Why We Need a Better Notion of Length

In the previous lesson, we defined a measure abstractly.

Now we want to construct the most important measure in all of analysis:

Lebesgue measure.

Its purpose is to formalize the intuitive idea of length.

For example:

$$m([0,1]) = 1$$

$$m([2,5]) = 3$$

$$m((a,b)) = b-a$$

At first glance this seems trivial.

However, extending the concept of length from intervals to extremely complicated subsets of the real line turns out to be one of the greatest achievements in mathematics.

---

What Properties Should Length Have?

Suppose:

$$m$$

represents length.

We would like:

Intervals have their usual lengths

$$m([a,b]) = b-a$$

---

Translation invariance

Moving a set should not change its length.

If:

$$A+t = {x+t : x\in A}$$

then:

$$m(A+t)=m(A)$$

Example:

$$m([0,1])=m([10,11])=1$$

Length should not depend on location.

---

Countable additivity

If:

$$A_1,A_2,\ldots$$

are disjoint,

then: $$m\left(\bigcup_{n=1}^{\infty}A_n\right)=\sum_{n=1}^{\infty}m(A_n)$$

This property allows infinite constructions.

---

The Problem

Defining length for intervals is easy.

Defining length for arbitrary sets is not.

Consider:

$$\mathbb{Q}\cap[0,1]$$

the rational numbers in the unit interval.

What should its length be?

The rationals are:

• infinitely many
• densely packed everywhere

Yet they occupy almost no space.

Our intuition suggests:

$$m(\mathbb{Q}\cap[0,1])=0$$

A successful theory should produce this result automatically.

---

Covering Sets by Intervals

Lebesgue’s key idea was ingenious.

Instead of asking:

What is the length of the set itself?

he asked:

How efficiently can the set be covered by intervals?

Suppose a set:

$$A$$

is contained inside intervals:

$$I_1,I_2,I_3,\ldots$$

Then the total covering length is:

$$\sum_{n=1}^{\infty}|I_n|$$

where:

$$|I_n|$$

denotes the interval length.

Different coverings give different totals.

Lebesgue’s idea was to take the smallest possible total.

---

Outer Measure

For any set:

$$A\subseteq\mathbb{R}$$

define:

This quantity is called the outer measure.

It represents the smallest total interval length needed to cover the set.

---

Example: A Single Point

Consider:

$$A={x}$$

A point can be covered by an interval of length:

$$\varepsilon$$

for any:

$$\varepsilon>0$$

Since we can make:

$$\varepsilon$$

arbitrarily small,

the infimum becomes:

$$m^*({x})=0$$

Thus a single point has zero length.

---

Example: A Finite Set

Consider:

$$A={1,2,3}$$

Each point has outer measure zero.

Therefore:

$$m^*(A)=0$$

Every finite set has length zero.

---

Example: Countable Sets

Consider:

$$\mathbb{N}$$

or:

$$\mathbb{Q}\cap[0,1]$$

Since the set is countable, we can enumerate its elements:

$$x_1,x_2,x_3,\ldots$$

Cover:

$$x_n$$

with an interval of length:

$$\frac{\varepsilon}{2^n}$$

The total covering length is:

$$\sum_{n=1}^{\infty}\frac{\varepsilon}{2^n}=\varepsilon
$$

Since:

$$\varepsilon$$

can be arbitrarily small,

we obtain:

$$m^*(\mathbb{Q}\cap[0,1])=0$$

This is one of the first major triumphs of Lebesgue measure.

---

Almost All Numbers Are Irrational

Since:

$$m([0,1])=1$$

and:

$$m(\mathbb{Q}\cap[0,1])=0$$

the entire measure comes from the irrational numbers.

In measure theory, we therefore say:

Almost every number in $$[0,1]$$ is irrational.

Notice:

• infinitely many rationals exist
• infinitely many irrationals exist

Yet the rationals contribute zero measure.

This illustrates how measure differs from counting.

---

Carathéodory’s Criterion

Outer measure works for every subset.

However, not every subset behaves nicely.

We need a way to decide which sets deserve to be called measurable.

A set:

$$E$$

is Lebesgue measurable if for every set:

$$A$$

we have:

This is called Carathéodory’s criterion.

It means:

The set splits every other set cleanly into two pieces without creating or destroying measure.

---

Lebesgue Measurable Sets

The collection of all sets satisfying Carathéodory’s criterion forms a sigma-algebra.

This sigma-algebra contains:

• open sets
• closed sets
• intervals
• Borel sets
• countable unions
• countable intersections

and many more.

These are the Lebesgue measurable sets.

---

Definition of Lebesgue Measure

Lebesgue measure is simply outer measure restricted to Lebesgue measurable sets.

We denote it:

$$m$$

Thus:

$$m([a,b])=b-a$$

but now the notion extends far beyond intervals.

---

Important Examples

Open interval

$$m((a,b))=b-a$$

---

Closed interval

$$m([a,b])=b-a$$

---

Half-open interval

$$m([a,b))=b-a$$

---

Single point

$$m({x})=0$$

---

Finite set

$$m(A)=0$$

---

Countable set

$$m(A)=0$$

---

Unit interval

$$m([0,1])=1$$

---

Entire real line

$$m(\mathbb{R})=\infty$$

---

Null Sets

A set with measure zero is called a null set.

Examples:

• finite sets
• countable sets
• rational numbers

Null sets play an enormous role in analysis.

Many theorems are true except on a null set.

This leads to one of the most important phrases in mathematics:

Almost everywhere.

---

Why Lebesgue Measure Changed Mathematics

Before Lebesgue, integration relied on geometric partitions.

Lebesgue introduced a fundamentally different perspective.

Instead of slicing the x-axis,

he effectively sliced according to the values taken by functions.

This idea made it possible to integrate functions that Riemann’s theory could not handle.

Modern:

• probability theory
• functional analysis
• stochastic processes
• PDEs
• ergodic theory
• quantum theory

all depend on Lebesgue measure.

---

Connection to Alain Connes

Classical measure theory studies measurable subsets of ordinary spaces.

Connes asks:

Can measure exist when the underlying space is no longer classical?

In noncommutative geometry:

• measurable sets disappear
• points may disappear
• ordinary length disappears

Yet integration survives through operator-theoretic objects called traces.

One way to view Connes’ program is:

Generalize Lebesgue measure beyond ordinary geometric spaces.

Everything begins with understanding Lebesgue measure first.

---

Key Concepts Learned

By the end of this lesson you should understand:

• Lebesgue measure generalizes ordinary length.
• Outer measure is built using interval coverings.
• Single points have measure zero.
• Finite sets have measure zero.
• Countable sets have measure zero.
• The rationals in $$[0,1]$$ have measure zero.
• Almost every real number is irrational.
• Carathéodory’s criterion defines measurable sets.
• Lebesgue measure is the most important measure in modern analysis.

---

Looking Ahead

In the next lesson:

Lesson 7: Measurable Functions

we will move from measuring sets to measuring functions, which is the critical step needed before defining the Lebesgue integral. This transition marks the beginning of modern integration theory.