Introduction

So far, every measure space we have studied has involved a single underlying space:

$$\left(X,\mathcal{F},\mu\right)$$

Examples include:

• $$([0,1],\mathcal{B},\lambda)$$
• $$(\mathbb{R},\mathcal{B},\lambda)$$
• Probability spaces

However, many important mathematical problems involve multiple variables.

Examples:

A function of two variables:

$$f(x,y)$$

A probability distribution of two random variables:

$$(X,Y)$$

A surface in three-dimensional space:

$$f(x,y,z)$$

To study such objects, we need a way to measure subsets of:

$$X\times Y$$

This leads to one of the most important constructions in measure theory:

the Product Measure.

---

Why We Need Product Measures

Suppose:

$$X=[0,1]$$

and:

$$Y=[0,1]$$

Both spaces have Lebesgue measure.

Now consider:

$$X\times Y=[0,1]^2$$

the unit square.

We know intuitively that:

$$\text{Area of }[0,1]^2=1$$

But how can we construct this area rigorously from the one-dimensional measures?

More generally:

If:

$$\mu$$

measures subsets of:

$$X$$

and:

$$\nu$$

measures subsets of:

$$Y$$

can we build a measure on:

$$X\times Y$$

using only:

$$\mu$$

and:

$$\nu$$

The answer is yes.

---

Cartesian Products

Recall:

$$X\times Y={(x,y):x\in X,\ y\in Y}$$

Examples:

$$\mathbb{R}\times\mathbb{R}=\mathbb{R}^2$$

$$\mathbb{R}^2\times\mathbb{R}=\mathbb{R}^3$$

If:

$$A\subseteq X$$

and:

$$B\subseteq Y$$

then:

$$A\times B$$

is called a rectangle.

---

Rectangles

A rectangle is a set of the form:

$$A\times B$$

where:

$$A\in\mathcal F$$

and:

$$B\in\mathcal G$$

Examples:

$$[0,1]\times[0,2]$$

$$[1,3]\times[5,7]$$

$$A\times B$$

for arbitrary measurable sets.

Rectangles are the building blocks of product measures.

---

Desired Property

Suppose:

$$A\subseteq X$$

and:

$$B\subseteq Y$$

If:

$$\mu(A)=3$$

and:

$$\nu(B)=5$$

then intuition suggests:

$$\text{Measure}(A\times B)=15$$

Just as:

$$\text{Area}=\text{Length}\times\text{Width}$$

we want:

$$m(A\times B)=\mu(A)\nu(B)$$

This becomes the defining property of product measures.

---

Product Sigma-Algebra

Before defining a measure, we need measurable sets.

Suppose:

$$\mathcal F$$

is a sigma-algebra on:

$$X$$

and:

$$\mathcal G$$

is a sigma-algebra on:

$$Y$$

The product sigma-algebra is written:

$$\mathcal F\otimes\mathcal G$$

and is defined as:

the smallest sigma-algebra containing every measurable rectangle:

$$A\times B$$

where:

$$A\in\mathcal F$$

and:

$$B\in\mathcal G$$

---

Why We Need the Product Sigma-Algebra

Just as Borel sets are generated from intervals,

product sigma-algebras are generated from rectangles.

Once the rectangles are measurable,

countable unions, intersections, and complements automatically become measurable.

This provides enough sets for serious analysis.

---

Definition of Product Measure

Let:

$$\left(X,\mathcal F,\mu\right)$$

and:

$$\left(Y,\mathcal G,\nu\right)$$

be σ-finite measure spaces.

Then there exists a unique measure:

$$\mu\times\nu$$

on:

$$\mathcal F\otimes\mathcal G$$

satisfying:

$$\left(\mu\times\nu\right)(A\times B)=\mu(A)\nu(B)$$

for every measurable rectangle.

This measure is called the product measure.

---

Why This Is Remarkable

The theorem says:

Once we know how to measure:

$$X$$

and:

$$Y$$

individually,

there is a unique consistent way to measure:

$$X\times Y$$

No ambiguity exists.

The measure is forced upon us.

---

Example 1: Area of a Rectangle

Take:

$$X=Y=\mathbb R$$

with Lebesgue measure.

Let:

$$A=[0,2]$$

and:

$$B=[0,3]$$

Then:

$$\lambda(A)=2$$

and:

$$\lambda(B)=3$$

Therefore:

$$\left(\lambda\times\lambda\right)(A\times B)=6$$

This recovers the familiar area formula.

---

Example 2: Volume

Suppose:

$$A=[0,2]$$

$$B=[0,3]$$

$$C=[0,4]$$

Then:

$$\lambda^3(A\times B\times C)=24$$

This is simply:

$$2\times3\times4$$

Product measures generalize ordinary volume.

---

Product Measures in Probability

Suppose:

$$X$$

and:

$$Y$$

are independent random variables.

Let:

$$P_X$$

and:

$$P_Y$$

be their probability measures.

The joint probability measure is:

$$P_X\times P_Y$$

For measurable sets:

$$A$$

and:

$$B$$

we obtain:

$$P(A\times B)=P_X(A)P_Y(B)$$

This is exactly the multiplication rule for independent events.

Thus product measures form the mathematical foundation of independence.

---

Why σ-Finiteness Appears

Notice that the theorem requires:

σ-finiteness.

This assumption ensures that:

• measures behave reasonably
• uniqueness holds
• later theorems such as Fubini’s Theorem work

Without σ-finiteness, pathological examples can occur.

---

Sections of a Set

Suppose:

$$E\subseteq X\times Y$$

Fix:

$$x\in X$$

Define:

$$E_x={y\in Y:(x,y)\in E}$$

This is called the vertical section of:

$$E$$

Similarly, fixing:

$$y\in Y$$

gives:

$$E^y={x\in X:(x,y)\in E}$$

called the horizontal section.

These sections become crucial in Tonelli’s and Fubini’s Theorems.

---

Example

Let:

$$E={(x,y)\in[0,1]^2:y\le x}$$

For fixed:

$$x$$

the section is:

$$E_x=[0,x]$$

Its measure is:

$$\lambda(E_x)=x$$

This simple observation eventually allows us to compute areas using iterated integrals.

---

Product Measure as Repeated Measurement

Suppose:

$$E\subseteq X\times Y$$

Imagine measuring:

1. Each vertical slice.
2. Then measuring how those slice-measures vary.

This intuition lies at the heart of Fubini’s Theorem.

Product measures make this procedure rigorous.

---

Connection to Lebesgue Measure in Higher Dimensions

One of the most important facts in analysis is:

$$\lambda^2=\lambda\times\lambda$$

on:

$$\mathbb R^2$$

Similarly:

$$\lambda^3=\lambda\times\lambda\times\lambda$$

on:

$$\mathbb R^3$$

Thus:

• area
• volume
• higher-dimensional volume

all arise from repeated applications of product measures.

---

Why Analysts Care

Without product measures we could not rigorously define:

• multivariable integration
• joint distributions
• stochastic processes
• Brownian motion
• Fourier analysis in higher dimensions

Product measures are one of the foundational constructions of modern mathematics.

---

Connection to Alain Connes

Product spaces represent one of the first examples of building larger spaces from smaller ones.

Later, in noncommutative geometry, Connes replaces ordinary spaces by operator algebras.

Tensor products of algebras play a role analogous to product spaces.

Thus product measures provide an early glimpse of a principle that reappears throughout advanced mathematics:

Complex structures are built by combining simpler ones in a consistent way.

---

Key Concepts Learned

By the end of this lesson you should understand:

• Product spaces have the form:

$$X\times Y$$

• Rectangles are sets of the form:

$$A\times B$$

• Product sigma-algebras are generated by measurable rectangles.
• Product measures satisfy:

$$\left(\mu\times\nu\right)(A\times B)=\mu(A)\nu(B)$$

• Product measures generalize area and volume.
• Independence in probability is built from product measures.
• Vertical and horizontal sections are fundamental tools.
• Product measures prepare the way for Tonelli’s and Fubini’s Theorems.

---

Looking Ahead

In the next lesson:

Measure Theory Lesson 22: Tonelli’s Theorem

we will prove one of the first major results about product measures. Tonelli’s Theorem explains when a double integral can be computed as repeated integrals and serves as the gateway to the even more powerful Fubini’s Theorem.