KO
Apr 15, 2026
What is the Baire Category Theorem and why is it important in analysis?
I am studying real analysis and I'm learning about the Baire Category Theorem. It states:
In a complete metric space, the intersection of countably many dense open sets is dense.
Equivalently: A complete metric space cannot be expressed as a countable union of nowhere dense sets.
My questions:
- What does "nowhere dense" mean? How is it different from "not dense"?
- What is the intuition behind the Baire Category Theorem?
- What are some stunning applications of the theorem?
- How do I prove that is not a set using Baire?
- How does Baire imply that there exist continuous functions that are nowhere differentiable?
I want to understand why this theorem is so fundamental.
1 answers223 views
Loading comments...
1 Answer
AO
Amara OkaforApr 15, 2026 Accepted**Nowhere Dense vs Not Dense:**
- **Not dense:** $\bar{A} \
eq X$ (closure is not the whole space). Example: $(0, 1)$ in $\mathbb{R}$.
- **Nowhere dense:** $\text{int}(\bar{A}) = \varnothing$ (the closure has empty interior). Example: Cantor set, any finite set, $\{1/n : n \in \mathbb{N}\}$.
A nowhere dense set is "full of holes" — it contains no open interval.
**Intuition:**
The Baire Category Theorem says complete metric spaces are "fat" — they cannot be covered by countably many skinny (nowhere dense) sets. Categories:
- **First category (meager):** Countable union of nowhere dense sets
- **Second category:** Not first category
Baire says every complete metric space is second category.
**Application 1: $\mathbb{Q}$ is not $G_\delta$**
A $G_\delta$ set is a countable intersection of open sets. $\mathbb{Q} = \bigcap_{n} U_n$ with each $U_n$ open and dense. But $\mathbb{R}$ is complete, so $\bigcap U_n$ must be dense. $\mathbb{Q}$ is not dense in the sense of containing a dense $G_\delta$... Wait, $\mathbb{Q}$ IS dense, so it could be $G_\delta$. Actually, if $\mathbb{Q} = \bigcap G_n$ with $G_n$ open, each $G_n$ is dense (contains $\mathbb{Q}$). Then each $\mathbb{R} \setminus \{q_n\}$ (rationals removed one at a time) is dense open, and their intersection is $\mathbb{R} \setminus \mathbb{Q}$ (irrationals), which is dense by Baire. So irrationals are dense — true, but not a contradiction.
The classic result: $\mathbb{Q}$ is NOT a $G_\delta$ set because if it were, then $\mathbb{R} \setminus \mathbb{Q}$ would be $F_\sigma$, and both being $G_\delta$ would contradict Baire.
**Application 2: Nowhere differentiable continuous functions exist.**
Let $E_n = \{f \in C[0,1] : \exists x \text{ with } |f(x+h)-f(x)| \leq n|h| \text{ for all small } h\}$.
Each $E_n$ is nowhere dense in $C[0,1]$ (with sup norm). Since $C[0,1]$ is complete, Baire says $\bigcup E_n \
eq C[0,1]$. Any $f$ not in the union is continuous but nowhere differentiable. So "most" continuous functions (in the Baire sense) are nowhere differentiable!
Loading comments...