Mathematics.

sigma algebras and measurable spaces

Borel Sets

Measure Theory60 minDifficulty7 out of 10

You should know: sigma algebras, topological space

Overview

Borel sets form the smallest sigma-algebra containing all open sets of a topological space. On the real line they are generated by the open intervals and include virtually every set one encounters in analysis: open and closed sets, countable unions and intersections thereof, and their iterated combinations. Borel sets are the natural domain for Borel measures and are central to real analysis, probability theory, and descriptive set theory.

Intuition

Start with the open intervals on the real line — these are the 'building blocks'. The Borel sets are everything you can build by taking countable unions, countable intersections, and complements, starting from open sets. Concretely: every open set is Borel, every closed set is Borel (complement of open), every countable union of closed sets (F-sigma) is Borel, every countable intersection of open sets (G-delta) is Borel, and so on through the Borel hierarchy. In practice almost every set you will ever write down is a Borel set.

Formal Definition

Definition

Let (X, tau) be a topological space. The Borel sigma-algebra B(X) is the smallest sigma-algebra on X that contains every open set U in tau. Equivalently, B(X) is the intersection of all sigma-algebras on X that contain tau. Elements of B(X) are called Borel sets.

B(X)=σ(τ)={F:F is a σ-algebra on X,τF}\mathcal{B}(X) = \sigma(\tau) = \bigcap \{ \mathcal{F} : \mathcal{F} \text{ is a } \sigma\text{-algebra on } X,\, \tau \subseteq \mathcal{F} \}
Borel sigma-algebra definition
B(R)=σ ⁣({(a,b):a<b,a,bR})\mathcal{B}(\mathbb{R}) = \sigma\!\left(\{ (a,b) : a < b,\, a,b \in \mathbb{R} \}\right)
Borel sets on the real line

Notation

NotationMeaning
B(X)\mathcal{B}(X)Borel sigma-algebra on the topological space X
GδG_\deltaCountable intersection of open sets
FσF_\sigmaCountable union of closed sets

Properties

Closed under countable operations

B(X) is closed under countable unions, countable intersections, and complements.\mathcal{B}(X) \text{ is closed under countable unions, countable intersections, and complements.}

Generated by closed sets

B(X)=σ({FX:F closed})\mathcal{B}(X) = \sigma(\{ F \subseteq X : F \text{ closed}\})

Condition: Equivalent to the open-set definition

Generated by half-open intervals on R

B(R)=σ({(,a]:aR})\mathcal{B}(\mathbb{R}) = \sigma(\{ (-\infty, a] : a \in \mathbb{R}\})

Countability of generators suffices

B(R)=σ({(a,b):a,bQ})\mathcal{B}(\mathbb{R}) = \sigma(\{ (a,b) : a,b \in \mathbb{Q}\})

Condition: Rationals suffice as endpoints

Theorems

Theorem 1: Borel Hierarchy
DefineΣ10=open sets,  Π10=closed sets,  Σn+10=countable unions of Πn0 sets,  Πn+10=countable intersections of Σn0 sets. Then B(R)=n<ω(Σn0Πn0).Define \Sigma^0_1 = \text{open sets},\; \Pi^0_1 = \text{closed sets},\; \Sigma^0_{n+1} = \text{countable unions of } \Pi^0_n \text{ sets},\; \Pi^0_{n+1} = \text{countable intersections of } \Sigma^0_n \text{ sets}. \text{ Then } \mathcal{B}(\mathbb{R}) = \bigcup_{n < \omega} (\Sigma^0_n \cup \Pi^0_n).
Theorem 2: Borel sets are measurable
EveryBorelsetinRisLebesguemeasurable.Theconversefails:thereexistLebesguemeasurablesetsthatarenotBorel.Every Borel set in \mathbb{R} is Lebesgue-measurable. The converse fails: there exist Lebesgue-measurable sets that are not Borel.

Worked Examples

  1. Express the singleton as a countable intersection of open intervals.

    {x}=n=1(x1n,x+1n)\{x\} = \bigcap_{n=1}^{\infty} \left(x - \tfrac{1}{n},\, x + \tfrac{1}{n}\right)
  2. Each open interval is in B(R) by definition, so the countable intersection is also in B(R).

    {x}B(R)\{x\} \in \mathcal{B}(\mathbb{R})
  3. A countable set is a countable union of singletons, and B(R) is closed under countable unions.

    Q=qQ{q}B(R)\mathbb{Q} = \bigcup_{q \in \mathbb{Q}} \{q\} \in \mathcal{B}(\mathbb{R})

Answer: Every singleton is a G-delta set and hence Borel; every countable set is a countable union of Borel singletons and is therefore Borel.

Practice Problems

Difficulty 6/10

Prove that the set of irrationals R \ Q is a Borel set, and determine its Borel class (G-delta or F-sigma).

Difficulty 7/10

Let f: R -> R be a monotone increasing function. Show that the set of continuity points of f is a G-delta (and hence Borel).

Difficulty 5/10

Which of the following is NOT necessarily a Borel set in R?

Common Mistakes

Common Mistake

Every subset of R is a Borel set.

False. Assuming the axiom of choice, there exist non-measurable (hence non-Borel) subsets of R such as Vitali sets. There are even Lebesgue-measurable sets that are not Borel.

Common Mistake

The Borel sigma-algebra equals the power set on uncountable spaces.

On R, B(R) is strictly smaller than the power set P(R); |B(R)| = 2^{aleph_0} = |R|, but |P(R)| = 2^{|R|} > |R|.

Common Mistake

Borel sets are only open or closed sets.

Borel sets form a rich hierarchy including G-delta, F-sigma, G-delta-sigma, and so on through transfinitely many levels. Most analytic sets encountered in practice are Borel.

Quiz

The Borel sigma-algebra B(X) is defined as:
A G-delta set is:
Is every Lebesgue-measurable set a Borel set?

Historical Background

Émile Borel introduced the notion in the 1890s while studying the foundations of measure theory. Henri Lebesgue later extended the construction to define Lebesgue measure, and Felix Hausdorff and Mikhail Suslin clarified the hierarchical structure of Borel sets via the Borel hierarchy.

  1. 1898

    Borel defines measurable sets in his thesis

    Émile Borel

  2. 1902

    Lebesgue extends Borel's ideas to construct Lebesgue measure

    Henri Lebesgue

  3. 1916

    Suslin discovers analytic sets, strictly beyond Borel sets in the hierarchy

    Mikhail Suslin

Summary

  • The Borel sigma-algebra B(X) is the smallest sigma-algebra containing all open sets of a topological space X.
  • On R it is generated equivalently by open intervals, closed sets, or half-open rays (-inf, a].
  • The Borel hierarchy organises sets by how many times one alternates between countable unions and intersections.
  • Every Borel set is Lebesgue-measurable, but not conversely.
  • Borel sets are the natural setting for regular Borel measures and most constructions in real analysis and probability.

References

  1. BookFolland, G. B. (1999). Real Analysis: Modern Techniques and Their Applications (2nd ed.). Wiley.
  2. BookRoyden, H. L., & Fitzpatrick, P. M. (2010). Real Analysis (4th ed.). Pearson.