Draw any continuous path from a point below the $x$-axis to a point above it, without lifting your pen. You must cross the axis somewhere.
The goal of this handout is to understand why it needs a proof, and what property of $\mathbb{R}$ makes it true.
Note. Every proof of IVT must use something $\mathbb{Q}$ lacks:
For each scenario, decide (without computing anything) whether the IVT guarantees a solution. If yes, state what it guarantees. If no, explain which hypothesis fails.
Let $f: [a,b] \to \mathbb{R}$ be continuous. If $L$ is a real number satisfying $f(a) < L < f(b)$ or $f(a) > L > f(b)$, then there exists a point $c \in (a,b)$ such that $f(c) = L$.
Why each hypothesis is necessary:
| Condition dropped | Counterexample | What fails |
|---|---|---|
| $f$ not continuous | ||
| Domain not a closed interval (e.g. open) | ||
| $\mathbb{R}$ replaced by $\mathbb{Q}$ |
We prove the special case $f(a) < 0 < f(b)$. The general case follows by replacing $f$ with $f - L$.
Proof sketch. Let $K = \{x \in [a,b] : f(x) \leq 0\}$.
The NIP gives an algorithmic proof: we bisect our way to the root.
Proof sketch. Again assume $f(a) < 0 < f(b)$. Let $I_0 = [a,b]$ and $z = \frac{a+b}{2}$.
Bisection rule:
If $f(z) \geq 0$, let $I_1 = \phantom{[a,b]}$
If $f(z) < 0$, let $I_1 = \phantom{[a,b]}$
In either case, $I_1 = [a_1,b_1]$ satisfies
Continuing inductively, we get nested closed intervals $I_0 \supseteq I_1 \supseteq I_2 \supseteq \cdots$ with $|I_n| = \frac{b-a}{2^{n}}$.
Strategy: Evaluate $p$ at a few well-chosen integers and apply IVT on each sign-change interval.
A fixed point of $f$ is a point $x_0$ such that $f(x_0) = x_0$ — the function maps the point to itself.
Let $f: [0,1] \to [0,1]$ be continuous. Then $f$ has at least one fixed point, i.e., there exists $x_0 \in [0,1]$ with $f(x_0) = x_0$.
Proof sketch. Define $g(x) = f(x) - x$ on $[0,1]$.
There is a more conceptual proof of IVT using the topology of $\mathbb{R}$.
Let $f: G \to \mathbb{R}$ be continuous. If $E \subseteq G$ is connected, then $f(E)$ is connected.
Proof. Suppose for contradiction that $f(E)$ is not connected. Then there exist nonempty sets $A, B$ with
, ,
Define the preimages $C = f^{-1}(A)\cap E$ and $D = f^{-1}(B)\cap E$.
A function $f$ has the intermediate value property (IVP) on $[a,b]$ if for all $x < y$ in $[a,b]$ and all $L$ between $f(x)$ and $f(y)$, there exists $c \in (x,y)$ with $f(c) = L$.
IVT says: continuous $\Rightarrow$ IVP. Is the converse true?
(a) If $f: [a,b] \to \mathbb{R}$ is continuous and $f(a) \cdot f(b) < 0$, then $f$ has at least one root in $(a,b)$.
(b) If $f: [0,1] \to [0,1]$ is continuous, then $f$ has exactly one fixed point.
(c) If $f: \mathbb{R} \to \mathbb{R}$ is continuous and $\lim_{x\to -\infty} f(x) = -\infty$ and $\lim_{x \to +\infty} f(x) = +\infty$, then $f$ has at least one root.