The Construction and Analysis of Real Numbers
The real numbers constitute the “continuum,” a mathematical structure and a complete ordered field that underlies the vast majority of analysis, calculus, and physics. While the rational numbers provide a dense set of ratios, they are fundamentally “holey”—failing to contain the limits of many convergent sequences (such as the sequence defining ). This lesson explores the rigorous formalizations used to “fill the gaps” in to arrive at .
The Supremum Property (Completeness Axiom)
The defining characteristic of the real numbers that distinguishes them from the rationals is the Least Upper Bound Property (or Supremum Property).
- Definition: A set is bounded above if there exists such that for all .
- Axiom: Every non-empty set of real numbers that is bounded above has a least upper bound (supremum) in .
Consider the set . In , this set is bounded above by or , but it has no least upper bound because . In , .
Construction 1: Dedekind Cuts
Richard Dedekind (1872) proposed a construction where real numbers are defined as partitions of the rational numbers.
Definition of a Dedekind Cut
A Dedekind Cut is a subset of satisfying:
- Non-triviality: and .
- Downward Closure: If and , where , then .
- No Maximum: For every , there exists such that .
The set of all such cuts defines . We identify each rational with the cut . Irrational numbers are cuts that do not correspond to any rational. For instance, is defined by the cut . Addition and multiplication are defined set-theoretically on these cuts.
Construction 2: Cauchy Sequences
Georg Cantor and Charles Méray independently formulated using the concept of completion.
Cauchy Sequences in
Recall that a sequence is Cauchy if for every , there exists such that for all : In , some Cauchy sequences do not converge to a rational number.
Equivalence Classes of Sequences
We define as the set of all Cauchy sequences of rational numbers, under an equivalence relation . Two sequences and are equivalent if: A real number is thus an equivalence class of Cauchy sequences. This formalizes the idea that a real number is “anything that a sequence of rationals can converge to.”
Algebraic and Topological Properties
An Ordered Field
is a field under addition and multiplication, satisfying the standard field axioms (associativity, commutativity, inverses). It is an ordered field because there is a total ordering consistent with the field operations:
- If , then .
- If and , then .
The Archimedean Property
For any , there exists an integer such that . This implies that the set of integers is not bounded above in , and that for any , there exists such that .
Density
Both (rational numbers) and (irrational numbers) are dense in . This means that between any two distinct real numbers , there exists a rational and an irrational such that and .
Topological Structure: The Metric Space
We view as a metric space with the distance function .
- Open Sets: A subset is open if for every , there exists such that .
- Compactness: A subset is compact if and only if it is closed and bounded (Heine-Borel Theorem).
- Connectedness: is connected, meaning it cannot be partitioned into two disjoint non-empty open sets. This reflects the “no gaps” nature of the continuum.
Cardinality: The Uncountability of R
While the set of rationals is countable (), the set of real numbers is uncountable. By Cantor’s Diagonal Argument, it can be shown that there is no bijection between and . The cardinality of is denoted by .
Computational Considerations: The Gap Between Model and Machine
In pure mathematics, consists of infinite precision numbers. In computer systems, we approximate using Floating-Point Representation (IEEE 754).
# Demonstrating the limitations of machine precision in representing Real Numbers
a = 0.1
b = 0.2
print(f"Mathematical 0.1 + 0.2 = 0.3")
print(f"Machine Output: {a + b}")
# Output: 0.30000000000000004
This discrepancy arises because and , while rational, have infinite repeating representations in binary, and must be truncated. Understanding the topology of (specifically error propagation and limits) is critical for numerical analysis and scientific computing.