Koffi1
Leslie Koffi
Philosophy 399
Gödel’s Incompleteness Theorem
Kurt Gödel proved, in a famous paper published in 1931, that in any true (consistent) axiomatic theory
sufficiently rich to enable the expression and proof of basic arithmetic propositions, it will be possible to
construct an arithmetical proposition G such that neither G, nor its negation, is provable from the given
axioms. Hence the system must be incomplete. Gödel developed a technique now known as Gödel
numbering, in an effort to prove this which codifies formal expressions as natural numbers. The
incompleteness theorems of 1930 are contrary results in Hilbert’s meta-mathematical program as the first
incompleteness theorem shows the impossibility of obtaining a complete axiomatization of arithmetic.
Since the extent of Gödel’s mathematical work is firmly rooted in philosophical inquiry, his use of metamathematics was his tool for finite reasoning. He viewed mathematical logic as a continual reflection on
the nature of mathematical knowledge. This provided a powerful measure of addressing core
epistemological concerns. Because the work of meta-mathematics encompasses the very nature of
mathematics itself, its roots within the principles of set-theory has shown to be subjected within its own
limitation. The first incompleteness theorem showed there are true propositions about whole numbers that
cannot be proved from the axioms. Hence, formal systems which, such as the following, are consistent are
not complete:
“Any self-consistent recursive axiomatic system powerful enough to describe the arithmetic of the natural
numbers (for example Peano arithmetic)” (Ferreirós, 2007).
Historical Context: The Foundation’s Crisis
In the early twentieth century, set theory, the concept of the continuum, the role of logic and the
axiomatic method versus the role of intuition were heavily debated among mathematicians and
philosophers. As these debated concepts contributed to the foundations crisis in the 1920’s, it questioned
Koffi2
the philosophical basis of mathematics where Hilbert devised a plan to investigation philosophical
theories by exploring the concerning nature of mathematics.
The following viewpoints including logicism, formalism and intuitionism questioned the basis of
foundational math given that the status of mathematical knowledge in this period led to the historical
debate known as the foundations crisis. The foundations crisis was rather a localized event in the 1920’s
where the debates between the partisans of “classical” (late-nineteenth century) mathematics, led by
Hilbert, and their critics, led by Brouwer, an intuitionist, advocated for stronger revisions of foundational
math (Ferreirós, 2007). Logicism, which Hilbert endorsed in 1899, was a unique reaction to the rise of
modern mathematics as the set-theoretic approach was an intellectual reaction to the concept of the
continuum. “Logicism was the thesis that the basic concepts of mathematics are definable by means of
logical notions, and that the key principles of mathematics are deducible from logical principles alone”
(Ferreirós, 2007). However, over time this thesis has become unclear because it seemed to be based on an
immature conception on logical theory, which became to be known as naïve set theory, confirmed later by
the paradoxes. Hilbert’s endorsement of logicism was rather a self-conscious endorsement of certain
modern methods in which he followed Richard Dedekind’s understanding of mathematics. This rise of
modern methods emerged in the nineteenth century and was associated with Göttingen mathematics.
Some of the modern approaches included the following:
(i)
(ii)
(iii)
a wholehearted acceptance of infinite sets and the higher infinite
a preference “to put thoughts in the place of calculations” (Dirichlet), and to concentrate on
“structures” characterized axiomatically
a reliance on “purely existential” methods of proof (Ferreirós, 2007)
Dedekind’s approach to the algebraic number enabled proven results such as the fundamental theorem of
unique decomposition as the proof was conceptually abstract. Dedekind insisted in the importance of a
general, conceptual theory. Hence, “he proved in full generality within any ring of algebraic integers,
ideals possess a unique decomposition into prime ideals” (Ferreirós, 2007). However, Leopold Kronecker,
a German mathematician who worked on number theory and algebra complained that Dedekind’s proofs
did not enable calculations and thus the proof was purely existential. Kronecker’s view was that these
Koffi3
modes of abstract methodology on algebraic properties were too remote for algorithmic treatment.
However, with Dedekind’s modern approach, “thoughts were in place of calculations”, as this principle
was also emphasized by Riemann’s theory. Hilbert was rather critical of Kronecker’s methodological
condemnation for mathematics. Certainly, Hilbert’s program can be seen, in many ways, as an attempt to
dismiss Kronecker’s ideas.
Kronecker also complained about the Balzano-Weistrass theorem, where every bounded sequence has
a convergent subsequence, because the theorem rests entirely on the completeness axiom for ℝ. Since real
numbers cannot be constructed in an elementary way from the rational numbers, one has to make use of
infinite sets, such as the set of all possible Dedekind cuts. Hence, Kronecker noted that the accumulation
point in the Bolzano–Weierstrass theorem cannot be constructed by elementary operations from the
rational numbers hence contributing to the concept of generality (Ferreirós, 2007).
Cantor and the Paradoxes
The idea of the set of ℝ or “the continuum” are non-constructive elements in modern mathematics.
Georg Cantor’s proofs in set theory also became essential examples of the modern methodology of
existential proof. An enumerable set ∑ can be listed off in some numerical order and the list may be
infinite, hence we now introduce the idea of Cantor’s Diagnolization argument. Cantor's theorem about
the indenumerability of the set of total functions from ℕ to ℕ uses the argument that if we assume that all
functions from ℕ to ℕ are denumerable—i.e. countable, then we should be able to form a new function
that is a bijection from the original one. Now Cantors theorem states that, “for any set A, the set of
all subsets of A (the power set of A) has a strictly greater cardinality than A itself” (Ferreirós, 2007). His
defense of the higher infinite and modern methods in a paper he wrote in 1883, prompted attacks from
Kronecker and in 1887, where he published an attempt to elaborate his foundational views. However,
Dedekind replied with a detailed set-theoretic and logistic theory of the natural numbers in 1888. Thus,
this early round criticism led to the victory for the modernist. Other powerful allies such as Felix Klein,
Koffi4
Giuseppe Peano, Hilbert and more join this new movement of modern thinking. Hence, during the 1890’s,
the modernist viewpoint for logic in particular expanded.
However, the dramatic change came with the logical paradoxes which were discovered by
Bertrand Russell, Ernst Friedrich Zermelo, Abraham Fraenkel and more. Their arguments showed the
assumption that certain existing sets do in fact lead to contradictions. These were later known as the settheoretic paradoxes which, in effect, destroyed attractive developments proposed by logicism. Russell’s
paradox states, “ the set of all sets are not elements of themselves”. Hence, the set is both a member of
itself and not a member of itself. Russell’s paradox relies on the naïve comprehension principle (∃A)
(∀x)(x ∈ A ≡ φ). Let φ be that x is not a member of itself. If A is a member of A, the condition φ does not
apply and A is not a member of A. However, if A is not a member of A, then condition φ does apply and
A is a member of A. Therefore, he showed that formal systems of arithmetic are not complete. Zermelo–
Russell tried to get out of Russel’s paradox by showing that the comprehension principle is contradictory,
by formulating a property that seems to be rooted in logical theory in stating that (∀A)( ∃B)(∀x)(x ∈ B ≡
((x ∈ A) & φ) which also led to a contradiction. Let φ be that x is not a member of itself and let A be a
universal set. If B is a member of B, then B is a member of the universal set and condition φ applies, so B
is not a member of itself. However, if B is not a member of B, then condition φ applies, thus B cannot be
a member of the universal set i.e. the universal set does not exist. This lead to another paradox.
Before the paradoxes, logicism was embraced by both mathematicians and philosophers, however, by
the 1920’s, logicism was of more interest to philosophers. This also led to more questions on the theory of
the rationale of foundational math itself as it became uncertain in that set theory was now shown to be
unstable.
Paris Problem
As Hilbert made the ideological transition from logicist to formalist, he founded the "formalist"
approach in Philosophy of Mathematics; Hilbert advocated in 1921 that the primary aim should be to
Koffi5
establish mathematics on a solid and provably consistent foundation of axioms. Therefore in principle, all
mathematical truths could be deduced by standard methods of first order or "predicate" logic. Hilbert’s
goal from this point forward would be to establish what mathematics should be in the twentieth century
by presenting his famous list of mathematical problems at the Paris International Congress of
Mathematics in 1900. He discussed Cantor’s continuum problem, and proposed the problem of whether
every set can be well-ordered. His second problem was to establish the consistency of the idea being the
set ℝ of real numbers. Hilbert then employed his fellow colleague Zermelo to show that ℝ (the
continuum) can be well-ordered and to establish the axiom of choice (Ferreirós, 2007).
Predicativity and the Axiom of Choice
Before the axiom of choice (AC) was formulated, Jules Henri Poincaré had critical viewpoint points
against logicism and formalism. His analysis of the paradoxes led him to formulate an important new
concept of predicativity, and maintain that impredicative definitions should be avoided in mathematics. In
an informal sense, a definition is impredicative when it introduces an element by reference to a totality
that already contains that element. A typical example is Dedekind’s definition of the set ℕ with a
sequence beginning with 1, with each number n associated with its successor (Ferreirós, 2007).
With the predicativity approach to foundational mathematics, all mathematical objects must be
introduced by explicit definitions. Hence, one begins with the idea of definable real numbers, for
example. Predicativity refers only to totalities that have already been established. Zermelo, however,
argued that this notion was limited in arguing that impredicative definitions were often used
unproblematically, not only in set theory but also in classical analysis (Ferreirós, 2007). The least upper
bound in real analysis is an example of the impredicative definition as the real numbers are not introduced
separately but rather as a completed whole.
Russell, who incorporated the theory of types, is a system of higher-order logic, with quantification
over properties or sets, over sets of sets and etc. It is based on the idea that the elements of any set should
Koffi6
always be objects of a certain similar or uniform type. For example, we can have individual sets such as
{a, b}, or sets of sets of individuals, such as {{a}, {a, b}}, but never a mixed set such as {a, {a, b}}
(Ferreirós, 2007). This system, together with axioms of infinity and choice, served as the development of
set theory and the number systems as it became the logical basis for the renowned Principia Mathematica
by Whitehead and Russell. They were considered in the development of the foundation for mathematics.
The main idea of Principia Mathematica was to develop an acceptance of the theory of the natural
numbers as they were considerably developed using classical logic.
A predicative foundation for mathematics lies within the foundation of modern methodology and it is
one of several approaches that do not fit into the conventional and not outdated forms of logicism,
formalism, and intuitionism.
The axiom of choice (AC), is the principle that, given any infinite family of disjoint nonempty sets,
there is a set, known as a choice set, that contains exactly one element from each set in the family. In
Hilbert’s proposition of well-ordered sets, the essential well-ordering of ℝ “exists” where ideally, the
completeness axiom assure that ℝ has no gaps and is non-empty. It seemed clear that this proposition
would be completely out of reach from any constructivist perspective. Thus, the axiom of choice
aggravated obscurities in previous conceptions of set theory, which forced mathematicians to introduce
clarifications. Moreover, AC was an explicit statement of prior interpretations about arbitrary subsets
which inherently set the stage was set for deep debate among critics. However ironically, AC in less
evident ways proved theorems in analysis (Ferreirós, 2007).
The ZFC system stands for Zermelo–Fraenkel with choice. It considerably categorized key traits of
modern mathematical methodology, offering a satisfactory framework for the development of
mathematical theories and the conduct of proofs. By allowing impredicative definitions and arbitrary
functions, this system of AC had strong existence principles. ZFC system asserted that axioms have
strong existential assumptions which sufficiently derived all of classical mathematics and Cantor’s theory
of the higher infinite.
Koffi7
Intuition
Brouwer’s philosophical version of constructivism presented his unique metaphysical views on the
foundation for mathematics in 1907. His philosophy of “intuitionism” derived from the view that
individual consciousness is the one and only source of knowledge. Hence, his proposition insisted on the
absolute independence of mathematics from language and logic.
Brouwer brought new ideas about questioning traditional ways of reasoning the natural numbers and
classical logic. For instance, the use of quantifiers and his famous rejection of the principle of the
excluded middle (PEM). Thus his critiques developed an alternative theory of analysis that was more
radical than his counterparts. Figures such as Klien and Poincare believed intuition has a necessary role to
play in mathematical knowledge. Hence the importance of logical properties in proofs and in the
development of mathematical theory infers that mathematics cannot be reduced to pure logic. Therefore,
proofs and theories are certainly organized in a logical sense, but their basic principles, axioms, are
grounded in intuition. Hence, there was a dramatic shift at this point of the foundations between Hilbert’s
attempts to justify “classical” mathematics vs. Brouwer’s reconstruction of intuitionistic mathematics.
There were initially high hopes that the development of intuitionism would end in an elegant and yet
simple demonstration of pure mathematics. However, Brouwer’s reconstruction developed in the 1920s
became clearer at some point that intuitionistic analysis was extremely complicated and extraneous
(Ferreirós, 2007).
Hilbert Program
The goal of Hilbert’s program was to eliminate all the skeptics by establishing the acceptability of the
classical theories of mathematics. This new perspective relied profoundly on formal logic that are
provable from given formulas (the axioms). It included the following:
Finding suitable axioms and primitive concepts for a mathematical theory T, such as that of the
real numbers.
Koffi8
Finding axioms and inference rules for classical logic, which makes the passage from given
propositions to new propositions a purely syntactic, formal procedure.
Formalizing T by means of the formal logical calculus, so that propositions of T are just strings of
symbols, and proofs are sequences of such strings that obey the formal rules of inference.
A finitary study of the formalized proofs of T that shows that it is impossible for a string of
symbols that expresses a contradiction to be the last line of a proof. (Ferreirós, 2007)
An Overview of Gödel’s First Incompleteness Result
An extremely clever development of metamathematical methods, the arithmetization of
metamathematics, allowed Gödel to produce contrary results to Hilbert’s program, thus proving that
systems such as axiomatic, set theory or Dedekind–Peano arithmetic are incomplete. Gödel’s first
incompleteness theorem states, in any language L strong enough and consistent to express 0, the
successor function S, +, <, and ∗ can express all primitive recursive functions, and thus no program can
prove or disprove all statements. In particular the language of the ℕ, Peano arithmetic, cannot have its
statements neither proved nor disproved with the system.
Gödel’s Theorem proves in a corollary of Proposition VI, that “there are arithmetical propositions
which are undecidable (i.e. neither provable nor disprovable) within their arithmetic system. Namely, the
proposition that g expressed by the formula is an arithmetic proposition. However, the proposition that g
is undecidable within the system is not an arithmetic proposition, since it is only concerned with
provability within an arithmetic system” (Gödel, 1962). This is a metamathematic principle and therefore
not an arithmetic concept. Hence Gödel results are not mathematical but rather metamathematical.
Fundamental notions of mathematics exhibited by Frege and Peano, Whitehead and Russell’s Principia
Mathematica including arithmetic as a deductive system starting from a limited number of axioms. Each
theorem is shown to be deducible by the axioms and theorems which precede it are in accord with a
limited number of rules of inference. Hence a represented deductive systems will entail a sequence of a
formulae, whereby a calculus. The initial formulae will express the axioms of the deductive system and
Koffi9
each of the other formulae. The purpose is to express the theorems which are obtained from the initial
formulae by a sequence of symbolic manipulations (Gödel, 1962).
Gödel had to be precise and detailed about the exact deductive system of arithmetic in order to prove
the undecidability of some arithmetical propositions. Hence for Gödel a proof representation is a series of
formulae is an immediate consequence of the formulae proceeding it. His proof schema included Gödel
numbering, in which the objective was to provide a unique decoded number for every possible wellformed sentence hence a formal language. The reason for this initiative was he wanted to be able to treat
any well-formed sentence as part of any other wff and hence wanted to establish proofs which required
this.
Godel assigned to every character allowable in the well-formed formula a number, a large number
construction from multiplying together pre-selected prime numbers. Prime numbers are special in the
following way in that, every number in existence can be factored down to the set of primes which when
multiplied together will generate that number and hence a prime number cannot be factored further down
than itself and 1. Gödel speaks of his rule as establishing a “one-to-one” correspondence as his rule of
arithmetization ensures that in every class of strings, there exist a correspondence in a unique class of
Gödel numbers, and vice versa (Gödel, 1962).
Interestingly he defines a gradual sequence of arithmetical concepts which correspond, according to
his rule of arithmetization. This is thus the metamathematical concepts expressed by the same arguments.
For example, he defines an arithmetical binary operation ∗ upon two numbers x and y. The result of
performing this operation is the Gödel number of the string obtained by taking the string whose
Godel number is x and placing the string whose Godel number is y immediately after it (Gödel,
1962).
The idea of recursion played an essential role in matamathematics because of Gödel’s clever
development. Hence, the method of the recursive definition is an extension of the method of
Koffi10
“mathematical induction” by which the natural numbers are define, step by step. Now mathematical
induction is not a rule of inference of predicate logic but is sometimes considered as a rule of "metalogic".
Thus, it is a rule we can use to prove properties about axiomatic systems in a more generalized from
rather than just proving statements solely within axiomatic systems.
“A recursive definition is the specification of the each number sequence of numbers by means of a
specification of the first number and of a rule which specified the (k+1)-th number in terms of the k-th
number and of k itself” (Gödel, 1962). Hence an arithmetical function is recursive if it is the last term in
a finite sequence. With each function being recursively defined by a rule where functions precede it, the
recursive structures of other arithmetic concepts is defined by means of a concept of a recursive function.
For metamathematics, the importance of recursiveness in general lies with reason that recursive definition
enables every number in a recursively defined infinite sequence to be constructed according to a rule.
Let Q(x,y) be the recursive relationship between the G-numbers of x and y, equivalent to Q’ (x,y(u))
by the modified arithmetization and x is a string whose G-number is x. Hence we have the following
proposition:
Proposition VI: If the formal system P satisfies certain condition of ‘consistency’, then there is at least
one recursive class-sign r(v) in P such that neither v Gen r(v) nor Neg[v Gen r(v)] is provable within P
(Gödel, 1962).
Conclusion
Despite the fact that almost all of Gödel’s proofs were explicitly finitary, he wanted to emphasize that
the “objectivistic conception of mathematics and meta-mathematics in general, and of transfinite
reasoning in particular, was fundamental to my other work in logic.” Gödel kept his focus on fundamental
questions, and with a concrete and deeply satisfying conception of incompleteness, he had the remarkable
capability to enhance our philosophical understanding. However, Hilbert was a more absolute and
consummate mathematician, believing with an unbounded faith in the ability of mathematics to solve all
Koffi11
problems. However, Gödel’s incompleteness result showed that mathematics is inexhaustible and thus
incomplete, and hence no matter which set of axioms is chosen as a foundation, one can always find
questions that those axioms cannot answer.
Koffi12
An Introduction to Gödel’s Theorems (Chap. 1-21)
i)
Truth vs. Provability
What is the meaning of truth vs. provable in mathematics? Why does the context of provability matter
if the truth of a given mathematic property of natural numbers hold for each of its sequences? The notion
of obvious truths in arithmetic, for example, provides us with a definite structure of sequential certainties.
However, that does not directly infer that its properties are automatically provable. Because axioms
contain enough information to enable the truth-value of any basic arithmetic proposition, any such
theorem β will be logically deduced to proclaim whether β is true or its negation – β is false. In this
particular realm of mathematics, the general idea that mathematically truth preserved structures can be
either logically deducible from axiomatic properties to be not only true or false but also provable or not is
the very core of what will be explored. The work of Gödel provides us with an outlook to notice that
certain arithmetic properties integrated within a truth based structure cannot be provable thus it is
incompleteable. Hence, most axiomatic systems have inherent limitations in proving arithmetic. Gödel’s
First Incompleteness Theorem shows that the entire natural idea that we can give a complete theory of
basic arithmetic with a tidy set of axioms is wrong. In this expedition of incompleteness, the set of natural
numbers will be our main consideration.
A set T has a nice set of true axioms and a reliably truth-preserving deductive logic--i.e. everything T
proves must be true, therefore T is a sound theory. Because the importance of sound or tautological sets
denote its inexplicable truth in every calculated sequence, it is critical to note that its properties do not
automatically imply that it is necessarily complete. An important understanding of what it means for a
sound property to be complete also means that its property is provable. Although Gödel provides us with
a sentence ℊ couched in the language of basic arithmetic such that assuming T is really sound, we can
show a sentence ℊ can’t be derived in T, and therefore ℊ must be true. A sentence � encodes the claim
that that very sentence is unprovable (Smith, 2013). Hence a theory T is either sound + provable = a
complete theory, or sound theory + un-provable = an incompleatable theory. Another interesting remark
Koffi13
to note is that the foundation of any such property/theory implying its provability should always meet the
basic need of being sound. Therefore, it’ll never be a question of how to justify that theory T is sound or
not but rather if it is provable. Hence, T being deductively sound however not provable in ℊ, means T is
not negation complete-i.e. provable to be true or provable to be false, nor a complete theory. Therefore T
is consistently true (sound) yet incompleatable.
ii)
Functions and Enumerations
Among the most powerful forms of relating a set of components to another set resides in functions.
Given a formal definition, we state that a function is a relation for which each value from the set of the
first components of ordered pairs is associated with exactly one value from the set of second components
of the ordered pair. The consideration of functions are critical to our understanding of numerical
descriptions as it denotes whether a given property holds true or false. For example given a function
� � = � 2 we will denote that if any given � produces an even number, it will be considered true for
some property of even numbers and false if otherwise. This is not to be confused with sound theory as
every enumeration of a function still produces a tautological sequence however if we limit our capacity to
which a certain set δ satisfies a given criterion then we are bounded within a particular property—e.g. an
even or odd result, that denotes whether it is true or false. This directly leads to the concept of a
characteristic functions which states that a numerical property/relation P is the one place total function
cp(n)=0 or 1 where P is a property of being either true or false (0, true) (1, false). Because characteristic
functions provide a descriptive result—i.e truth or otherwise, we note that total functions map each and
every element of its domain to some unique corresponding value in its codomain i.e. range. Examples of
total functions are subjective (onto), injective (one-to-one), and bijective (one-to-correspondence. As
previously mentioned, every enumeration of a function produces a tautological sequence where a set ∑ is
enumerable iff its members can be listed off in some numerical order and the list may be infinite hence the
set ∑ surjective where f: ℕ→ ∑ (ℕ maps to ∑)—i.e every codomain has at least one matching domain.
Koffi14
iii)
Cantors Diagnolization Argument
Because an enumerable set ∑ can be listed off in some numerical order and the list may be infinite, we
now introduce the idea of Cantor’s Diagnolization argument. Now Cantor's theorem about the
indenumerability of the set of total functions from ℕ to ℕ uses the argument that if we assume that all
functions from ℕ to ℕ are denumerable—i.e. countable, then we should be able to form a new function
that is a bijection from the original one. Recall section: Cantor and the Paradoxes.
How does this relate to the idea of natural numbers? Let’s suppose we have a finite set where we could
simply count the elements of some set £ and find the cardinality—i.e. the number of elements in the set
and now let’s suppose we have an infinite set. With an infinite set we are not able to simply list its
number of elements because it is unlimited. However a different way to compare their sizes would be to
use the concept of a total function—i.e. a bijection. We recall that a bijection is a one-to-one
correspondence where we compare one set of elements between sets—i.e. we compare one set of
elements to another and vice versa. The result of a bijection between two sets will yield the conclusion
that they have the same size—i.e. they are countable; otherwise, one set will be bigger or more infinite
than the other. For example, a set of ℕ say Ҩ = {1,3,5} will directly correspond with set Ҕ={2,4,6}
provided we have a function say � � = � +
as its inverse function that will map the elements in Ҕ
back to Ҩ. Provided we have a bijection between two sets, we denote they are countably infinite, however
let’s look at a set that is uncountable—i.e. Cantor’s Argument. We will use the argument of real numbers
ℝ. Let S be a set which contains every infinite sequence of 0 and 1—i.e. a decimal expansion of a real
number. Suppose S is countable, so we must be able to order them. As we present different layers of S1=
{0 1 0 1 0 0 1 1….}, S2= { 1 1 0 1 0 0 1….} and so on, we form a diagonal from an x number of layers
where we have a set Sp= { 1 0 0 1 0 1 1….} where the flipped diagonal yields a different set of strings Sn=
{ 0 1 1 0 1 0 0….} which is now different every other sequence in the ordering—i.e. we cannot formulate
a one-to-one correspondence between the two sets. However, our original assumption was the claim that
Koffi15
every element of S is in the set Sp, and Sp is surely within the set S but it is not a bijection with the other
subsets, thus we have a contradiction where S is uncountable and therefore infinitely uncountable. We
notice that we have the same occurrence with any set of ℕ’s, where we can create a power set Ƥ (ℕ)
which contains all the subsets of all ℕ which are also uncountable—i.e. infinitely uncountable. Hence,
taking the power set always results in a bigger set, and we conclude that there has to be infinitely many
sizes to infinity, and therefore indenumerable.
iv)
Effectively Computable and Effectively Enumerable
As previously mentioned, the consideration of functions are critical to our understanding of numerical
descriptions as it denotes whether a given property holds true or false. Hence, a set can be enumerated by
an effectively computable function. The central idea of an effective computable function involves the
execution of an algorithm which successfully terminates and hence involves an entirely determinate
sequence of discrete step-by-step procedures. The set ∑ is effectively enumerable (e.e.) iff either ∑ is
empty or there is an effectively computable function that enumerates it. For example, an e.e. infinite set,
the computable function � � = n2 effectively enumerates the natural numbers which are perfect squares.
The importance the word effective merely signifies that there exist an algorithm where given enough time,
it would successfully terminate. Hence, we denote that a total functions f: £→ β is effectively computable
iff there is an algorithm which can be used to calculate in a finite number of steps. This is particularly
important within the study of incompleteness because the central idea resides in whether there is an
algorithm that will successfully terminate any theorem in arithmetic as will see that it does not. Since
algorithmic sequences not only compute functions but also decides whether a property holds, we begin to
actually answer if a given property is either true or false. The term decidability holds a significant weight
to numerical functions because it will be seen how important property/relation applies to any
appropriately given item. Therefore, a numerical property or relation is effectively decidable iff its
characteristic function is effectively computable. This directly leads to the theorem that if is an effectively
decidable set of numbers, it is effectively enumerable, however its converse is not true because any
Koffi16
effectively enumerable function will successfully terminate however it will not automatically infer that a
given property is decidable true or false—e.g. even vs. odd determinates of a function where a function is
decidable to be true only if it is even, in this case.
v)
Effectively Axiomatized Theories
Within the assessment of complex arguments we need to develop a formalized structure that is
completely free from any obscurity where a suitable artificial language reveals a logical structure. The
infusion between mathematical and logical structure aims to show perfect clarity and honest inferences as
a theory regimented by a tidy set of axioms will lead to the idea of an axiomatized formal theory—i.e. a
theory built in a formalized language. Within the correctness of a stated proof, we want to show that it is
explanatory—i.e. clear with no ambiguity. Because our study of incompleteness entails the notion of an
incomplete theory, Gödel notes the absolute importance of a deducible and formalized language. Since we
have defined a formal language L as being a pair <ф, µ>, where ф is a syntactically defined system of
expressions (words) and µ gives the interpretation (image) of these expressions, its semantics effectively
tells us the condition under which a given sentence is true or false. Hence, if we think of a language
strictly from a logical perspective, a claim or statement written within a formal language follows that
there is an effective way of conveying whether a condition actually holds and whether the sentence really
is true—i.e. valid or sound.
An effectively axiomatized theory being effectively enumerated is not to say that the theory is
decidable. This is very interesting because the concept of a decidable property holds whether a property is
true or not but effective enumeration of an axiomatized theory does not automatically guarantee that a
property will automatically be decidable to be true or false. However, a theory built in a formalized
language will be deducible to generalize any mechanical method for any theorem eventually. This leads to
one the most important theorems which states, Any consistent, effectively axiomatized, negation-complete
theory T is effectively decidable. Because T, in this case, is negation-complete, its theory is therefore
either true of false. So to decide whether a sentence Ҩ of T’s language is a T-theorem—i.e. where Ҩ is a
Koffi17
theorem (hypothesis) of T, we effectively list theorems until we capture either Ҩ or –Ҩ. It must be noted
that every modern mathematical theory begins from a given set of non-logical axioms (i.e. hypothesis)
where a developed principle theory could be axiomatized and formalized within the language of logical
formulas (Smith, 2013).
vi)
Expressive vs. Capturing
It is noted that a theory T is expressive as it depends on the richness of the theory’s language. Hence,
an interpreted formal L is sufficiently expressive iff it can express every effectively computable one-place
numerical function. Because an effective computable function successfully terminates we note that the
richness of its language depends on whether it can be express T characteristic function. Another important
theorem which states, the set of truths of a sufficiently expressive language L is not effectively
enumerable, leads to this example. Say K is the set of all even numbers and K’s complement is the set of
all odd numbers. Because a function � � = � example only enumerates even outcomes however not
odd ones. Although we remember that language L is still sufficiently expressed because it can express
every computable one-place numerical function. The central importance of denoting capturabilty entails
that fact that if we merely capture a sentence P in a sound theory T, then Ҩ(x) expresses P—i.e. T is
provable to be true or false. We therefore have a complete theory. In terms of performing how to succumb
to an incomplete theory we must weaken one assumption from soundness to mere consistency, where a
formal theory is still sufficiently strong if it captures all effectively decidable numerical properties.
However, in terms of an undecidable theorem we use the theorem which states, no consistent, sufficiently
strong, effectively axiomatized theory is decidable.
Because everything depends, for starters, on whether the ideas of a ‘sufficiently expressive’ arithmetic
language and a ‘sufficiently strong’ theory of arithmetic are in good order, Gödel’s proof tells us how to
take a consistent effectively axiomatized theory T and actually construct a true but unprovable-in-T
sentence. Because Gödel did this by finding a general method that enabled him to take any theory T
strong enough to capture a modest amount of basic arithmetic and construct a sentence ℊ, it thereby
Koffi18
encoded the claim that the sentence ℊ itself is unprovable in the theory T (p.53, Smith). So ℊ is true if
and only if T can’t be prove it. The methods of how Gödel constructs his sentences will be constructed by
mapping a set of symbols where a Gödel number will appear and we will see how his construct will
include a sentence ℊ.
vii)
Two Formalized Arithmetics
Because the language of a formal system is an essential step of Gödel's proof, it precisely defines and
fixes a correspondence of a particular type between the expressions of that language and the system of
natural numbers. Hence, a code sequence or Gödel numbering of the language will suffice for wellformed formulas and consequently as sequences of primitive symbols were each are assigned a unique
number. BA’s language of non-logical vocabulary LA = < LB, IB >, where LA is the first-order language of
basic arithmetic, is the same as that of LA. Non-logical vocabulary signifies a constant that is not a logical
constant nor a symbol of grouping i.e. logical variable. For instance, a variable standing for a number is a
non-logical constant because it cannot be replace by a statement. Hence, in the context of mathematical
logic, a non-logical symbol signifies a symbol (Smith, 2007).
Since the language of BA, LB, is a subset of the language LA, they have the same non-logical
vocabulary, namely the symbols 0, S, + and ×. Hence, it only knows about the successor, addition and
multiplication functions. Although LB lacks the apparatus of quantification i.e. missing numerical
variables and quantifiers, it still contains the identity and equality symbol. The corresponding components
of LA and the interpretation of LB are identical. Because the inference rules of propositional logic can be
used to deductively deduce the language of BA together with rules for the equality predicate, Leibniz's
law enables us to infer the sentence obtained by substituting in φ for every occurrence of ρ by τ for
example, where the identity ρ = τ, where ρ and τ are terms (Smith, 2007).
The natural number sequence and the addition and multiplication functions are defined by BA 6 nonlogical axioms. “Since BA does not contain quantifiers nor variables, each axiom is a template
Koffi19
or schema using the Greek letters α and β as placeholders for numerical expressions.” Hence for each
generated sequence, a schema generates an infinite number of sentences, where one is a possible
substitution of the Greek letter(s).
For instance, the first sentences generated below demonstrate the first schema 0 ≠ S0, 0 ≠ SS0, 0 ≠
SSS0, 0 ≠ S(S0 × SS0), etc.” The following are the 6 axiom schemata of BA:
Schema 1: 0 ≠ Sα
0 is not the successor of any number
Schema 2: (Sα = Sβ) → (α = β)
by contraposition, any two distinct numbers have
distinct successors
Schema 3: α + 0 = α
base case for the definition of addition
Schema 4: α + Sβ = S(α + β)
recursive case for the definition of addition
Schema 5: α × 0 = 0
base case for the definition of multiplication
Schema 6: α × Sβ = (α × β) + α
recursive case for the definition of multiplication
As shown, in an axiomatized formal theory, the requirements for BA are met. “Hence, it is decidable
whether or not a sentence of LB is an instance of an axiom schema of BA. Hence, Smith proves that BA is
negation-complete i.e. for any sentence φ in LB, either BA ⊢ φ or BA ⊢ ¬ φ. In addition, Smith proves
that BA correctly decides all sentences in LB.” (Smith) Therefore, we can inevitably infer that BA is
decidable, since BA is negation-complete.
In conclusion, BA is complete: it can prove every true statement that its language can express.
However, its language is rather limited. Specific statements, such as: 3 + 2 = 5, 7 ≠ 2 × 3, can be express,
the successor of 0 is not 0, the successor of 1 is not 0, the successor of 2 is not 0, and etc. However, since
it lacks the apparatus of quantification, it cannot express general statements, such as: all natural numbers
have a successor that differs from 0. Nonetheless, stronger AFTs of arithmetic can be acquired by
extending BA.
Koffi20
In order to make BA more expressive, its logical vocabulary must allow for variables and quantifiers.
By doing this, we merely obtain the interpreted language LA. Because Gödel’s proof requires the
expressions of formal language, quantification must be solidified. Hence, variables and quantifiers can
substitute the schemata of BA with regular axioms. The formal system of arithmetic is known as QRobinson Arithmetic. By definition, the interpreted language of Robinson Arithmetic is the language of
LA plus the interpretation IA.
The inference rules of Robinson Arithmetic will be some form of first-order logic which will include the
identity. The following are the axioms of Robinson Arithmetic:
Axiom 1: ∀x (0 ≠ Sx)
0 is not the successor of any number
Axiom 2: ∀x∀y (Sx = Sy → x = y)
by contraposition, any two distinct numbers have
Axiom 3: ∀x (x ≠ 0 → ∃y (x = Sy))
no element besides zero that does not have a
Axiom 4: ∀x (x + 0 = x)
base case for the definition of addition
Axiom 6: ∀x (x × 0 = 0)
base case for the definition of multiplication
distinct successors
successor
Axiom 5: ∀x∀y (x + Sy = S(x + y))
recursive case for the definition of addition
Axiom 7: ∀x∀y (x × Sy = (x × y) + x)
recursive case for the definition of multiplication
Although Q is not negation-complete, letting φ be the sentence that ∀x (0 + x = x) determines that Q is
sound since its logic is truth-preserving. However, Q cannot derive φ and Q cannot derive ¬ φ.
In proving this statement, an alternative interpretation for LA should make Q's axioms true but φ false
because if Q is sound, then it cannot derive φ. Under the original interpretation IA, Q is sound and ¬ φ is
false.
Koffi21
For Smith, “an extension of IA with two new elements a and b in the domain such that a is its own
successor, b is its own successor, and addition and multiplication are extended in such a way that 0 + a is
equal to b and 0 + b is equal to a, thereby falsifying φ.”
Since the interpretation of φ shows the addition of what Q cannot prove, it can be inferred that Q is a
weak theory of arithmetic. However it’ll be proven that Q is sufficiently strong as this property is defined
in terms of capturing properties. For instance, Q can derive all of the sentences 0 + 0 = 0, 0 + S0 = S0, 0 +
SS0 = SS0, etc. However, it cannot derive the universally quantified sentence φ. Smith mentions that any
consistent theory that extends Q will be sufficiently strong and hence negation incomplete.
Since any system in which Q can be deduced is essentially incomplete, then any such theory in
which Q is interpretable can be proved directly to be incomplete. For example, in set theory, coding
formulas and deriving sets i.e. Gödel sets, and proceeding in the same fashion. However, in order to
establish the deductibility of Q, the system must be an axiomatizable extension of Q where “a certain
amount of elementary arithmetic can be carried out within a system”. (Smith) Hence Q is considered to be
a weak theory as it cannot justify statements such as ∀x (0 + x = x).
viii)
Primitive Recursion
The formal theories of arithmetic seen thus far are the successor function, addition and multiplication
built in. Primitive recursive functions are considered as a subclass of the effectively computable functions.
Primitive recursive properties are also defined in terms relations i.e. those with a p.r. characteristic
function which can effectively decide when the relations and properties holds. Since primitive recursive
functions can be defined by a chain of definitions by recursion and arrangement.
“A full definition for the p.r. function f is a specification of a sequence of functions f 0,f1,f2,...,fk where
each fj is either an initial function or is defined from previous functions in the sequence by composition or
recursion, and fk = f.” Because every p.r. function has a simple method of proving that every p.r. function
has some property P implies that p.r. functions are in fact total functions. “According to Theorem 19:
Koffi22
Primitive recursive functions are effectively computable by a series of (possibly nested) bounded ‘for’
loops.” (Smith, 2007)
LA can express the following factorial function:
(Sx)! = x! × Sx
“Hence, this is the p.r. definition for the factorial where It guides us to construct a sequence of numbers
0!, 1!, 2!, . . . , x!” (Smith, 2007).
For loops are used to execute a lump of code a specific number of times. They are considered a main
programming structure, used to be produce computation by a program where, given a primitive recursive
function, its specified value can be determined and every loop iterates a specified number of times. This is
very important because this tool is what is needed to show that negation-complete theories are decidable.
This process of enumerating Theorem 19 will generate which of ϕ or ¬ϕ leads to the conclusion of a
computational decision procedure (Smith, 2014).
Because computer languages allow some process for functions to be iterated until a given condition is
satisfied, in this case no prior limit was made for a specified number of iterations to be executed. Hence, if
searches where in fact made to be unbounded searches as computations then it is plausible that not
everything computable will be primitive recursive.
Hence we have, “Theorem 21 There are effectively computable numerical functions which aren’t
primitive recursive.”
ix)
Peano Arithmetic
In order to derive the statement ∀x(0+x = x) which is beyond Q’s theory, a formal arithmetic that
incorporates some stronger axiom(s) for proving quantified wffs will suffice by using Peano Arithmetic.
Peano Arithmetic is considered as, “a formal arithmetic that incorporates some stronger axiom(s) for
proving quantified wffs.” Since Peano borrows its axioms from Dedekind which states that, “for every
Koffi23
partition of all the points… there is a point of one set which lies between every other point of that set and
every point of the other set”, the notion of natural numbers is defined "in itself. PA, in this case, it known
as the first order arithmetic where its objects that can assume only distinct, separated values i.e. discrete
math (Smith, 2007).
“PA – First-order Peano Arithmetic – is the result of adding to the axioms of Q the universal closures of
all instances of the Induction Schema. Plainly, it is decidable whether any given wff has the right shape to
be one of the new axioms, so PA is a legitimate formalized theory” (Smith, 2014).
Considering that the wff ∀x(0+x = x) is unprovable in Q, it can however be proven in PA by induction.
Assuming the natural style of deduction yields rules from UI (Universal Instantiation), UG (Universal
Generalization on ‘arbitrary’ names) and CP (Conditional Proof). Hence, starting with the instance of the
Induction Schema where ϕ(x) is replaced by (0+x = x) will help to prove the two conjuncts in the
antecedent of that instance in order to conclude ∀x(0+x = x). The formal version of the proof, “makes the
axioms of Q true while making ∀x(0+x = x) false has the object as a self-successor.” Hence, what is
needed is the absolute fundamental axioms in order to get standard arithmetical results. (Smith, 2014)
It’s been noted that in “Theorem 19: Every p.r. function is intuitively computable – moreover it is
computable using only ‘for’ loops and not open-ended ‘do until’ loops.” However, not all computable
numerical function is primitive recursive (Smith, 2014).
x)
Capturing Functions
The notion of computability in general concludes that, “addition can be defined in terms of repeated
applications of the successor function as multiplication can be defined in terms of repeated applications of
addition” (Smith, 2014). Hence, the idea of defining one function in terms of repeated application is
essentially the idea of defining a function by primitive recursion. Since the idea of repeated applications
of the successor function via programming structure executes a lump of code a specific number of times,
we can use this to see how functions are captured. Recall that a formal theory is still sufficiently strong if
Koffi24
it captures all effectively decidable numerical properties which is equivalent to capturing one place
functions. However, capturing a function, in this case, will consist of Q capturing any p.r. function using
Σ1 wffs (Theorem 17.2). Therefore, we note that with sufficiently strong theories, functions can be
defined in specifying a whole class of primitive recursive functions (Smith, 2007).
Hence, we want to show that Q and PA can at best capture all primitive recursive computable functions
which will prove sufficient in Gödel’s argument for incompleteness. In using a coding/decoding
technique, we create a sequence using the (primitive recursive) decoding function exp(b,i) i.e. Gödel’s βfunction:
This single number b can be thought of as encoding the whole sequence
k0, k1, k2, ..., kn…for this function returns the exponent of the prime
number πi (series of primes) in the factorization of b. By the construction
of b, then, exp(b,i) = ki for i ≤ n. Now let’s generalize. We’ll say A twoplace β-function is a function of the form β(b,i) such that, for any finite
sequence of natural numbers k0,k1,k2,...,kn there is a code b such that for
every i ≤ n, β(b,i) = ki. (Smith, 2007)
So the idea essentially is selecting a corresponding code number b in being the primary argument for β,
where the functions will interpret it and give a required sequence to the members in an orderly manner.
An example could be the use of the division algorithm, � = � + , where the idea of multiplication and
addition on the remainder of division can be elementarily. Hence, L A can express all p.r. functions by
using the β concept.
Koffi25
Gödel’s Incompleteness Theorem
Are there contradictions in mathematics? Since the idea of proofs are supposed to establish certainty, a
single contradiction would eliminate all certainty. Suppose we have a system of logical rules which tells
us we can make deductions. While Hilbert’s ultimate goal was to prove that the foundations of
mathematics was free of all contradiction, Gödel’s results would prove otherwise. Hence, in a consistent
mathematical system, there are results which are true but not provable. Using a clever coding system by
way of representing mathematics by integers, Gödel created a sentence g, stating “this sentence is not
provable”. However, if that statement could be proven, we will show in his proof that a contradiction
arises and therefore the statement cannot be proven, but it is true. Gödel created his own version of the
liar’s paradox and proved that it is impossible to obtaining a complete axiomatization of arithmetic
In retrospect, the fundamental logical concepts seen thus far are the following:
Systems of Formal Languages
The original idea of a formal language is that, Gödel’s proof can be completely and precisely defined.
Hence, it requires that every symbol used must be defined in a language which is included but not limited
to commas, periods, colons and etc.
Proofs residing in a Formal Language
We must understand that a formal language, in and of itself, does not prove anything. Thus, we must
define exactly what a proof in the formal language means; which will consist of formal sentences, wff’s,
decidable properties, capturing functions and etc. In a proof, it is rather evident that sentences interact
with other sentences in which there must be some direct association among the sentences being proven.
Koffi26
Rules of Inference
Rules clearly stating how one sentence can follow from one or more other sentences are call rules of
inference. We use formal expressions only as a combination of symbols where the rules are only justified
in how to evaluate one or more expressions i.e. combinations of symbols. For example, if we had
sentences “le garçon va” and “à l'école”, then we can create a rule saying that you can combine the first
three words (symbols) of the first sentence with the second two (words) symbols of the other sentence.
Axioms of Formal Systems
Since axioms are considered as fundamental expressions of the formal language requiring no proof, every
proof has to use at least one of them.
Proof Sequences in a Formal System
In the beginning of a formal language proof, starting with at least one axiom, a set of rules is then
followed to make a new sentence. By doing that several times, we get a new sentence and etc. where
every sentence constructed is proved. Hence, every sentence that has been proven will have a list of
sentences that go together to prove it.
Consistency
In order to avoid inconsistencies and contradictions, we must built a formal language where the proofs
are well defined, such that one cannot prove a sentence as well as the opposite of that sentence. Hence,
Gödel’s proof only applies to formal languages that do not result in contradictions. Can we then use the
system to show that there cannot be a contradiction among results we can prove using the system? Gödel
would say yes! A logical system can prove that it itself is consistent if and only if it is not inconsistent.
Hence, if a system is consistent and it doesn’t contain any contradictions then we cannot prove that it is
inconsistent. However, if our system is not inconsistent, then we can prove that it is consistent.
Koffi27
Incompleteness and Completeness
By defining a consistent formal language referring to numbers and expressions that can be interpreted,
the goal is to make propositions about those numbers. Hence, a formal language proof system is
considered complete such that every proposition about numbers can be made and interpreted in the formal
language, and there will always be a proof of that proposition or of its negation. Gödel’s proof thus shows
that any formal language, regardless of the axioms chosen, and proof rules, can always define a formal
expression such that the sentence nor its negation can be proven by those axioms. In addition, if Gödel’s
proof is correct, then it infers that every formal language proof system must be ‘incomplete’.
Gödel’s Numbering/Coding
Gödel’s numbering system represents any sentence of the formal language as a number. This implies
that in a formal language, every sentence can be represented as a number. Hence, Gödel fixes a
correspondence, i.e. diagonalization, of a particular type between the expressions of that language and the
system of natural numbers. Gödel’s code sequence of symbols is matched to a unique number, such that
no two combinations/arrangements can have the same number and also, no two numbers can match to the
same combination. This is equivalent to an earlier statement where a code sequence or Gödel numbering
of the language will suffice for well-formed formulas and consequently as sequences of primitive symbols
were each are assigned a unique number.
The process goes as follows. This list has not been exhausted.
¬
1
3
5
→
7
↔
9
∀
11
∃
13
=
15
(
17
)
19
0
21
S
23
+
25
×
27
x
2
y
4
For example, suppose we take an example of a formal sentence (x + y) where we match number to its
symbols:
(
17
x
27
z
6
Koffi28
+
25
y
4
)
19
The following sequence are numbers, 17, 27, 25, 4, and 19 where putting them together makes
172725419. However, this method could be distorted in which the original sequence could become 1, 7,
2, 72, 54, 19. In order to retrieve the original information, Gödel devised a clever way of converting these
sequences of numbers in a way that retains all the original information. Gödel’s system takes the prime
numbers in order, so that for example, to encode a sequence of five numbers, we start off with the first
five prime numbers – which are 2, 3, 5, 7 and 11. Hence, we have the following:
217, 327, 525, 74, 1119
Multiplying 217 × 327 × 525 × 74 × 1119, we get:
which gives the matching Gödel number for the sentence (x + y). In ordering to derive the original
numbers 217 × 327 × 525 × 74 × 1119 first divide by the first prime number, which is 2. Continue to divide
by 2 until you don’t get an even dividend.
Creating a Special Correspondence with Numbers
The idea, essentially, is as follows: any formal system can be linked with a Gödel number and thus any
formal proof can prove that sentence. Hence, there exist a matching Gödel number for that proof. Gödel’s
system of corresponding relationships between that proof and the sentence that it proves is always well
defined.
Symbols signify a name or meaning for particulars where in this case, we can represent symbols for
sentences and variables. For instance, stating names for particular characters such as Jack and Jill is
essentially equivalent to saying ‘3 and 5’, where 3 and 5 are names for particular numbers. Names can
also signify variables such that saying a goat and a pig is also equivalent to saying a number x and a
number y, provided that x and y are variables for numbers.
Koffi29
What would then constitute as a Gödel proof can be defined in general terms, for instance:
The relationship in the sentence P is the proof of the sentence Q where P and Q are variables for formal
sentences is the particular correspondence of interest. Thus, given any two formal sentences, one can
substitute them for variables P and Q, where the proposition is either correct or incorrect. Let’s suppose
that P is (y + 7), and Q is (3 + x) which gives the expression (y + 7) is the proof of (3 + x). This
proposition is wrong however we are simply illustrating a principle.
This Gödel proof has a particular relationship with numbers where the relationship between the wellformed formal sentences P and Q means that GN(P) is the G-proof of GN(Q) and it is a relationship
corresponding to the expression P is the proof of the sentence Q. Note that in reference to Gödel’s
numbering system, we will use GN. F is a free variable meaning the value of any formal symbol
sentence, where the free variable of the Gödel numbering system is GN(F). The name GN(F) is simply an
expression of a Gödel number that matches to the formal sentence F.
In the earlier example (x + y) has the matching Gödel number:
The GN expression of this sentence is GN((x + y)) and now we can notice a relationship between a formal
sentence and a formal proof.
Matching Formal Sentences with Numeric Relationships
Any numeric relationship with at least one free variable will always have a matching formal sentence
with at least one free variable. That formal sentence will then express the same theory as that numeric
relation. Since that formal sentence has a matching Gödel number, that number relation will also have a
matching Gödel number.
Koffi30
For any pure numeric relationship, it is a rather simple relationship between numbers that is expressed
in a language that is not necessarily formal. For instance, there are many different ways of expressing
binary relations such as multiplication and addition. For instance, the result of x and x is four, x by x is
equal to four, x plus x is four and so on. Therefore, a numeric relationship should be clearly defined such
that it can only be translated into the formal language in which the formal sentence has a matching Gödel
number.
Take the number relationship ‘x plus x is equal to four’ and suppose that in our formal language the
matching formal sentence is:
(x + x = SSSS0)
In retrospect, the formal system 0 = zero, S0 = one, SS0 = two, SSS0 = three, and etc. That means the
matching Gödel number for ‘x plus x is equal to four’ or (x + x = SSSS0) is:
217 × 327 × 525 × 727 × 1115 × 1323 × 1723 × 1923 × 2323 × 2921 × 3119
Since a number relationship with a free variable is not a proposition, that variable substituted by some
specific value will validate a proposition. Substitution takes one value and gives another matching value.
Hence, substituting the free variable by some number and then substituting the free variable from the
formal sentence by the same number will yield a certain number relationship that is a proposition about
numbers having a matching formal sentence. Substituting variable x by two, for instance, would
yield ‘two plus two is equal to four’, and the matching formal sentence would be (SS0 + SS0 = SSSS0).
Gödel would then stipulate the following: “If a formal sentence is provable, then the matching
number relationship must be true and for any sentence of the formal language, either there’s a formal
proof of that sentence or there isn’t.”
An Alternative Numbering System
Koffi31
At this point in the game, a new function is needed such that it is similar to the Gödel numbering
function, which will be used in Gödel’s proof. Gödel’s numbering function is not a numeric relationship,
meaning that the symbols of the formal language are represented as the symbol for equals, plus, and etc.
Let’s suppose we would like to use Gödel’s numbering function but only for symbols of the formal
language that in fact are numbers. Consequently, we’d like to call Gödel’s numbering function GN(x),
such that substituting x only by symbols of the formal language would represent numbers. However, this
still wouldn’t infer that GN(x) is a numeric relationship since the definition of the Gödel numbering
function still refers to symbols rather than numbers. Hence, in order to establish a working proof for
Gödel, we would need a function that is similar to the GN(x) function, but one which is purely a number
relationship.
Therefore, we need a new function which will only take numeric values for free variable, provided
that entering only numbers would mean that the function would only result in number values. The
goal would be to get a new function that could do the exact same thing as the Gödel numbering function.
Since this new function refers to numbers and variables representing numbers, then it is a number
relationship. Let’s call the new function a Simple Numbering function, or SN(x), where x is a free
variable. In Gödel’s original proof (relation 17), Gödel calls this function Z(n). This implies that for
whichever symbols would represent symbols for a number, the SN function would gives us the correct
Gödel number.
The Proof
In retrospect, Gödel’s proof has a particular relationship with numbers where the relationship between
the wff sentences P and Q (also having a number relationship) means that
GN(P) is the G-proof of GN(Q) ; which leads to the expression, P is the proof of the sentence Q where it
is possible that there could be a formal proof of that sentence. Equivalently, if P is the matching Gödel
number for a sequence α, and Q is the matching Gödel number for sentence β, there is matching
relationship saying that, P is the proof of the sentence Q and there is possibly a number that is the
Koffi32
G-proof of the sentence that is either true or false. However, if a code sequence of a formal sentence isn’t
a Gödel number, then there cannot be a G-proof of that sentence.
Recall that substitution takes one value and gives another matching value. An example of substitution
could be the sentence (x + 3) where substituting x by 5 produces the sentence (x + 3). This expression
took two values, one for the sentence to be replaced and the other for the number that’s used for the
substitution. This would result in two free variables where we can call it
Substitution (Sentence, Number). Hence, if the sentence has only one free variable, then we’re provided
a specific substitution. For our example, we have: Substitution ((x + 3), 5) where the value of that
is (5 + 3).
Gödel’s formulation of his proof involves the concept of substitution; where taking a number
relationship, and substituting it by its own matching Gödel number would then generate with a new
relationship which appears to self-reference its own Gödel number. First, a relationship of substitution
G-Substitution (Numeral , Numeral ) needs to be defined.
A relationship that isn’t a numeric relationship would imply there is no G-proof of G-Substitution (x,
GN(x)). Because we want a relationship that is equivalent, we define another relationship in which its
equivalent statement also says there is no G-proof of G-Substitution (x, SN(x)). Recall that SN is the
Simple Number function which is a numeric relationship, defined in terms of numbers and variables only.
Also, since we have a numeric relationship, there will always be a corresponding formal sentence for this
relationship. Hence, we can call SN a formal sentence FS. Finally we say that the formal sentence FS will
be a corresponding Gödel number GS.
We can see that all of SN, FS, and GS are self-referential however, SN and FS both have a free
variable where GS is only a number. The trick Gödel uses at this point is to substitute the free variable of
the Simple Numeric relationship SN by the Gödel Sentence GS. This would result in another relationship
stating the following: there is no G-proof of G-Substitution (GS, SN(GS)), equivalently SN(GS), where
Koffi33
there is a corresponding formal language/sentence FS(GS). At this particular point we want to show if
there is a proof in the formal language FS(GS) or not.
Suppose that there is a proof of FS(GS) in the formal language such that the corresponding numeric
relationship SN(GS) states, there is no G-proof of G-Substitution (GS, SN(GS)) must be true! However,
if that is the case that there cannot be a formal sentence where the free variable of the formal sentence FS
is replaced by the GS number by virtue of GS being only a number then we have a contradiction! There
cannot in fact be any proof of the formal sentence FS(GS) that would mean the free variable of the formal
sentence FS is replaced by the number GS. Therefore, we can say this relationship is true by semantics,
but not provable and thus incomplete.
Koffi34
Works Cited
Ferreirós, José. "II.7 The Crisis in the Foundations of Mathematics." The Princeton Companion
to Mathematics (2007): n. pag. Web.
Gödel, Kurt. On Formally Undecidable Propositions of Principia Mathematica and Related
Systems. New York: Basic, 1962. Print.
Meyer, James R. "A Simplified Explanation of Gödel's Proof - Part 8." LL: Gödel Simplified: 8.
Web. 04 Feb. 2016.
Smith, Peter. "Gödel Without (Too Many) Tears." Logic Matters (n.d.): n. pag.
Http://www.logicmatters.net/. 20 Feb. 2014. Web.
Smith, Peter. An Introduction to Gödel's Theorems. Cambridge: Cambridge UP, 2007. Print.
Smith, Peter. An Introduction to Gödel's Theorems. Cambridge, England: Cambridge UP, 2013.
Print.
"Summer of Gödel." Summer of Gödel. N.p., n.d. Web. 26 Jan. 2016.
.
Koffi35