Goedel's Theorem /godel's Theorem

A Simple Proof of
G
odels Incompleteness Theorems
Arindama Singh, Department of Mathematics, IIT Madras, Chennai-600036
Email: [email protected]
Introduction
G
odels incompleteness theorems are considered as achievements of twentieth century mathematics. The theorems say that the natural number
system, or arithmetic, has a true sentence which cannot be proved and the
consistency of arithmetic cannot be proved by using its own proof system;
see [1]. Though the ideas involved in their proofs are very complex, they
can be presented in a simple and comprehensible way.
Background
We assume a theory of arithmetic, say N = (N, +, ) to be consistent. Write

` X for X is a theorem in N . The usual theorems or laws of logic hold
true in this theory. We will be using explicitly the laws of Double Negation,
Contradiction, Distribution of implication, Contraposition, Modus Ponens
and Hypothetical Syllogism, as spelled out below.
` X X.
(1)
` X (X Y ).
(2)
` (X (Y Z)) ((X Y ) (X Z)).
(3)
If ` X Y, then ` Y X.
(4)
If ` X and ` X Y, then ` Y.
(5)
If ` X Y and ` Y Z, then ` X Z.
(6)
Besides the logical laws, there are some more theorems specific to arithmetic, which are obtained by encoding formulas as natural numbers. The
1
encoding is the so called Godel numbers. We define Godel number g() of

symbols, formulas (in general, strings), and proofs as follows.
Enumerate the symbols such as connectives, quantifiers, punctuation
marks, predicates, function symbols, and variables as:
>, , , , , , , (, ), P1 , f1 , x1 , P2 , f2 , x2 , . . .
Define g() = n, where the symbol comes as the n-th in the above list.
Extend g to strings of these symbols by
g(m )
g(1 2 m ) = 2g(1 ) 3g(2 ) pm
where pi is the i-th prime number. This defines g of terms and formulas.
Next, extend g to proofs of formulas by
g(Xm )
g(X1 X2 . . . Xm ) = 2g(X1 ) 3g(X2 ) pm
.
where again pi is the i-th prime number.

Due to prime factorization theorem in N , the function g has the following
properties:
(a) g is a computable function.
(b) g(uv) can be computed from those of g(u) and g(v).
(c) Given n N, if n = g(X) and X is known to be a symbol, or a formula,
or a proof, then X can be computed from n.
Let P roof (x, y) be a binary predicate that translates x is the Godel
number of a proof of a formula whose Godel number is y. Write
P r(y) = xP roof (x, y).
That is, we interpret P r as the subset (unary relation) of N which is the
set of G
odel numbers of all provable (in N ) formulas. P r(g(X)) means
that there is a natural number which is the Godel number of a proof of a
formula whose G
odel number is g(X). Which, in turn, means that there is
a natural number which is the Godel number of a proof of X. We thus use a
2
further abbreviation such as P (X) = P r(y) = P r(g(X)). The predicate P r

is the provability predicate. We may also say that P is a predicate whose
arguments are formulas, and that P (X) means that X is provable in N .
P (X) is a formula in the theory N . We thus loosly call P as the provability
predicate.
Provability Predicate
The provability predicate P has the following properties:

If ` X, then ` P (X).
(7)
` P (X Y ) (P (X) P (Y )).
(8)
` P (X) P (P (X)).
(9)
Since 0 6= 1 in N , P (0 = 1) expresses inconsistency of N . Therefore,

consistency of N may be formulated by asserting that the sentence P (0 = 1)
is not a theorem of N . Our assumption of consistency of N thus gives
6 ` P (0 = 1).
(10)
Let B1 (n), B2 (n), . . . be an enumeration of all formulas in N having

exactly one free variable. Consider the formula P (Bn (n)). This is one in
the above list, say Bk (n). Since ` p p, we have ` Bk (n) P (Bn (n)).
Then. ` n(Bk (n) P (Bn (n))), by universal generalization. And
` Bk (k) P (Bk (k)),
by universal specification with [n/k]. Abbreviating the sentence Bk (k) to A,
we obtain:
` A P (A).
(11)
The statement (11) says that the sentence This sentence is not provable
is expressible and is a theorem in N ; ingenuity of Godel.
For a formal proof of (11), start with the formula B(x), having exactly
one free variable. Let the diagonalization of B(x) be the expression
x(B(x) (x = g(B(x)))).
Since g a computable function, the relation
diag(m, n) : n is the Godel number of the diagonalization of the
formula having exactly one free variable with Godel number m
is recursive and hence representable in N by some binary predicate, say,
C(x, y). Next, define
the formula F (x) as y(C(x, y) B(y))
the sentence G as x(y(C(x, y) B(y)) (x = g(F (x)))).
Finally, show that ` G B(g(G)).
You have thus proved the Diagonalization Lemma:
for each formula B(y) with exactly one free variable there exists
a sentence G such that ` G B(g(G)).
Next, take B(y) as P r(y) to obtain (11); see for example, [2].
Here are some more properties of this special sentence A.
` P (A) P (A)
Proof: ` A P (A)
` P (A) A
` P (P (A) A)
` P (P (A) A) (P (P (A)) P (A))
` P (P (A)) P (A)
` P (A) P (P (A))
` P (A) P (A)
(12)
(11)
(4)
(7)
(8)
(5)
(9)
(6)
The sentence A is connected to inconsistency as the following property

shows:
` P (A) P (0 = 1).
4
(13)
Proof: ` A (A (0 = 1))
` P (A (A (0 = 1)))
` P (A (A (0 = 1)))
(P (A) P (A (0 = 1)))
` P (A) P (A (0 = 1))
` P (A (0 = 1)) (P (A) P (0 = 1))
` P (A) (P (A) P (0 = 1))
` (P (A) (P (A) P (0 = 1)))
((P (A) P (A)) (P (A) P (0 = 1)))
` (P (A) P (A)) (P (A) P (0 = 1))
` P (A) P (A)
` P (A) P (0 = 1)
(2)
(7)
(8)
(5)
(8)
(6)
(3)
(5)
(12)
(5)
Incompleteness Theorems
Theorem 1 There exists a sentence C in N such that 6 ` C and 6 ` C.

Proof: Take C = A, the sentence used in (11). If ` A, then by (7), ` P (A).
But by (11), ` A implies that ` P (A). This is a contradiction. On the
other hand, if ` A, then by (11), ` P (A). By (13) and (5), ` P (0 = 1).
This contradicts (10). Therefore, neither ` A nor ` A.

Theorem 2 There exists a true sentence in N which is not provable in N .
Proof: Consider the sentence A used in (11). Either A is true or A is true.
But neither is provable by Theorem 1. Whichever of A or A is true serves
as the sentence asked in the theorem.

Theorem 3 6 ` P (0 = 1).
Proof: Suppose ` P (0 = 1). Due to (1), (4) and (13), ` P (A). By (11),
` A. By (7), ` P (A). By (13), ` P (0 = 1). This contradicts (10).

Notice that the sentence P (0 = 1) can also serve the purpose of the
sentence asked in Theorems 1-2.
5
Theorems 1-2 are called as Godels First Incompleteness theorem; they

are, in fact one theorem. Theorem 1 shows that Arithmetic is negation
incomplete. Its other form, Theorem 2 shows that no axiomatic system for
Arithmetic can be complete. Since axiomatization of Arithmetic is truly
done in second order logic, it shows also that any axiomatic system such as
Hilberts calculus for second order logic will remain incomplete.
In (9-12), we could have used any other inconsistent sentence instead of
(0 = 1). Since P (0 = 1) expresses consistency of Arithmetic, its unprovability in Theorem 3 proves that consistency of Arithmetic cannot be proved
using the proof mechanism of Arithmetic. It shattered Hilberts program
for proving the consistency of Arithmetic. Herman Wyel thus said:
God exists since mathematics is consistent,
and the Devil exists since its consistency cannot be proved.
References
[1 ] K. G
odel, Uber
formal unentscheidbare Satze der Principia Mathematica und verwandter Systeme, I, Monatshefte f
ur Mathematik und
Physik 38: 173-98, 1931.Trans: On formally undecidable propositions
of Principia Mathematica and related systems I, In: Solomon Feferman, ed., 1986. Kurt Gdel Collected works, Vol. I. Oxford University
Press: 144-195.
[2 ] A. Singh and C. Goswami, Fundamentals of Logic, Indian Council of
Philosophical Research, New Delhi, 1998.

Goedel's Theorem /godel's Theorem

Uploaded by

Copyright:

Available Formats

Goedel's Theorem /godel's Theorem

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Goedel's Theorem /godel's Theorem

Uploaded by

Copyright:

Available Formats

How does the document define Gödel numbers?

What are some properties of the provability predicate P?

A Simple Proof of

We assume a theory of arithmetic, say N = (N, +, ) to be consistent. Write

` (X (Y Z)) ((X Y ) (X Z)).

encoding is the so called Godel numbers. We define Godel number g() of

where again pi is the i-th prime number.

further abbreviation such as P (X) = P r(y) = P r(g(X)). The predicate P r

The provability predicate P has the following properties:

Since 0 6= 1 in N , P (0 = 1) expresses inconsistency of N . Therefore,

Let B1 (n), B2 (n), . . . be an enumeration of all formulas in N having

The sentence A is connected to inconsistency as the following property

Theorem 1 There exists a sentence C in N such that 6 ` C and 6 ` C.

Theorems 1-2 are called as Godels First Incompleteness theorem; they

You might also like