Introduction to Lebesgue Outer Measure

TL;DR

Lebesgue outer measure is a way to assign a "size" to any subset of real numbers, extending the intuitive idea of length. It's built by covering sets with open intervals and then taking the smallest possible total length of these coverings. This allows us to measure sets that are too complex for standard interval-based length.

1. The Mental Model

Imagine you have a weird, scattered collection of points and intervals on a line. You want to measure its total "length." Lebesgue outer measure is like trying to cover this collection with an infinite number of tiny, clear plastic rulers (open intervals) and then finding the most efficient way to do so, summing up the lengths of those rulers.

2. The Core Material

You know how to measure the length of an interval, say $[a,b]$, which is simply $b-a$. But what about more complex sets? What if you have a set that's a union of many disjoint intervals, or even something much stranger like the Cantor set? Lebesgue outer measure, denoted $\mu^*(E)$, gives us a way to assign a size to any subset $E$ of the real numbers.

The Basic Idea: Covering with Intervals

Photo by Sergey Meshkov on Pexels

The core idea is to cover your set $E$ with a countable collection of open intervals. Why open intervals? Because they are the simplest sets whose "length" is clearly defined, and they behave nicely with limits.

Let $E$ be any subset of $\mathbb{R}$. We want to find its outer measure, $\mu^*(E)$. We do this by:
1. Choosing a countable collection of open intervals: Let $(I_k)_{k=1}^\infty$ be a sequence of open intervals such that $E \subseteq \bigcup_{k=1}^\infty I_k$. This means the union of these intervals completely covers your set $E$.
2. Summing their lengths: For each such covering, we calculate the sum of the lengths of the intervals: $\sum_{k=1}^\infty \text{length}(I_k)$.
3. Taking the infimum: The outer measure $\mu^*(E)$ is the infimum (the greatest lower bound) of all possible sums you can get from all such countable coverings.

In mathematical terms:
$$ \mu^*(E) = \inf \left\{ \sum_{k=1}^\infty \text{length}(I_k) : E \subseteq \bigcup_{k=1}^\infty I_k, \text{ each } I_k \text{ is an open interval} \right\} $$

Properties of Lebesgue Outer Measure

Photo by Thirdman on Pexels

• Non-negativity: $\mu^*(E) \geq 0$ for all sets $E$. Lengths can't be negative!
• Monotonicity: If $A \subseteq B$, then $\mu^*(A) \leq \mu^*(B)$. If a set is contained within another, its "size" can't be larger.
• Countable Subadditivity: For any countable sequence of sets $(E_k)_{k=1}^\infty$,
  $$ \mu^*\left(\bigcup_{k=1}^\infty E_k\right) \leq \sum_{k=1}^\infty \mu^*(E_k) $$
  This is a crucial property. It means that the measure of a union of sets is less than or equal to the sum of their individual measures. It's not always equality (that's where measurability comes in, for later topics).
• Measure of an interval: For any interval $I$ (open, closed, half-open, bounded), $\mu^*(I) = \text{length}(I)$. This confirms that Lebesgue outer measure agrees with our usual notion of length for intervals.
• Measure of a countable set: If $E$ is a countable set (like $\mathbb{N}$, $\mathbb{Z}$, or $\mathbb{Q}$), then $\mu^*(E) = 0$. This is because you can cover each point with an arbitrarily small interval.

graph TD
    A["Start with a set E you want to measure"] --> B["Cover E with open intervals (I_k)"]
    B --> C["Calculate sum of lengths of these intervals: Sum(length(I_k))"]
    C --> D{"Are there other coverings?"}
    D -- Yes --> B
    D -- No --> E["Take the infimum (greatest lower bound) of all such sums"]
    E --> F["Result is the Lebesgue Outer Measure μ*(E)"]

3. Worked Example

Let's find the Lebesgue outer measure of a single point, say $E = \{x_0\}$.

1. Choose a covering: We need to cover the single point $x_0$ with open intervals.
   A simple covering is $I_1 = (x_0 - \epsilon/2, x_0 + \epsilon/2)$ for any $\epsilon > 0$.
   The length of this interval is $(x_0 + \epsilon/2) - (x_0 - \epsilon/2) = \epsilon$.
   The sum of the lengths is just $\epsilon$ (since it's only one interval).

2. Take the infimum: We need to consider all possible coverings and take the infimum of their total lengths.
   Since we can make $\epsilon$ arbitrarily small, we can cover the point $x_0$ with an interval of length $\epsilon$, where $\epsilon$ can be any positive number.
   The set of all possible sums of lengths would include values like $1, 0.1, 0.01, 0.001, \dots$ (by choosing different $\epsilon$).
   The infimum of the set $\{ \epsilon : \epsilon > 0 \}$ is $0$.

Therefore, $\mu^*(\{x_0\}) = 0$. This makes sense: a single point has no "length."

4. Key Takeaways

• Lebesgue outer measure extends the concept of length to any subset of real numbers.
• It's defined by covering a set with open intervals and finding the smallest possible total length of such coverings.
• The mathematical formulation involves an infimum over sums of interval lengths.
• Outer measure is non-negative, monotonic, and countably subadditive.
• The outer measure of any single point or any countable set is 0.
• For any standard interval, its outer measure equals its usual length.

Common mistakes you should avoid:
- Confusing outer measure with the actual Lebesgue measure (which requires measurability).
- Assuming countable additivity instead of subadditivity for general sets.
- Forgetting that the covering intervals must be open.
- Not understanding that the infimum step is crucial to get the "tightest" possible covering.

5. Now Try It

Calculate the Lebesgue outer measure of the set $E = \{1, 2, 3\}$.

What to do:
1. Think about how you would cover these three points with open intervals.
2. Consider how you can make the total length of these covering intervals as small as possible.
3. Apply the definition of Lebesgue outer measure (the infimum).

What success looks like:
You should be able to show that $\mu^*(E) = 0$ by constructing a sequence of coverings whose total lengths approach zero.