The Timpson Lyceum seal

The Timpson Lyceum · Real Analysis (2020)

Lecture 7 (Real Analysis 2020) Part 1

Overview

This lecture connects sequences to the sets that contain them. We first prove that every convergent sequence is bounded, then develop a powerful sequential characterization of dense sets: a set is dense in R\mathbb{R} exactly when every real number is the limit of some sequence drawn from the set. We establish an order-preservation result — a convergent sequence of non-negative terms has a non-negative limit — and use it to show that closed intervals contain the limits of their sequences. This motivates the general definition of a closed set as one that contains all of its sequential limits, and we test that definition against several examples. These ideas are foundational for the topology of R\mathbb{R} studied later.

Key Concepts

Bounded sequence. A sequence (an)(a_n) is bounded if there exists a real number MM such that ∣an∣≤Mfor all n,|a_n| \le M \quad \text{for all } n, equivalently −M≤an≤M-M \le a_n \le M for all nn. The value MM "sandwiches" the entire sequence.

Convergent ⇒\Rightarrow bounded. If (an)(a_n) converges, then (an)(a_n) is bounded. The proof splits the sequence into a tail (controlled by convergence) and a finite head (controlled by taking a maximum):

A set containing a sequence. A set SS contains the sequence (an)(a_n) if an∈Sa_n \in S for every nn — equivalently, the range (image) of the sequence is a subset of SS. A sequence is a function on N\mathbb{N}, not itself a set, so "contains" refers to its range.

Dense set — sequential characterization. Recall SS is dense in R\mathbb{R} if every open interval (a,b)(a,b), however small, contains a point of SS. Equivalently: S is dense in R  ⟺  for every x∈R there is a sequence (an) in S with lim⁡n→∞an=x.S \text{ is dense in } \mathbb{R} \iff \text{for every } x \in \mathbb{R} \text{ there is a sequence } (a_n) \text{ in } S \text{ with } \lim_{n\to\infty} a_n = x. Example: Q\mathbb{Q} is dense in R\mathbb{R}, so any real number — even an irrational like 2\sqrt{2} — is the limit of a sequence of rationals.

Order limit theorem (non-negativity). If an≥0a_n \ge 0 for all nn and lim⁡an=a\lim a_n = a, then a≥0a \ge 0. The limit cannot "fall below" a bound that every term respects.

Closed interval contains its limits. If (cn)(c_n) is a sequence in the closed interval [a,b][a,b] and cn→cc_n \to c, then c∈[a,b]c \in [a,b]. This fails for open intervals: (1n)\left(\tfrac1n\right) lies in (0,2)(0,2) but converges to 0∉(0,2)0 \notin (0,2).

Closed set (definition). A set SS is closed if, for every sequence (an)(a_n) in SS that converges, the limit also lies in SS. The result above says every closed interval is a closed set (under the standard topology on R\mathbb{R}).

Worked Example

Claim. If an≥0a_n \ge 0 for all nn and lim⁡an=a\lim a_n = a, then a≥0a \ge 0.

Proof (by contradiction). Suppose instead a<0a < 0. Then −a2>0-\tfrac{a}{2} > 0, so we may take it as our ε\varepsilon. By convergence there is an index NN with ∣an−a∣<−a2|a_n - a| < -\tfrac{a}{2} for all n≥Nn \ge N, which unfolds and rearranges as:

∣an−a∣<−a2a2<an−a<−a23a2<an<a2<0.\begin{aligned} \left|a_n - a\right| &< -\tfrac{a}{2} \\ \tfrac{a}{2} &< a_n - a < -\tfrac{a}{2} \\ \tfrac{3a}{2} &< a_n < \tfrac{a}{2} < 0. \end{aligned}

Thus an<0a_n < 0 for all n≥Nn \ge N — contradicting an≥0a_n \ge 0. Hence a≥0a \ge 0. ■\blacksquare

Application. To show [a,b][a,b] is closed, take (cn)(c_n) in [a,b][a,b] with cn→cc_n \to c. Then b−cn≥0b - c_n \ge 0 and cn−a≥0c_n - a \ge 0 for all nn, so by the theorem b−c=lim⁡(b−cn)≥0b - c = \lim(b - c_n) \ge 0 and c−a=lim⁡(cn−a)≥0c - a = \lim(c_n - a) \ge 0. Therefore a≤c≤ba \le c \le b, i.e. c∈[a,b]c \in [a,b].

Summary