Start Learning Reals 2 | Completeness Axiom [dark version]

Q: Why does the triangle inequality matter for connecting convergence and Cauchy sequences?

The triangle inequality provides a bound on the distance between two sequence terms via a third point. If a sequence converges to a, then for large n and m, both |x_n − a| and |x_m − a| are small. The triangle inequality gives |x_n − x_m| = |(x_n − a) − (x_m − a)| ≤ |x_n − a| + |x_m − a|. That turns “each term is close to a” into “each pair of late terms is close to each other,” which is exactly the Cauchy condition.

Q: How does the ε′ = ε/2 trick work in the proof that convergence implies Cauchy?

Given any ε > 0, the proof defines ε′ so that 2·ε′ = ε, i.e., ε′ = ε/2. Convergence guarantees an index N such that for all n > N, |x_n − a| N, both |x_n − a| and |x_m − a| are less than ε′, so |x_n − x_m| ≤ |x_n − a| + |x_m − a| < ε′ + ε′ = ε. This matches the Cauchy definition for the chosen ε.

Q: What exactly is the difference between a Cauchy (Kösi) sequence and a convergent sequence?

A convergent sequence has a specified target a: eventually every term lies within ε of a. A Cauchy sequence does not require naming any limit; it only requires that eventually all terms are mutually close: for every ε > 0, there exists N such that for all n, m > N, the distance |x_n − x_m| < ε. In pictures, convergence means terms enter an ε-neighborhood around a point, while Cauchy means terms enter a shrinking cluster even without identifying its center.

Q: Why does “Cauchy ⇒ convergent” fail in the rationals ℚ but hold in the reals ℝ?

In ℚ, there are Cauchy sequences whose terms get arbitrarily close to each other but approach a limit that is not rational. Since ℚ lacks those missing limit points, the sequence cannot converge within ℚ. ℝ fixes this by adding the completeness axiom: every Cauchy sequence converges to a real number, so the “holes” on the number line disappear.

Q: What does completeness mean in the axiomatic description of ℝ?

Completeness is stated as: every Kösi (Cauchy) sequence is also a convergence sequence, with distance measured by the absolute value. In other words, the axioms don’t just guarantee algebraic and order properties; they also guarantee that the metric notion of “eventually close to each other” always produces an actual limit in ℝ.

TL;DR

Absolute value provides the distance notion used to define both convergence and the Cauchy (Kösi) property.

Briefing Cornell Notes

Briefing

The real numbers are built around a single decisive idea: every Cauchy (Kösi) sequence must settle down to a limit. That “completeness axiom” is what distinguishes ℝ from the rationals ℚ, where Cauchy sequences can converge only to numbers that aren’t rational. The discussion starts with the absolute value as the distance tool, then uses the triangle inequality to connect two ways of describing “numbers getting arbitrarily close.”

Absolute value is treated as a distance measure on rational numbers, written with bars around a rational. Two key properties drive the later arguments. First, absolute value is multiplicative: |xy| = |x||y|, meaning multiplication behaves cleanly under distance. Second, absolute value satisfies the triangle inequality: |x + y| ≤ |x| + |y|. That inequality is crucial because it provides an upper bound on how far two points can be apart using intermediate steps—exactly what’s needed to relate “closeness to a limit” with “closeness to each other.”

Next comes the comparison between two sequence notions. A sequence is called a Kösi (Cauchy) sequence if, for every ε > 0, there is an index N such that all later terms are within ε of each other. Convergence is defined differently: a sequence converges to a number a if, for every ε > 0, there is an index N such that all later terms lie within ε of a. Visually, convergence means eventually all terms fall inside an ε-neighborhood around a point on the number line; Cauchy means eventually all terms fall inside a shrinking “cluster” even without naming a target point.

The relationship between these definitions is asymmetric. Convergence always implies the Cauchy property in any ordered number system where the triangle inequality holds. The proof uses a standard ε-splitting trick: given ε, choose ε′ = ε/2. Once terms are within ε′ of the limit a, the triangle inequality shows that the distance between any two sufficiently late terms is at most |x_n − a| + |x_m − a| < ε′ + ε′ = ε. That establishes “convergent ⇒ Cauchy.”

The reverse direction—“Cauchy ⇒ convergent”—is where ℝ earns its special status. Over the rationals ℚ, a Cauchy sequence need not converge to a rational number, so the implication fails. Over the reals ℝ, it is enforced as an axiom: every Cauchy sequence converges, with distance measured by the absolute value. This axiom is presented as the final ingredient that turns a field with an order into a complete ordered field, filling the “holes” on the number line.

The transcript then lists the axioms of the real numbers in four groups: addition forms an abelian group (associativity, inverses, commutativity), multiplication forms a group on nonzero elements, distributivity links addition and multiplication, and the order is a total order compatible with operations. Finally comes the Archimedean property and, most importantly, completeness: every Kösi sequence converges. The payoff is that ℝ behaves like a full number line with no gaps, enabling later calculations and, eventually, an explicit construction of the real numbers.

Cornell Notes

The real numbers ℝ are characterized by completeness: every Cauchy (Kösi) sequence converges. Using absolute value as a distance, convergence means terms eventually lie within ε of a limit a, while the Cauchy property means terms eventually lie within ε of each other. Convergence always implies the Cauchy property, and the proof relies on the triangle inequality plus an ε-splitting step (choose ε′ = ε/2). The reverse implication fails in the rationals ℚ, so ℝ is defined/axiomatized so that Cauchy sequences always have limits. This completeness axiom is what removes “holes” from the number line and makes ℝ a complete ordered field.

Why does the triangle inequality matter for connecting convergence and Cauchy sequences?

The triangle inequality provides a bound on the distance between two sequence terms via a third point. If a sequence converges to a, then for large n and m, both |x_n − a| and |x_m − a| are small. The triangle inequality gives |x_n − x_m| = |(x_n − a) − (x_m − a)| ≤ |x_n − a| + |x_m − a|. That turns “each term is close to a” into “each pair of late terms is close to each other,” which is exactly the Cauchy condition.

How does the ε′ = ε/2 trick work in the proof that convergence implies Cauchy?

Given any ε > 0, the proof defines ε′ so that 2·ε′ = ε, i.e., ε′ = ε/2. Convergence guarantees an index N such that for all n > N, |x_n − a| < ε′. Then for any n, m > N, both |x_n − a| and |x_m − a| are less than ε′, so |x_n − x_m| ≤ |x_n − a| + |x_m − a| < ε′ + ε′ = ε. This matches the Cauchy definition for the chosen ε.

What exactly is the difference between a Cauchy (Kösi) sequence and a convergent sequence?

A convergent sequence has a specified target a: eventually every term lies within ε of a. A Cauchy sequence does not require naming any limit; it only requires that eventually all terms are mutually close: for every ε > 0, there exists N such that for all n, m > N, the distance |x_n − x_m| < ε. In pictures, convergence means terms enter an ε-neighborhood around a point, while Cauchy means terms enter a shrinking cluster even without identifying its center.

Why does “Cauchy ⇒ convergent” fail in the rationals ℚ but hold in the reals ℝ?

In ℚ, there are Cauchy sequences whose terms get arbitrarily close to each other but approach a limit that is not rational. Since ℚ lacks those missing limit points, the sequence cannot converge within ℚ. ℝ fixes this by adding the completeness axiom: every Cauchy sequence converges to a real number, so the “holes” on the number line disappear.

What does completeness mean in the axiomatic description of ℝ?