality:
The real numbers cannot be countered
Proof: assume the opposite
Cantor

r1 = 0. x11 x12 x13 L r2 = 0. x21 x22 x23 L r3 = 0. x31 x32 x33 L M

H.J. Jensen, Dept. of Math., IC

r = 0. x11 x22 x33

r = 0. ~11 ~22 ~33 x x x

25

Cardinality:
So clearly: # reals > # integers
But: is there a set with a number of elements in between? Cantor said: No - but could not prove it.
26
H.J. Jensen, Dept. of Math., IC

Cantor

Consequences continued:
The continuum hypothesis:
Was the first problem in Hilberts list of 23 unsolved important problems.

Paul Cohen showed in 1963 that the continuum hypothesis is undecidable

27
H.J. Jensen, Dept. of Math., IC

More consequences:
Truth cannot be identified with provability

Roger Penrose: Creative mathematicians do not think in a mechanistic way. They often have a kind of insight into the Platonic realm which exists independently of us.
28
H.J. Jensen, Dept. of Math., IC

More consequences:
We cannot build one all embracing explanation of everything based on one finite set of axioms. There exist more true statements than the countable number of truths that can be recursively deduced from a finite set of axioms.
29
H.J. Jensen, Dept. of Math., IC

The world is too complex for a `finitistic axiomatic approach to suffice.

Creativity is needed at all levels of description.

30
H.J. Jensen, Dept. of Math., IC

So where does this leave

The Theory of Everything

?
31
H.J. Jensen, Dept. of Math., IC