IOSR Journal of Mathematics (IOSR-JM)

e-ISSN: 2278-5728, p-ISSN: 2319-765X. Volume 13, Issue 5 Ver. I1 (Sep. - Oct. 2017), PP 32-40
www.iosrjournals.org

Real Analysis of Real Numbers- Cantor and Dedekind Real

Number Structuring
1
Ahlam Khalaf Jabbar, 2Chandana Kumar, 3*S.Balavidhya
1
(Department of Mathematics, Acharya Institute of Graduate Studies, Bangalore, Karnataka, India)
2
(Assistant Professor, Department of Mathematics, Acharya Institute of Graduate Studies, Bangalore,
Karnataka, India)
*3
(Assistant Professor, Department of Mathematics, Gopalan College Of Engineering and Management,
Bangalore, Karnataka, India)

Abstract: In this paper, the real number system is reconstructed with the Peano’s axioms, to create systems of
elements that demonstrate the properties of integer and rational numbers. Construction of two real number
systems named Dedekind Real Number System that shows the completeness of the order and Cantor Real
Number System that shows the Cauchy completeness are done and studied. During this process of systematic
construction of real numbers, the integers and rational numbers can also be obtained. The process of
constructing a real number system is made through mathematical concepts.
Keywords: Peano’s Axiom, Archimedean Property, Order Completeness and Cauchy Completeness.
----------------------------------------------------------------------------------------------------------------------------- ----------
Date of Submission: 12-09-2017 Date of acceptance: 22-09-2017
----------------------------------------------------------------------------------------------------------------------------- ----------

I. Introduction
The real number system possesses the Archimedean property and it is Cauchy complete. The real
number system such as the natural number system reveals that the fundamental properties play an essential role
in the construction process. This work enables to show that the real number system is logically necessary and
gives out the importance of the real number system.
The real number system is reconstructed with the Peano‟s axioms to demonstrate the properties of
integer and rational numbers. The real number system includes all the rational numbers including integers,
fractions, irrational numbers, transcendental numbers and the square root of 2 also get included in the real
numbers.
Peano‟s axiom describes the natural number system in an efficient manner. Axioms are used to build a
concrete natural system; they also do not pose any problem with reference to the definition of the Integer
System and Field of Rational. This can be well explained with a preceding system. Initially real numbers
consisted only of the rational numbers because irrational numbers were not derived. A system which is closed
with basic mathematical operations is possible by carrying out it from the natural system which is not closed.
Richard Dedekind‟s (1831-1916) and Georg Cantor‟s (1845-1918) intuition can be relied upon to help
with the construction of real numbers. Therefore, these are classified as the Dedekind Real Number System and
Cantor Real Number System. In this paper Peano‟s axiom, Dedekind Real Number System and Cantor Real
Number System are used for reconstruction of real number system. Organization of the paper is with respective
sections: Preliminaries, Properties of Dedekind Real Number Systems,

II. Preliminaries
Definition2.1: The five axioms of peano: We assume the existence of a set N with the following properties:
(i) There exists an element 1∈N.
(ii) For every n ∈ N, there exists an element S (n) ∈N such that {(n, S (n))| n ∈ N} is a function.
(iii) 1∉ S (N).
(iv) S is one-one.
(v) If P is any subset of N such that 1∈ P and S (n) ∈ P ∀n ∈ P, then P=N.

Definition 2.2: The Dedekind’s cut: A subset α of Q is called a cut if the following conditions are satisfied:
(i) α ≠ ∅, α ≠ Q.
(ii) For every r ∈ α and s ∈ Q\α, r < s.
DOI: 10.9790/5728-1305023240 www.iosrjournals.org 32 | Page
 Real Analysis of Real Numbers- Cantor and Dedekind Real Number Structuring

(iii) Max α does not exist.

Definition 2.3: Rational Cuts: If r ∈ Q, then the set can be defined by α r = {x ∈ Q | x < r} is a cut. We call αr a
rational cut.
Definition 2.3: The negative cut: For any α ∈ R, the set defined by -α = {-s ∈ Q | s ∉ α, s ≠ min (Q\α)} is also a
cut. This is called the negative cut of α.

Definition 2.4: The Dedekind Real Number System: An ordered field (RD, ⊕, ⊙, ≻) is called a Dedekind
real number system if
(i) There exist a subfield (QD, ⊕, ⊙, ≻) which is isomorphic to (Q, +,>).
(ii) (RD, ⊕, ⊙, ≻) is order complete.

III. Properties Of Dedekind Real Number Systems

Most of the results proven here are usually regarded as properties fundamental to the real numbers.
Order completeness of Dedekind real number systems are invoked in most of the proofs, and this may be how
Dedekind stumble upon the idea that order completeness may just be the „essence‟ of the real numbers. It is now
started off with the Archimedean Property, which incidentally also employs the fact that Dedekind real number
systems are order complete.

3.1: Archimedean Property

Theorem 3.1.1: Let x, y ∈ R, x>0. Then there exist n ∈ N such that 𝑛𝑥 > 𝑦. Let A = {nx| n ∈ N}.By
contradiction assume, A is bounded above by y then Sup A exists (By order completeness). Since 𝑥 > 0,
sup 𝐴 − 𝑥 < 𝑠𝑢𝑝𝐴 and sup 𝐴 − 𝑥 is not an upper bound for A. Hence m ∈ N such that sup 𝐴 − 𝑥 < 𝑚𝑥 , that
is sup 𝐴 < 𝑚 + 1 𝑥 . But 𝑚 + 1 𝑥 ∈ A and this contradict the fact that Sup A is an upper bound for A. Since
the falsity of the claim leads to a contradiction, the claim must be true.

Lemma 3.1.2: Let x, y ∈ R. Then the following holds:

(i) There exist n ∈ N such that 𝑛 > 𝑦.
1
(ii) If 𝑥 > 0, there exist n ∈ N such that 𝑥 > .
𝑛
(iii) If 𝑦 ≥ 0, there exist n ∈ N such that 𝑛 − 1 ≤ 𝑦 < 𝑛.
Proof:
(i) This is a special case of Archimedean Property for 𝑥 = 1 > 0.
1
(ii) By Archimedean Property, for 𝑦 = 1, there exist n ∈ N such that 𝑛𝑥 > 1, that is 𝑥 > .
𝑛
(iii) Consider the set A = {m ∈ N | y<m}. (i) Ensures that A is not empty.
Hence, since N is well-ordered, Min A = n exist. Then n-1∉ A.

𝑹
Lemma 3.1.3: is not empty.
𝑸
Proof: Take Q ⊆ R, then Q = R. Hence, (Q, +,>) = (R, +,>) which is order complete.
𝑹 𝑹
But (Q, +,>) is not order complete. Hence, cannot be empty. We call elements in as irrational points.
𝑸 𝑸

𝑹
Theorem3.1.4: Density theorem: Let x, y ∈ R be such that 𝑥 < 𝑦. Then there exist r ∈ Q, z ∈ such that
𝑸
𝑥 < 𝑟, 𝑧 < 𝑦
1 1
Proof: Assume 𝑥 > 0 and if > 0 then < 𝑛 (by Archimedean property) there exist n ∈ N. Therefore
(𝑦−𝑥) (𝑦 −𝑥)
𝑛𝑦 − 𝑛𝑥 > 1. Since 𝑛𝑥 > 0, there exist m ∈ N such that 𝑚 − 1 ≤ 𝑛𝑥 + 1 < 𝑚 (by Archimedean property).
𝑥<𝑚 𝑚
Then 𝑚 ≤ 𝑛𝑥 + 1 < 𝑛𝑦 that is 𝑚 < 𝑛𝑥 < 𝑛𝑦. Hence and 𝑟 = ∈ Q. If 𝑥 = 0, then 𝑦 > 0 and by
𝑛<𝑦 𝑛
0<1
Archimedean property, there exist n ∈ N such that 𝑥 = and it is simply let r=1/n ∈ Q. If 𝑥 < 0 and𝑦 > 0,
𝑛 <𝑦
then simply let 𝑟 = 0 ∈ 𝑄. Finally, if 𝑥 < 𝑦 ≤ 0, then – 𝑥 > −𝑦 ≥ 0 and we have proved that there exist r′ ∈ Q
𝑹 𝑹
such that–x>r′ >-y, that is 𝑥 < −𝑟′ < 𝑦 and we let 𝑟 = −𝑟′ ∈ 𝑄. Since is not empty there exist z′ ∈ . By
𝑸 𝑸
above result there exist r ∈ Q such that 𝑥 + 𝑧′ < 𝑟 < 𝑦 + 𝑧′, that is 𝑥 < 𝑟 − 𝑧′ < 𝑦. we claim 𝑧 = 𝑟 − 𝑧 ′ ∉
𝑄. Suppose not. Then there exist r′ ∈ Q such that 𝑟 − 𝑧′ = 𝑟′, that is 𝑧′ = 𝑟 − 𝑟′ ∈ 𝑄, a contradiction. The
density theorem is proven.

DOI: 10.9790/5728-1305023240 www.iosrjournals.org 33 | Page

 Real Analysis of Real Numbers- Cantor and Dedekind Real Number Structuring

3.2. Cauchy Completeness of R

Theorem 3.2.1: Let {xn} in R. Then the following holds:
(i) If {xn} is monotonically increasing and bounded by M, then {𝑥𝑛 } is convergent.
Proof: Since | xn | ≤ M ∀ n ∈ N, the set A = {xn | n ∈ N} is bounded above by M, x1∈A so A ≠ φ. x = SupA
exists (by order completeness). Let any ε > 0 be given. Then there exist 𝑥𝑘 ∈ 𝐴 such that𝑥𝑘 > 𝑆𝑢𝑝𝐴 – 𝜀.Since
{𝑥𝑛 } is increasing, we have 𝑥𝑛 ≥ 𝑥𝑘 ∀ 𝑛 ≥ 𝑘. Hence ∀ 𝑛 ≥ 𝑘, Sup A – ε < xk ≤ xn ≤ Sup A < Sup A + ε, |
xn – SupA | < ε. Therefore 𝑥𝑛 → 𝑆𝑢𝑝𝐴.
(ii) If {xn} is monotonically decreasing and bounded by M, then {xn} is convergent.
Proof: Since| 𝑥𝑛 | ≤ 𝑀 ∀ 𝑛 ∈ 𝑁, the set A = {xn | n ∈ N} is bounded below by –M. Let x1∈A so A ≠ φ. Then
–A = {-xn | n ∈ N} ≠ φ is bounded above and since {-xn} is increasing, we have{−𝑥𝑛} → 𝑆𝑢𝑝 (−𝐴). 𝑖𝑛𝑓𝐴 =
−𝑆𝑢𝑝 (−𝐴). Since lim {-xn} exists, we have 𝑚 𝑥𝑛 = 𝑙𝑖𝑚 ((−1) {−𝑥𝑛 }) – 𝑙𝑖𝑚 {−𝑥𝑛 } – 𝑆𝑢𝑝 (−𝐴) −
(−𝑖𝑛𝑓𝐴) 𝑖𝑛𝑓𝐴 , that is {𝑥𝑛 } → 𝑖𝑛𝑓𝐴.
(iii) If {xn} is Cauchy and has a subsequence which converges to x, then we also have {x n} → x.
Proof: Let some subsequence (x n(k)) → x. Let any ε > 0 be given. Since {xn} is Cauchy there exist N ∈ N s.t
| 𝑥𝑛 – 𝑥𝑚 | < 𝜀/2 ∀ 𝑛, 𝑚 ≥ 𝑁
𝑆𝑖𝑛𝑐𝑒 (𝑥𝑛 (𝑘)) → 𝑥, , 𝑡𝑕𝑒𝑟𝑒 𝑒𝑥𝑖𝑠𝑡 𝑇 ∈ 𝑁 𝑠𝑢𝑐𝑕 𝑡𝑕𝑎𝑡
| 𝑥𝑛 𝑘) – 𝑥 | < 𝜀/2 ∀𝑘 ≥ 𝑇
If taken as S = max (N, T), note that 𝑛𝑆 ≥ 𝑛𝑁 ≥ 𝑁. Hence,
| 𝑥𝑛 – 𝑥 | = | (𝑥𝑛 – 𝑥𝑛(𝑆)) + (𝑥𝑛 (𝑆) – 𝑥) |
≤ 𝑥𝑛 – 𝑥𝑛 𝑆 + 𝑥𝑛 𝑆 – 𝑥
𝜀 𝜀
< + = 𝜀, 𝑛 ≥ 𝑁 Therefore {xn} → x.
2 2
(iv) If {xn} is bounded, ,there exist a subsequence which is convergent (Bolzano – Weierstrass property)
Proof: There exist a subsequence of {xn} which is monotone. Since {xn} is bounded, this subsequence is also
bounded. Hence, by (i) and (ii), we can conclude that this subsequence must be convergent.
(v) If {xn} is Cauchy, it is convergent (Cauchy Completeness).
Proof: Now {xn} has a convergent subsequence. Since {xn} is also Cauchy, we can conclude, that {xn} is also
convergent.
Result 3.2.2:
(i) {r | r ∈ Q, r < x + y} = {s + t | s, t ∈ Q, s < x, t < y.
(ii) {r | r ∈ Q, 0 < r < xy} = {st | s, t ∈ Q, 0 < s < x, 0 < t < y}, x, y > 0.
(iii) Sup{r | r ∈ Q, r < x} = Sup{r | r ∈ Q, 0 < r < x}, x > 0.

3.3 Dedekind Real Number Systems Are Unique

Consider any 2 Dedekind Real number system (𝑅, +, . , >) and 𝑅′ , +′ , .′ , >′ , by transitivity of isomorphism,
(𝑄, +, . , >) ≃ (𝑄′, +′, . ′, > ′) and we let 𝜑: 𝑄 → 𝑄′ be that isomorphism. Define the mapping 𝜓: 𝑅 → 𝑅′ by
𝜓(𝑥) = 𝑆𝑢𝑝 𝐴𝑥 ∀ 𝑥 ∈ 𝑅. where 𝐴𝑥 = {𝜑(𝑟) | 𝑟 < 𝑥, 𝑟 ∈ 𝑄}.
To prove: ψ is well-defined.
Proof: For any 𝑥 ∈ 𝑅, the Archimedean property for R demands that there exist 𝑛 ∈ 𝑁 (hence in Q) such that
𝑛 > 𝑥. Hence, 𝜑(𝑟) ∈ 𝐴𝑥 => 𝑟 < 𝑥 < 𝑛 => 𝜑(𝑟) < ′ 𝜑(𝑛) (by isomorphism). Hence, Ax is bounded
above (𝑏𝑦 𝜑(𝑛)) and so by order completeness of R′, Sup Ax exists, i.e. 𝜓(𝑥) ∈ 𝑅′. Also, for any 𝑥, 𝑦 ∈ 𝑅 such
that 𝑥 = 𝑦, we have 𝜓(𝑥) = 𝑆𝑢𝑝{𝜑(𝑟) | 𝑟 < 𝑥, 𝑟 ∈ 𝑄}
= 𝑆𝑢𝑝{𝜑(𝑟) | 𝑟 < 𝑥 = 𝑦, 𝑟 ∈ 𝑄} (note that suprema is unique)
= 𝜓(𝑦)
Hence ψ is well-defined.
Suppose now that there exist 𝑥, 𝑦 ∈ 𝑅 such that 𝜓(𝑥) = 𝜓(𝑦) but𝑥 ≠ 𝑦. Without loss of generality, it may
be assumed𝑥 < 𝑦. By density theorem for R there exist r1 ∈ Q such that𝑥 < 𝑟1 < 𝑦. Applying the density
theorem for R on r1 and y, we obtain some r2∈Q such that𝑥 < 𝑟1 < 𝑟2 < 𝑦. Now𝜑(𝑟) ∈ 𝐴𝑥 => 𝑟 < 𝑥 <
𝑟1 < 𝑟2 => 𝜑(𝑟) < ′ 𝜑(𝑟1) < ′ 𝜑(𝑟2 ) (by isomorphism). Hence, both φ (r1) and φ (r2) are upper bound for
Ax. Since𝜑(𝑟1) < ′ 𝜑(𝑟2 ), we cannot have 𝜑(𝑟2 ) = 𝑆𝑢𝑝 𝐴𝑥 and hence𝜓(𝑥) < ′ 𝜑(𝑟2 ). But𝜑 (𝑟2 ) ∈ 𝐴𝑦, so
𝜓(𝑥) < ′ 𝜑(𝑟2 ) ≤ ′ 𝜓(𝑦), contradicting𝜓(𝑥) = 𝜓(𝑦)
Hence ψ must be one-one.
Now take any 𝑥′ ∈ 𝑅′. Consider the element 𝑥 = 𝑆𝑢𝑝{𝑟 𝜖 𝑄 | 𝜑(𝑟) < 𝑥′}.
To claim: 𝑥 ∈ 𝑅.
By Archimedean property for R′, there exist 𝑛′ ∈ 𝑁′ (hence in Q′) such that 𝑛′ >′ 𝑥 ′ , 𝑎s φ is onto, there exist
𝑛 ∈ 𝑄 such that 𝑛′ = 𝜑(𝑛). Then 𝜑(𝑟) < ′ 𝑥′ => 𝜑(𝑟) < ′ 𝑥′ < ′ 𝜑(𝑛) => 𝑟 < 𝑛 (by isomorphism).
Hence n is an upper bound for {𝑟 ∈ 𝑄 | 𝜑(𝑟) < ′ 𝑥′} and so by order completeness of R, 𝑥 exists
To claim: 𝑆𝑢𝑝 𝐴𝑥 = 𝑥′.

DOI: 10.9790/5728-1305023240 www.iosrjournals.org 34 | Page

 Real Analysis of Real Numbers- Cantor and Dedekind Real Number Structuring

If x′ is not an upper bound for Ax, then there exist 𝜑(𝑟) ∈ 𝐴𝑥 such that 𝜑(𝑟) > ′ 𝑥′. Then 𝑟 ∉ {𝑟 ∈ 𝑄 | 𝜑(𝑟) <
′ 𝑥′}. Since 𝑟 < 𝑥, r is not an upper bound for {𝑟 ∈ 𝑄 | 𝜑(𝑟) < ′ 𝑥′} and so there exist 𝑟1 ∈ { 𝑟 ∈ 𝑄 | 𝜑(𝑟) <
′ 𝑥′} such that 𝑟 < 𝑟1 . By isomorphism, 𝜑(𝑟) < ′ 𝜑(𝑟1 ), i.e. 𝜑(𝑟) < ′ 𝜑(𝑟1) < ′ 𝑥′ and so 𝑟 ∈ {𝑟 ∈
𝑄 | 𝜑(𝑟) < ′ 𝑥′}, a contradiction! Hence x′ is an upper bound for Ax. Now take any 𝑦′ ∈ 𝑅′ such that 𝑦′ < ′ 𝑥′.
Then by density theorem for R′, there exist 𝑟1′ ∈ 𝑄′ such that 𝑦′ < ′ 𝑟1′ < ′ 𝑥′. Applying the density theorem
on r1′ and x′, we obtain r2′∈Q′ such that 𝑦′ < ′ 𝑟1′ < ′ 𝑟2′ < ′ 𝑥′. Since φ is onto there exist r1, r2∈Q such that
𝜑(𝑟1 ) = 𝑟1 ′, 𝜑(𝑟2 ) = 𝑟2 ′, i.e. 𝑟1 , 𝑟2 ∈ {𝑟 ∈ 𝑄 | 𝜑(𝑟) < ′ 𝑥′} and so 𝑟1 , 𝑟2 ≤ 𝑥. By isomorphism 𝜑(𝑟1 ) <
′ 𝜑(𝑟2 ) => 𝑟1 < 𝑟2 and so we have r1 < x. Then 𝑟1 ′ = 𝜑(𝑟1 ) ∈ 𝐴𝑥 and 𝑦′ < ′ 𝑟1 ′ < ′ 𝑥′. Hence 𝑥′ =
𝑆𝑢𝑝 𝐴𝑥, i.e. 𝜓(𝑥) = 𝑥′. Hence ψ is onto.
Take 𝑥, 𝑦 ∈ 𝑅. ∀ n > k.
(i) 𝜓(𝑥 + 𝑦) = 𝑆𝑢𝑝{𝜑(𝑟) | 𝑟 < 𝑥 + 𝑦, 𝑟 ∈ 𝑄}
= 𝑆𝑢𝑝{𝜑(𝑠 + 𝑡) | 𝑠 < 𝑥, 𝑡 < 𝑦, 𝑠, 𝑡 ∈ 𝑄}
= 𝑆𝑢𝑝{𝜑(𝑠) + ′ 𝜑(𝑡) | 𝑠 < 𝑥, 𝑡 < 𝑦, 𝑠, 𝑡 ∈ 𝑄} (by isomorphism)
= 𝑆𝑢𝑝{𝜑(𝑠) | 𝑠 < 𝑥, 𝑠 ∈ 𝑄} + ′ 𝑆𝑢𝑝{𝜑(𝑡) | 𝑡 < 𝑦, 𝑡 ∈ 𝑄}
= 𝜓(𝑥) + ′ 𝜓(𝑦).
(ii) Assume 𝑥, 𝑦 > 0. Then 𝛹 (𝑥. 𝑦) = 𝑆𝑢𝑝{𝜑(𝑟) | 𝑟 < 𝑥. 𝑦, 𝑟 ∈ 𝑄}
= 𝑆𝑢𝑝{𝜑(𝑟) | 0 < 𝑟 < 𝑥. 𝑦, 𝑟 ∈ 𝑄}
= 𝑆𝑢𝑝{𝜑(𝑠. 𝑡) | 0 < 𝑠 < 𝑥, 0 < 𝑡 < 𝑦, 𝑠, 𝑡 ∈ 𝑄}
= 𝑆𝑢𝑝{𝜑(𝑠) . ′ 𝜑(𝑡) | 0 < 𝑠 < 𝑥, 0 < 𝑡 < 𝑦, 𝑠, 𝑡 ∈ 𝑄} (by isomorphism)
= 𝑆𝑢𝑝{𝜑(𝑠) | 0 < 𝑠 < 𝑥, 𝑠 ∈ 𝑄} . ′ 𝑆𝑢𝑝{𝜑(𝑡) | 0 < 𝑡 < 𝑦, 𝑡 ∈ 𝑄}(
Note that by isomorphism, φ(s), φ(t)>′ 0′)
= 𝑆𝑢𝑝{𝜑(𝑠) | 𝑠 < 𝑥, 𝑠 ∈ 𝑄} ∗ ′ 𝑆𝑢𝑝{𝜑(𝑡) | 𝑡 < 𝑦, 𝑡 ∈ 𝑄}
= 𝜓(𝑥) . ′ 𝜓(𝑦).
To claim: One of 𝑥, 𝑦 is zero. Without loss of generality (due to commutative), it is assumed x = 0.
Then 𝜓 (𝑥. 𝑦) = 𝜓(0. 𝑦)
= 𝜓(0)
= 0′
= 0′ . ′ 𝜓(𝑦)
= 𝜓(𝑥) . ′ 𝜓(𝑦)
(b) 𝑥, 𝑦 < 0 𝜓 𝑥. 𝑦 = 𝜓 −𝑥 . −𝑦
= 𝜓 −𝑥 ∗′ 𝜓 −𝑦 (∵ −𝑥, −𝑦 > 0)
= −𝜓 𝑥 ∗ ′ (−𝜓(𝑦))
= 𝜓 𝑥 ∗ ′ 𝜓(𝑦)
(c) 𝑥 > 0, 𝑦 < 0
= −𝜓 𝑥. 𝑦 = 𝜓 − 𝑥. 𝑦 = 𝜓(𝑥. (−𝑦))
= 𝜓 𝑥 .′ 𝜓 −𝑦 ∵ 𝑥, −𝑦 > 0 = 𝜓(𝑥) . ′ (−𝜓(𝑦)) − 𝜓 𝑥 .′ 𝜓 𝑦 𝑖. 𝑒. 𝜓 𝑥. 𝑦
= 𝜓 𝑥 .′ 𝜓 𝑦

IV. Cantor Real Number System

Definition 4.1: Let C denote the set of all Cauchy sequences in Q. It is (rn ), (𝑠𝑛 ) ∈ 𝐶 are equivalent and we
write (rn ) ∼ (sn ) if given any rational𝜀 > 0, there exist 𝑘 ∈ 𝑁 such that | (rn ) – sn | < 𝜺.
Theorem 4.2: The relation ∼ is an equivalence relation on C.
Proof: Take any(rn ) ∈ 𝐶. Given any 𝜀 > 0, take 1 ∈ 𝑁.Then | (rn ), – (rn ), | = 0 < 𝜀 ∀ 𝑛 ≥ 1.Hence,
(rn ) ∼ (rn ). Hence ∼ is reflexive.
Take any (rn ), (sn ) ∈ 𝐶 such that (rn ), ∼ (sn ). Given any𝜀 > 0, there exist 𝑘 ∈ 𝑁 𝑠. 𝑡 | rn – sn | < 𝜀 ∀ 𝑛 ≥
𝑘 i.e. | sn – (rn ), | = |(rn ), – sn | < 𝜀 ∀ 𝑛 ≥ 𝑘. Hence, sn ∼ rn . hence, ∼ is symmetric.
Take any (rn ), (sn ), (t n ) ∈ 𝐶 such that (rn ) ∼ (sn ) and (sn ) ∼ (t n ). Then given any > 0 , there exist
𝑘1, 𝑘2 ∈ 𝑁s.t| sn – t n | < 𝜀/2 ∀ 𝑛 ≥ 𝑘1 | sn – t n | < 𝜀/2∀ 𝑛 ≥ 𝑘2.
Take 𝑘 = 𝑚𝑎𝑥(k1 , k 2 ). Then| rn – sn | < 𝜀/2, | sn – t n | < 𝜀/2 ∀ 𝑛 ≥ 𝑘. But
rn – t n = (rn ) – sn + sn – t n
≤ rn – sn + sn – t n
𝜀 𝜀
< + ∀𝑛 ≥ 𝑘
2 2
= 𝜀 ∀ 𝑛 ≥ 𝑘.
Hence (rn ) ∼ (t n ). Hence, ∼ is transitive. Hence, ∼ is an equivalence relation on C.

DOI: 10.9790/5728-1305023240 www.iosrjournals.org 35 | Page

 Real Analysis of Real Numbers- Cantor and Dedekind Real Number Structuring

Definition 4.3: Rational Convergence Point: Let (rn ) ∈ 𝐶 be such that(rn ) → 𝑟 ∈ 𝑄. Then (rn ) ∼ (sn ) only
if (sn ) → 𝑟. It is called [(rn )] a rational convergence point. In this case, it is denoted [(rn )] simply as [r].

Theorem 4.4: (rn ) ∼ (sn ).

𝜀
Proof: Suppose(rn ) ∼ (sn ). Let any ε > 0 be given. Then there exist 𝑘1 ∈ 𝑁 such that rn – sn < ∀𝑛 ≥
2
𝜀
𝑘1. Since rn → 𝑟, there exist 𝑘2 ∈ 𝑁 s.t rn – 𝑟 < ∀ 𝑛 ≥ 𝑘2.
2
Take 𝑘 = 𝑚𝑎𝑥(𝑘1, 𝑘2) and we have | rn – sn | < 𝜀/2, | rn – 𝑟 | < 𝜀/2, ∀ 𝑛 ≥ 𝑘
But | 𝑠𝑛 – 𝑟 | = | (sn – rn ) + (rn – 𝑟) |
≤ | sn – rn | + | rn – 𝑟 |
< 𝜀/2 + 𝜀/2 ∀ 𝑛 ≥ 𝑘
= 𝜀 ∀ 𝑛 ≥ 𝑘.
Hence, (sn ) → 𝑟.
Now suppose (sn ) → 𝑟. Note that this means(sn ) ∈ 𝐶. Then given 𝜀 > 0, there exist 𝑘1 ∈ 𝑁
s.t | sn – 𝑟 | < 𝜀/2 ∀ 𝑛 ≥ 𝑘1 Since (rn ) → 𝑟 there exist 𝑘2 ∈ 𝑁 s.t | rn – 𝑟 | < 𝜀/2 ∀ 𝑛 ≥ 𝑘2
Take 𝑘 = 𝑚𝑎𝑥(𝑘1, 𝑘2) and we have | rn – 𝑟 | < 𝜀/2, | sn – 𝑟 | < 𝜀/2 ∀ 𝑛 ≥ 𝑘
But |rn – 𝑠𝑛 | = | (rn – 𝑟) + (𝑟 – sn ) |
≤ | rn – 𝑟 | + | 𝑟 – sn |
𝜀 𝜀
< + ∀𝑛 ≥ 𝑘
2 2
= 𝜀 ∀ 𝑛 ≥ 𝑘.
Hence (𝑟𝑛) ∼ (𝑠𝑛).

Definition 4.5: Order on R: For any (rn ) ∈ 𝐶, we say that (rn) is a positive sequence if ,there exist some
rational 𝑟 > 0 and 𝑎 𝑘 ∈ 𝑁 such that rn > 𝑟 ∀ 𝑛 ≥ 𝑘.

Definition 4.6: The subset PR of R by 𝑃𝑅 = { [(rn )] ∈ 𝑅| (rn ) 𝑖𝑠 𝑎 𝑝𝑜𝑠𝑖𝑡𝑖𝑣𝑒 𝑠𝑒𝑞𝑢𝑒𝑛𝑐𝑒 }

Theorem 4.7: The set PR is a well-defined set.

Proof: Let[(rn )] ∈ 𝑃𝑅. It must be shown that if (sn ) ∼ (rn ), then [(sn )] ∈ 𝑃𝑅 also, that is (sn ) is a positive
sequence. Since (rn ) is a positive sequence there exist some rational 𝑟 > 0 and 𝑘1 ∈ 𝑁 such that rn > 𝑟 ∀ 𝑛 ≥
𝑘1 Also as (sn ) ∼ (rn ), for 𝜀 = 𝑟/2 > 0 there exist 𝑘2 ∈ 𝑁 s.t∀ 𝑛 ≥ 𝑘2 . Take𝑘 = 𝑚𝑎𝑥(k1 , k 2 ).
Then∀ 𝑛 ≥ 𝑘, have got| rn – sn | < 𝑟/2 , rn > 𝑟 => rn – sn < 𝑟/2 , rn > 𝑟 => 𝑟 − 𝑟/2 < rn – 𝑟/2 <
sn , that is 0 < 𝑟/2 < sn . Hence, (sn ) is also a positive sequence. Hence, PR is a well-defined set.

Theorem 4.8: For any [(rn )] ∈ 𝑅, one and only one of the following
holds: rn = 0 , rn ∈ 𝑃𝑅, − rn ∈ 𝑃𝑅. PR is closed under ⊕ and ⊙.
Proof:
Case 1:Take any [(rn )] ∈ 𝑅. It is first shown that one of the cases must hold. If [(rn )] ≠ [0], then it cannot
have(𝑟𝑛) → 0. Hence, there exist a rational 𝑟 > 0 and 𝑘1 ∈ 𝑁 such that|rn | ≥ 𝑟 ∀ 𝑛 ≥ 𝑘1 For 𝜀 = 𝑟/2 > 0 ,
there exist k2∈N s.t | rn – 𝑟𝑚 | < 𝑟/2 ∀ 𝑛, 𝑚 ≥ k 2 . Take𝑘 = 𝑚𝑎𝑥(k1 , k 2 ). Then it is | rn | ≥ 𝑟| rn – rm | <
𝑟
𝑟/2 ∀ 𝑛, 𝑚 ≥ 𝑘. In particular, have got 𝑟𝑘 ≥ 𝑟 rn – 𝑟𝑘 < ∀ 𝑛 ≥ 𝑘.Hence ∀ 𝑛 ≥ 𝑘, it is −(𝑟/2) <
2
rn – 𝑟𝑘 < 𝑟/2 rk – 𝑟/2 < rn < rk + 𝑟/2
Case 2: | rk | ≥ 𝑟
(a) 𝑟𝑘 ≥ 𝑟
Then 𝑟/2 = 𝑟 – 𝑟/2 ≤ 𝑟𝑘 – 𝑟/2 < 𝑟𝑛.This means rn > 𝑟/2 ∀ 𝑛 ≥ 𝑘 and so (rn ) is a positive sequence,
i.e. [(rn )] ∈ 𝑃𝑅.
(b) 𝑟𝑘 ≤ −𝑟
Thenrn < 𝑟𝑘 + 𝑟/2 ≤ −𝑟 + 𝑟/2 = −(𝑟/2).This means − rn > 𝑟/2 ∀ 𝑛 ≥ 𝑘 and so (−rn ) is a positive
sequence.
Hence−[(rn )] = [(rn )] ∈ 𝑃𝑅. If [(rn )] ∈ 𝑃𝑅, then, there exist 𝑟 > 0, 𝑘 ∈ 𝑁 such that rn > 𝑟 ∀ 𝑛 ≥ 𝑘.i.e.
− rn < −𝑟 < 0 ∀ 𝑛 ≥ 𝑘 and hence it is impossible that −[(rn )] = [(rn )] ∈ 𝑃𝑅.By symmetry, it can be
claimed that [(rn )] ∈ 𝑃𝑅 𝑎𝑛𝑑 − [(𝑟𝑛)] ∈ 𝑃𝑅 never be held together. If [(rn )] = [0], then(rn ) → 0. Hence for
any rational 𝑟 > 0, there exist 𝑘 ∈ 𝑁 such that| rn | < 𝑟 ∀ 𝑛 ≥ 𝑘 i.e.rn ≤ | rn | < 𝑟 ∀ 𝑛 ≥ 𝑘 and hence it
is impossible that [(rn )] ∈ 𝑃𝑅
Hence [(rn )] = [0] 𝑎𝑛𝑑 [(rn )] ∈ 𝑃𝑅 never be held together. Since [(rn )] = [0] ⇔ −[(rn )] = [0], it may
be claimed from preceding result that [(rn )] = [0] 𝑎𝑛𝑑 − [(rn )] ∈ 𝑃𝑅 never hold together also. Hence, only

DOI: 10.9790/5728-1305023240 www.iosrjournals.org 36 | Page

 Real Analysis of Real Numbers- Cantor and Dedekind Real Number Structuring

one of the cases is true. Take any [(rn )], [(sn )] ∈ 𝑃𝑅. Then there exist rational 𝑟, 𝑠 > 0 such that, there exist
𝑘 1 𝑘 2 ∈ 𝑁 where 𝑟𝑛 > 𝑟 > 0 ∀ 𝑛 ≥ k 2 sn ) > 𝑠 > 0 ∀ 𝑛 ≥ 𝑘 2
Take 𝑘 = 𝑚𝑎𝑥(𝑘 1 𝑘 2 ) and have got rn > 𝑟 > 0, sn > 𝑠 > 0 ∀ 𝑛 ≥ 𝑘 i.e. rn + sn > 𝑟 + 𝑠 >
0, rn sn > 𝑟𝑠 > 0 ∀ 𝑛 ≥ 𝑘.Hence, (rn + sn ) ), (rn sn ) ) are both positive sequence.
Then [(rn )] ⊕ [(sn )] = [(rn + sn )] ∈ 𝑃𝑅and [(rn )] ⊙ [(sn )] = [(rn sn )] ∈ 𝑃𝑅.
Hence, PR is closed under ⊕ and ⊙.

V. The Cantor Real Number System

Theorem 5.1: (RQ, ⊕, ⊙, ≻) is a subfield of (R, ⊕, ⊙, ≻). Furthermore, (RQ, ⊕, ⊙, ≻) ≃ (Q, +, ., >).
Take any [r], [s]∈ RQ. Then
[r] ⊕ (−[𝑠]) = [(𝑟)] ⊕ (−[(𝑠)])
= [(𝑟)] ⊕ [(−𝑠)]
= [(𝑟 – 𝑠)]
= [𝑟 – 𝑠] (∈ 𝑅𝑄)

[r] ⊙ ([𝑠]) − 1 = [(𝑟)] ⊙ ([(𝑠)]) − 1

= [(𝑟)] ⊙ [(1/𝑠)] ∀ [𝑠] ≠ [0], note this means s ≠ 0
= [(𝑟/𝑠)]
𝑟
= ∈ 𝑅𝑄
𝑠
[1] ∈ 𝑅𝑄. Hence (𝑅𝑄,⊕,⊙, ≻) is a subfield of (𝑅,⊕,⊙, ≻).
Consider the mapping 𝜑: 𝑅𝑄 → 𝑄 𝑔𝑖𝑣𝑒𝑛 𝑏𝑦 𝜑([𝑟]) = 𝑟 ∀ [𝑟] ∈ 𝑅𝑄.
[𝑟1 ] = [𝑟2 ] => 𝑙𝑖𝑚(𝑟1 ) = 𝑙𝑖𝑚(𝑟2 ) => 𝑟1 = 𝑟2 . Hence, φ is well-defined.
Let r1 = r2. Then 𝑙𝑖𝑚(𝑟1 ) = 𝑙𝑖𝑚(𝑟2 ) and so (𝑟1 ) ∼ (𝑟2 ), i.e., [𝑟1 ] = [𝑟2 ]. Hence, φ is one – one.
For any𝑟 ∈ 𝑄, take [𝑟] ∈ 𝑅𝑄 and we will have 𝜑([𝑟]) = 𝑟. Hence φ is onto.
Hence φ is bijective.
For any [𝑟], [𝑠] ∈ 𝑅𝑄, 𝜑([𝑟] ⊕ [𝑠]) = 𝜑( [(𝑟)] ⊕ [(𝑠)] )
= 𝜑( [(𝑟 + 𝑠)] )
= 𝜑([𝑟 + 𝑠])
= 𝑟 + 𝑠
= 𝜑([𝑟]) + 𝜑([𝑠])
𝜑([𝑟] ⊙ [𝑠]) = 𝜑( [(𝑟)] ⊙ [(𝑠)] )
= 𝜑( [(𝑟. 𝑠)] )
= 𝜑([𝑟. 𝑠])
= 𝑟. 𝑠
= 𝜑([𝑟]). 𝜑([𝑠])
[r] ≻ [𝑠] => [(𝑟)] ≻ [(𝑠)]
=> (𝑟 – 𝑠) is a positive sequence.
=> rn – sn > 𝑡 ∀ 𝑛 ≥ 𝑘 For some 𝑡 ∈ 𝑄, 𝑡 > 0, 𝑘 ∈ 𝑁.
=> 𝑟 – 𝑠 > 𝑡 Since rn = 𝑟, sn ) = 𝑠 ∀ 𝑛 ∈ 𝑁.
 𝑟 > 𝑡 + 𝑠, i.e. 𝑟 > 𝑠 since 𝑡 > 0.
Hence φ is an isomorphism from RQ to Q and so (RQ, ⊕, ⊙, ≻) ≃ (Q, +, ., >).

Theorem 5.2: Denseness of rational:

Let 𝑠𝑛 , t n ∈ 𝑅 be such that 𝑠𝑛 ≺ t n .Then there exist 𝑟 ∈ 𝑅𝑄 such that sn ) ≺ 𝑟 ≺ t n .
since 𝑠𝑛 ≺ t n , t n – 𝑠𝑛 is a positive sequence and so there exist a rational r′ > 0 and a k1∈N s.t
t n – sn > 𝑟 ′ ∀ 𝑛 ≥ k1 . 𝐴𝑠 𝑠𝑛 , t n are Cauchy, ,there exist k 2 , k 3 ∈ 𝑁 𝑠. 𝑡
| 𝑡𝑛 – 𝑡𝑚 | < 𝑟′/3 ∀ 𝑛, 𝑚 ≥ k 2 . | 𝑠𝑛 – 𝑠𝑚 | < 𝑟′/3 ∀ 𝑛, 𝑚 ≥ k 3 .
Take 𝑘 = 𝑚𝑎𝑥(k1 , k 2 , k 3 ) and have got, ∀ 𝑛 ≥ 𝑘, | t n – t k | < 𝑟′/3, | 𝑠𝑛 – sk | < 𝑟′/3, 𝑖. 𝑒. 𝑡𝑘 – 𝑟′/3 <
t n < t k + 𝑟′/3 t n – sn ) > 𝑟′, sk – 𝑟′/3 < sn ) < sk + 𝑟′/3, t n – sn ) > 𝑟′
Now,( t k + sk )/2 – 𝑠𝑛 = ½ (( t k – 𝑠𝑛 ) + ( sk – 𝑠𝑛 ))
> ½ (( t k – sk – 𝑟′/3) – 𝑟′/3)
𝑟′ 𝑟′ 𝑟′
> ½ 𝑟′ – – =
3 3 6
> t n – ( t k + sk )/2 = ½ (( t n – t k ) + ( t n – sk ))
> ½ (−𝑟′/3 + ( t k – 𝑟′/3 – sk ))
> ½ (– 𝑟′/3 + (– 𝑟′/3 + 𝑟′)) = 𝑟′/6

DOI: 10.9790/5728-1305023240 www.iosrjournals.org 37 | Page

 Real Analysis of Real Numbers- Cantor and Dedekind Real Number Structuring

Let 𝑟 = ( t k + sk )/2. Then both (𝑟 – 𝑠𝑛 ) 𝑎𝑛𝑑 ( t n – 𝑟) are positive sequence, i.e. [𝑟] ≻ [(𝑠𝑛 )] 𝑎𝑛𝑑 [( t n )] ≻
[𝑟]. Hence [(𝑠𝑛 )] ≺ [𝑟] ≺ [( t n )] for some [𝑟] ∈ 𝑅𝑄.

Theorem 5.3: Let 𝑟𝑛 , 𝑟 ∈ 𝑄. Then the following holds:

(i) | [𝑟] | = [| 𝑟 |]
(ii) ([rn ]) is Cauchy iff (rn ) is Cauchy.
(iii) ([rn ]) → [r] iff (rn ) → r
(iv) (rn )∈[r] iff ([rn ]) → [r]
(v) Let 𝛼 ∈ 𝑅 and (rn ) ∈ 𝛼. Then ([rn ]) is Cauchy and furthermore, ([rn ]) → 𝛼.
By isomorphism, 𝑟 ≻ 0 => 𝑟 > 0, 𝑖. 𝑒. 𝑟 = 𝑟; 𝑟 ≺ 0 => 𝑟 < 0, 𝑖. 𝑒. 𝑟 = −𝑟 𝑟 = 0 =>
𝑟 = 0, 𝑖. 𝑒. 𝑟 = 𝑟.Hence | [𝑟] | = [𝑟] = [| 𝑟 |] 𝑖𝑓 [𝑟] ≽ [0], −[𝑟] = [−𝑟] = [| 𝑟 |], 𝑖𝑓 [𝑟] ≺
[0], 𝑖. 𝑒. | [𝑟] | = [| 𝑟 |].
𝜀
Suppose(rn ) ∼ (𝑠𝑛 ). Let any ε > 0 be given. Then there exist k1 ∈ 𝑁 such that rn – sn < ∀ 𝑛 ≥ k1 .
2
𝜀
Since rn → 𝑟, ,there exist k 2 ∈ 𝑁 s.t rn – 𝑟 < ∀ 𝑛 ≥ k 2 .Take 𝑘 = 𝑚𝑎𝑥(k1 , k 2 ) and we have
2
| rn – 𝑠𝑛 | < 𝜀/2, | rn – 𝑟 | < 𝜀/2, ∀ 𝑛 ≥ 𝑘
But | sn – 𝑟 | = | (sn ) – rn ) + (rn – 𝑟) |
≤ | sn ) – rn | + | rn – 𝑟 |
< 𝜀/2 + 𝜀/2 ∀ 𝑛 ≥ 𝑘
= 𝜀 ∀ 𝑛 ≥ 𝑘. (sn ) → 𝑟.
Now suppose (sn ) ) → 𝑟. Note that this means(sn ) ) ∈ 𝐶. Then given 𝜀 > 0 there exist k1 ∈ 𝑁 s.t
| sn – 𝑟 | < 𝜀/2 ∀ 𝑛 ≥ k 2 Since (rn ) → 𝑟 there exist k 2 ∈ 𝑁 s.t | rn – 𝑟 | < 𝜀/2 ∀ 𝑛 ≥ k 2 Take 𝑘 =
𝑚𝑎𝑥(k1 , k 2 ) and we have | rn – 𝑟 | < 𝜀/2, | sn ) – 𝑟 | < 𝜀/2 ∀ 𝑛 ≥ 𝑘
But | rn – 𝑠𝑛 | = | (rn – 𝑟) + (𝑟 – sn ) ) |
≤ | rn – 𝑟 | + | 𝑟 – sn ) |
𝜀 𝜀
< + ∀𝑛 ≥ 𝑘
2 2
= 𝜀 ∀ 𝑛 ≥ 𝑘. Hence rn ∼ sn .

Theorem 5.4: (𝑅,⊕,⊙, ≻) is Cauchy Complete.

Proof: Let (𝛼𝑛 ) be any Cauchy sequence in R. Now, by denseness of rational, for each𝑛 ∈ 𝑁 , there exist
[rn ] ∈ 𝑅𝑄 s.t
𝛼𝑛 ⊕ (−[1/𝑛]) ≺ [rn ] ≺ 𝛼𝑛 ⊕ [1/𝑛], i.e. | [rn ] ⊕ (−𝛼𝑛 ) | ≺ [1/𝑛].
Let any real 𝜺 ≻ [0] be given. Since (𝛼𝑛 ) is Cauchy there exist 𝑘1 ∈ 𝑁 such that | 𝛼𝑛 ⊕ (−𝛼𝑚 ) | ≺
𝜺/3 ∀ 𝑛 ≥ 𝑘. By denseness of rational, there exist [𝜀] ∈ 𝑅𝑄 such that[0] ≺ [𝜀] ≺ 𝜺/3. By Archimedean
property for Q there exist k 2 ∈ 𝑁 s.t 1/k 2 < 𝜀, hence 1/𝑛 < 𝜀 ∀ 𝑛 ≥ k 2 . By isomorphism, It can then be
claimed [1/𝑛] ≺ [𝜀] ≺ 𝜺/3 ∀ 𝑛 ≥ k 2 .
Take 𝑘 = 𝑚𝑎𝑥(k1 , k 2 ). Then
rn ⊕ − r = rn ⊕ −𝛼𝑛 ⊕ 𝛼𝑛 ⊕ −𝛼𝑚 ⊕ 𝛼𝑚 ⊕ − rm
≼ | [rn ] ⊕ (−𝛼𝑛 )| ⊕ | 𝛼𝑛 ⊕ (−𝛼𝑚 ) | ⊕ | 𝛼𝑚 ⊕ (−[𝑟𝑚]) |
≺ [1/𝑛] ⊕ | 𝛼𝑛 ⊕ (−𝛼𝑚 ) | ⊕ [1/𝑚]
≺ 𝜺/3 ⊕ 𝜺/3 ⊕ 𝜺/3 ∀ 𝑛, 𝑚 ≥ 𝑘
= 𝜺 ∀ 𝑛, 𝑚 ≥ 𝑘
Hence ([rn ]) and (rn ) is a Cauchy sequence in Q. 𝛼 = [(rn )] will be in R, therefore ([rn ]) → 𝛼.
Now let any real 𝜺 ≻ [0] be given. There exist k1 ∈ 𝑁 s.t | [rn ] ⊕ (−𝛼) | ≺ 𝜺/2 ∀ 𝑛 ≥ k1 .There exist
k 2 ∈ 𝑁 s.t
[1/𝑛] ≺ 𝜺/2 ∀ 𝑛 ≥ k 2 . Take 𝑘 = 𝑚𝑎𝑥(k1 , k 2 ) and it is
𝛼𝑛 ⊕ −𝛼 = 𝛼𝑛 ⊕ − rn ⊕ rn ⊕ −𝛼
≼ 𝛼𝑛 ⊕ − rn ⊕ rn ⊕ −𝛼
1
≺ ⊕ rn ⊕ −𝛼
𝑛
𝜺 𝜺
≺ ⊕ ∀𝑛 ≥ 𝑘
2 2
= 𝜺 ∀ 𝑛 ≥ 𝑘. Hence(𝛼𝑛 ) → 𝛼.
(𝑅,⊕,⊙, ≻) is Cauchy Complete.

Theorem 5.5: Archimedean Property

For every[(rn )], [(sn )] ∈R such that[(rn )] ≻ [0] there exist [𝑛] ∈ 𝑅𝑁 such
that[𝑛] ⊙ [(rn )] ≻ [(sn )].If[(rn )] ≻ [(sn )], then since [1]∈RN, there is nothing to prove. Hence, it can be

DOI: 10.9790/5728-1305023240 www.iosrjournals.org 38 | Page

 Real Analysis of Real Numbers- Cantor and Dedekind Real Number Structuring

assumed that[(sn )] ≽ [(𝑟𝑛)] ≻ [0]. By denseness of rational there exist [r], [𝑠] ∈ 𝑅𝑄 such that[(sn )] ⊕
[1] ≻ [𝑠] ≻ [(sn )] ≽ [(rn )] ≻ [𝑟] ≻ [0].Under isomorphism, it is invoked that the Archimedean property
of Q so ,there exist [n]∈RN such that[n] ⊙ [r] ≻ [s]. It is then[𝑛] ⊙ [(𝑟𝑛)] ≻ [𝑛] ⊙ [𝑟] ≻ [𝑠] ≻ [(sn )]
𝑖. 𝑒. [𝑛] ⊙ [(rn )] ≻ [(sn )]. Hence, the Archimedean property holds.
Definition 5.6: For any ordered field (𝑅𝐶,⊕,⊙, ≻), we say that it is a Cantor Real Number System if
(i) there exist a subfield (𝑄𝐶,⊕,⊙, ≻) that is isomorphic to (𝑄, +, . , >).
(ii) (𝑅𝐶,⊕,⊙, ≻) is Cauchy Complete
(iii) (RC, ⊕, ⊙, ≻) has the Archimedean property.
Hence it is Cantor Real Number System and so Cantor Real Number System does exist.

Theorem 5.7: (ℝ, +, . , >) is order complete.

Take any non-empty subset of A of ℝ that is bounded above by some u0.Let U = {𝑢 ∈ ℝ | u is an upper bound
of A}. Since
𝐴 ≠ 𝜑, ∃𝑎 0 ∈ 𝐴. ∃ 𝑚 ∈ ℕ 𝑠. 𝑡 𝑚 > −𝑎 0 𝑖. 𝑒. 𝑎0 < −𝑚 𝑠𝑜 − 𝑚 ∉ 𝑈. (By Archimedean property) it is
defined 2 sequences (xn ), (yn) as such: 𝑥1 = −𝑚 (∉ 𝑈) 𝑦1 = 𝑢0 (∈ 𝑈). Suppose thatxn ∉ 𝑈 𝑎𝑛𝑑 𝑦𝑛 ∈ 𝑈.
Define xn+1 = ½(xn + yn) if ½(xn + yn) ∉U xn otherwise 𝑦𝑛 + 1 = ½(xn + 𝑦𝑛 )if ½ xn + 𝑦𝑛 ∈ 𝑈 𝑦𝑛 otherwise.
By definition, note that then will be xn ∉ 𝑈, 𝑦𝑛 ∈ 𝑈 ∀ 𝑛 ∈ 𝑁. Hence, for every n, there exist 𝑎𝑛 ∈ 𝐴 such
that𝑥𝑛 < 𝑎𝑛 ≤ 𝑦𝑛 ), i.e. xn < 𝑦𝑛 . Let 𝑁 = 𝑦1 – 𝑥1 > 0. First a few observations shall be made. If
½ xn + 𝑦𝑛 ∈ 𝑆, then 𝑦𝑛 ) + 1 – xn + 1 = ½ xn + 𝑦𝑛 ) – xn = ½ 𝑦𝑛 ) – xn . If½ xn + 𝑦𝑛 ∉ 𝑆, then
𝑦𝑛 + 1 – xn + 1 = 𝑦𝑛 – ½ xn + 𝑦𝑛 ) = ½ 𝑦𝑛 – xn . Hence 𝑦𝑛 + 1 – xn + 1 = ½ xn + 𝑦𝑛 ∀ 𝑛 ∈
𝑁.Hence𝑦𝑛 – xn = ½ xn − 1 + 𝑦𝑛 ) − 1 = ½(½(xn − 2 + 𝑦𝑛 ) − 2)) = 𝑁/2𝑛 − 1 ∀ 𝑛 ∈ 𝑁.
For every 𝑛 ∈ 𝑁, either𝑦𝑛 = 𝑦𝑛 + 1 𝑜𝑟 𝑦𝑛 – 𝑦𝑛 + 1 = 𝑦𝑛 – ½(xn + 𝑦𝑛 )
= ½(𝑦𝑛 – xn )
= N/2n
> 0 i.e. 𝑦𝑛 + 1 ≤ 𝑦𝑛 and so (yn ) is decreasing.
For every 𝑛 ∈ 𝑁, either xn = xn + 1 𝑜𝑟xn + 1 – xn = ½(xn + 𝑦𝑛 ) – xn
= ½(𝑦𝑛 – xn )
= 𝑁/2𝑛
> 0 i.e.xn + 1 ≥ xn and so {xn } is increasing. If m, n ∈ N are such that𝑚 < 𝑛, have got 0 <
𝑦𝑚 – 𝑦𝑛 < 𝑦𝑚 – xn (∵ 𝑦𝑛 xn < 𝑦𝑚 – 𝑥𝑚 (∵ xn > 𝑥𝑚) = 𝑁/2𝑚 − 1 i.e. | 𝑦𝑚 – 𝑦𝑛 | < 𝑁/2𝑚 − 1. Let
any 𝜀 > 0 be given by Archimedean property, there exist 𝑘′ ∈ 𝑁 such that𝑁 < 𝑘′𝜀.By the exponentiation
version of Archimedean property for N , there exist 𝑘 ∈ 𝑁 such that2𝑘 > 𝑘′.Hence, ∀ 𝑛 ≥ 𝑘, have got
2𝑛 ≥ 2𝑘 > 𝑘′ i.e. 𝜀2𝑛 ≥ 𝜀2𝑘 > 𝜀𝑘′ > 𝑁, 𝑖. 𝑒. 𝑁/2𝑛 < 𝜀 Hence | 𝑦𝑚 – 𝑦𝑛 | < 𝑁/2𝑚 − 1 ( by
symmetry of absolute order, it can always be assumed 𝑚 < 𝑛. m = n, | 𝑦𝑚 – 𝑦𝑛 | = 0 < 𝑁/2𝑚 − 1) <
𝜀∀ 𝑛, 𝑚 ≥ 𝑘 + 1. Hence, it is shown that (yn ) is Cauchy. By Cauchy Completeness, (𝑦𝑛) converges to some
𝑦 ∈ 𝑅. Suppose 𝑦 ∉ 𝑈. Then there exist 𝑎 ∈ 𝐴 such that 𝑎 > 𝑦. By density theorem there exist 𝑧 ∈ 𝑅 such
that 𝑎 – 𝑦 > 𝑧 > 0. Since it is always been𝑦𝑛 ≥ 𝑎, it will be 𝑦𝑛 – 𝑦 ≥ 𝑎 – 𝑦 > 𝑧 > 0. But(𝑦𝑛 ) → 𝑦, i.e.
for there exist 𝑘 ∈ 𝑁 s.t | 𝑦𝑛 – 𝑦 | < 𝑧/2∀ 𝑛 ≥ 𝑘. In particular,𝑦𝑘 – 𝑦 ≤ | 𝑦𝑛 – 𝑦 | < 𝑧/2 < 𝑧 < 𝑦𝑘 – 𝑦,
a contradiction! Hence, 𝑦 ∈ 𝑈.Now suppose there exist 𝑢 ∈ 𝑈 such that 𝑢 < 𝑦. Note first that similar to above,
by Archimedean property and its exponentiation version for N, given any ε > 0 there exist 𝑘 ∈ 𝑁 such that
𝑁
𝑁/2𝑛 < 𝜀 ∀ 𝑛 ≥ i.e. 𝑦𝑛 – xn – 0 = 𝑦𝑛 – 𝑥𝑛 ∵ 𝑦𝑛 – xn > 0 = − 1 < 𝜀∀ 𝑛 𝑘 + 1 i.e. (𝑦𝑛 – xn ) →
2𝑛
0. If there exist 𝑦𝑘 such that 𝑦𝑘 < 𝑦, then it is 𝑦𝑛 ≤ 𝑦𝑘 < 𝑦 ∀ 𝑛 ≥ 𝑘 as (𝑦𝑛 ) is decreasing. But (𝑦𝑛 ) → 𝑦 so
for 𝜀 = 𝑦 – 𝑦𝑘 > 0 there exist N ∈ N such that | 𝑦𝑛 – 𝑦 | < 𝑦 – 𝑦𝑘 ∀ 𝑛 ≥ 𝑁.For𝑁′ = 𝑚𝑎𝑥(𝑘, 𝑁), have got
in particular | 𝑦𝑁, – 𝑦 | = −(𝑦𝑁, – 𝑦), | 𝑦𝑁, – 𝑦 | < 𝑦 – 𝑦𝑘 => −(𝑦𝑁, – 𝑦) < 𝑦 – 𝑦𝑘 => 𝑦𝑁, > 𝑦𝑘 ,
contradicting (yn) being decreasing. Hence, we always have 𝑦𝑛 – 𝑦 ≥ 0. since (𝑦𝑛 – 𝑥𝑛) → 0, for ε = y – u >
0 there exist 𝑘 ∈ 𝑁 such that 𝑦𝑛 – xn = 𝑦𝑛 – xn < 𝑦 – 𝑢, 𝑖. 𝑒. 𝑦𝑛 – 𝑦 < 𝑥𝑛 – 𝑢 ∀ 𝑛 ≥ 𝑘 .In
particular0 ≤ 𝑦𝑘 – 𝑦 < 𝑥𝑘 – 𝑢, 𝑖. 𝑒. 𝑥𝑘 > 𝑢. But this means 𝑥𝑘 > 𝑢 ≥ 𝑎 ∀ 𝑎 ∈ 𝐴, making 𝑥𝑘 an upper
bound, which is a contradiction! Hence, 𝑦 ≤ 𝑢 ∀ 𝑢 ∈ 𝑈 and so 𝑆𝑢𝑝 𝐴 = 𝑦 exists. Since (𝑅, +, . , >) has the
least upper bound property, it is hence order complete.

VI. Conclusion
Peano‟s axiom allows creating natural number system, integer system, field of rational, and the real
number system. The existence of the real number system was proceeded to pursue the instinct of Richard
Dedekind and Georg Cantor. These two approaches congregate distinctive number system. Dedekind built a
system that is not only reliable with normal operation but also free from the predicament of the immeasurable
DOI: 10.9790/5728-1305023240 www.iosrjournals.org 39 | Page
 Real Analysis of Real Numbers- Cantor and Dedekind Real Number Structuring

triangle from the rational numbers. Cantor with his natural approach resulted with the same system that has
given more confidence on this observation. The real number system is a complete ordered field that is therefore
Dedekind real number system.
This paper enunciates two things about the system, the real number system is distinctive and any
complete ordered field is the real number system. An ordered field is Cauchy complete and possesses the
Archimedean property only if the order is complete. Through this paper, a better insight on how the number
system actually works. The relationship between the fundamental subsets such as the integer and the natural
numbers could be seen well.

Acknowledgements
I acknowledge Dr. Balasubramanian Sathyamurthy, Professor, Department of Biochemistry, Ramaiah College
of Arts, Science and Commerce, Bangalore for his intensive help for pursuing and publishing this research
article

Reference
[1]. Richard Courant and Herbert Robbins, What is Mathematics., Oxford University Press, (1941)
[2]. Ivan Niven, Numbers: Rational and Irrational., New Mathematical Library, The Mathematical Association of America, (1961)
[3]. C D Olds, Continued Fractions:, New Mathematical Library, The Mathematical Association of America, (1963)
[4]. Elaine J. Hom, “Real Numbers: Properties and Definition.”. Livescience.com
[5]. Michael Ian Shamos, Shamos‟s Catalog of the Real Numbers., (2011)
[6]. John Stillwell, the Real Numbers: An Introduction to Set Theory and Analysis. Springer.
[7]. Anthony W. Knapp, Basic Real Analysis. Birkhäuser (2016).
[8]. B. Lafferriere, G. Lafferriere, N. Mau Nam Introduction to Mathematical Analysis. Portland State University Library (2015).
[9]. Shanti Narayan, Scand And Company, A Course Of Mathematical Analysis. (1962)
[10]. Lee Larson, Introduction to Real Analysis, University of Louisville (2014).
[11]. Robert Rogers, How We Got From There to Here: A Story of Real Analysis. Eugene Boman, Open SUNY Textbooks (2013).
[12]. Bruce K. Driver, Undergraduate Analysis Tools, University of California, San Diego (2013).
[13]. Terrence Tao, An Introduction to Measure Theory, American Mathematical Society (2011).
[14]. Martin Smith-Martinez, et al, Real Analysis. Wikibooks (2013).

Ahlam Khalaf Jabbar. “Real Analysis of Real Numbers- Cantor and Dedekind Real Number Structuring.”
IOSR Journal of Mathematics (IOSR-JM), vol. 13, no. 5, 2017, pp. 32–40.

DOI: 10.9790/5728-1305023240 www.iosrjournals.org 40 | Page