From Sets to Measurable Spaces
In the previous lesson, we learned that not every subset of a set can be measured.
To avoid contradictions and pathological objects, we introduced a sigma-algebra:
$$\mathcal{F}$$
which specifies exactly which subsets are measurable.
At this point we have two objects:
- A set of points
- A collection of measurable subsets
Together they form one of the most fundamental objects in modern mathematics.
Definition of a Measurable Space
A measurable space is a pair:
$$\left(X,\mathcal{F}\right)$$
where:
- $$X$$ is a set
- $$\mathcal{F}$$ is a sigma-algebra on $$X$$
That’s it.
No measure yet.
No probability yet.
No integration yet.
Just a set together with a specification of which subsets are measurable.
Why This Matters
Suppose:
$$X=[0,1]$$
By itself, this interval tells us nothing about which subsets are measurable.
We must also specify:
$$\mathcal{F}$$
Only then do we know which sets we are allowed to discuss mathematically.
The measurable space is therefore the stage upon which measure theory takes place.
An Analogy
Think of:
$$X$$
as a country.
The sigma-algebra:
$$\mathcal{F}$$
acts like a map showing the regions that can be legally surveyed.
Without the map:
- locations exist
- but measurement rules do not
The measurable space provides the framework within which measurements become possible.
Example 1
Let:
$$X=\{1,2,3\}$$
and
$$\mathcal{F}={\emptyset,X}$$
Then:
$$\left(X,\mathcal{F}\right)$$
is a measurable space.
Only two sets are measurable:
- the empty set
- the whole space
Nothing else can be assigned a measure.
Example 2
Let:
$$X=\{1,2,3\}$$
and
$$\mathcal{F}=\mathcal{P}(X)$$
where:
$$\mathcal{P}(X)$$
is the power set.
Then every subset is measurable.
This is also a measurable space.
The Same Set Can Produce Different Measurable Spaces
This is an important idea.
The underlying set may remain unchanged while the sigma-algebra changes.
For example:
$$X=[0,1]$$
can be paired with different sigma-algebras.
Thus:
$$\left(X,\mathcal{F}_1\right)$$
and
$$\left(X,\mathcal{F}_2\right)$$
may be completely different measurable spaces.
The measurable structure matters as much as the underlying points.
The Borel Sigma-Algebra
The most important sigma-algebra in analysis is the Borel sigma-algebra.
It begins with all open sets.
Examples:
$$(-1,1)$$
$$(0,2)$$
$$(3,5)$$
We then repeatedly apply:
- complements
- countable unions
- countable intersections
until no new sets can be generated.
The resulting sigma-algebra is called:
$$\mathcal{B}(\mathbb{R})$$
the Borel sigma-algebra.
Why the Borel Sigma-Algebra Is Important
Most sets encountered in analysis are Borel sets.
Examples include:
- intervals
- open sets
- closed sets
- countable unions of intervals
- countable intersections of intervals
Probability theory is largely built upon Borel sets.
Generated Sigma-Algebras
Often we begin with a collection of sets:
$$\mathcal{A}$$
which is not necessarily a sigma-algebra.
We then ask:
What is the smallest sigma-algebra containing these sets?
This sigma-algebra is denoted:
$$\sigma(\mathcal{A})$$
and is called the sigma-algebra generated by:
$$\mathcal{A}$$
Example
Suppose:
$$X=\{1,2,3,4\}$$
and:
$$\mathcal{A}=\{\{1,2\}\}$$
The generated sigma-algebra is:
$$\{\emptyset,\{1,2\},\{3,4\},X\}$$
because closure under complements and unions forces those additional sets to appear.
Why Generated Sigma-Algebras Matter
Later, when studying random variables, we will encounter statements such as:
Let $$\sigma(X)$$ denote the sigma-algebra generated by the random variable.
This idea appears everywhere:
- probability
- stochastic processes
- ergodic theory
- information theory
and eventually in parts of noncommutative geometry.
To generate a sigma-algebra:
- Start with your given sets.
- Add all complements.
- Add all countable unions.
- Add all countable intersections.
- Keep repeating until no new sets appear.
Measurable Functions: A Preview
Suppose:
$$f:X\to Y$$
is a function.
If:
$$X$$
and
$$Y$$
are measurable spaces, we want the function to respect their measurable structures.
This leads to the concept of a measurable function.
A measurable function is the measure-theoretic analogue of a continuous function in topology.
It is one of the most important ideas in all of analysis.
We will study it in detail soon.
Why Measure Theory Starts Here
Many students think measure theory begins with measures.
It does not.
The logical order is:
- Sets
- Sigma-algebras
- Measurable spaces
- Measures
- Integration
A measure cannot exist until a measurable space already exists.
The measurable space is the foundation.
Connection to Probability
A probability space is:
$$\left(\Omega,\mathcal{F},P\right)$$
Notice what appears first:
$$\left(\Omega,\mathcal{F}\right)$$
This is a measurable space.
Probability is simply a measure added afterward.
Thus every probability space begins as a measurable space.
Connection to Alain Connes
At first glance, measurable spaces may seem innocent.
However, one of the deepest questions in modern mathematics is:
How much of a space can be recovered from the functions defined on it?
Classical measure theory starts with:
$$\left(X,\mathcal{F}\right)$$
Connes eventually replaces ordinary spaces with operator algebras.
In his framework:
- points may disappear
- sets may disappear
- geometry survives through algebraic structure
The measurable space is therefore the first step toward understanding what information about a space is truly fundamental.
Key Concepts Learned
By the end of this lesson you should understand:
- A measurable space is:
$$\left(X,\mathcal{F}\right)$$
- A sigma-algebra determines which sets are measurable.
- Different sigma-algebras create different measurable spaces.
- The Borel sigma-algebra is the most important sigma-algebra in analysis.
- Generated sigma-algebras are built from smaller collections of sets.
- Probability spaces are measurable spaces equipped with a measure.
- Measurable spaces come before measures.
Looking Ahead
In the next lesson:
Lesson 5: Measures
we finally introduce the central object of measure theory:
$$\mu$$
which assigns a size, mass, length, area, volume, or probability to measurable sets.
This is where measure theory truly begins.
Leave a Reply