Visual walkthrough — Intermediate Value Theorem

We prove the clean special case first (called Bolzano's Theorem) and lift it to the full statement at the end. Everything is built from scratch below.

Step 0 — The words before the symbols

Before any letter appears, let us agree on plain-language meanings. A function $f$ is a machine: feed it a number $x$, it returns a height $f(x)$. Picture the input $x$ running left-to-right along a horizontal ruler (the ==$x$-axis==), and the output $f(x)$ as a height measured up (positive) or down (negative) from that ruler.

WHAT the picture shows: a wiggly chalk curve with a start dot at height $f(a)$ and an end dot at height $f(b)$. WHY it matters: the whole theorem is a claim about heights the pen must pass through. Fix that image now.

Step 1 — Reduce the goal to "find a zero"

The full theorem asks: given a target height $N$ between $f(a)$ and $f(b)$, is there an input $c$ with $f(c)=N$? That extra letter $N$ is clutter. Let us remove it.

WHAT we did: replaced "hit height $N$" with "hit height $0$", i.e. find a root $g(c)=0$. WHY: a root is easier to reason about — it is just where the curve crosses the ruler. And $f(a)<N<f(b)$ becomes $g(a)<0<g(b)$: the shifted curve starts below the ruler and ends above it. PICTURE: the same curve, dropped down by $N$; the dashed target line becomes the ruler.

From here we solve the Bolzano problem: $g$ continuous, $g(a)<0<g(b)$, prove some $c$ has $g(c)=0$. We rename $g$ back to $f$ for tidiness — so assume $f(a)<0<f(b)$.

Step 2 — Collect every point that is still "below"

We need a way to point at where the crossing happens without knowing it yet. Trick: gather all the inputs where the curve is below the ruler.

WHAT: $S$ is the shaded stretch of the ruler sitting under the below-water part of the curve. WHY this set? Its right-hand edge is exactly where "below" gives way to "not below" — the suspect crossing point. Note two facts we will need:

• $S$ is not empty: $a$ is in it, because $f(a)<0$.
• $S$ is bounded above: nothing in $S$ exceeds $b$.

PICTURE: chalk-blue shading over the sub-ruler region; the right edge marked with a "?".

Step 3 — Summon the edge point $c$ (this is where the reals earn their keep)

The right edge of $S$ looks obvious in the picture — but does a real number actually sit there? For the rationals $\mathbb{Q}$ it might not (an edge can fall into a "gap"). The Completeness Axiom of Real Numbers guarantees it does for $\mathbb{R}$.

WHAT: we named the edge $c$. WHY: this is the entire reason IVT is true on $\mathbb{R}$ and false on $\mathbb{Q}$ — see Completeness Axiom of Real Numbers. Without completeness the edge could be a phantom. PICTURE: the "?" replaced by a solid pale-yellow tick at $x=c$, with the tightest ceiling line resting on $S$.

Step 4 — Squeeze from the left: $f(c)\le 0$

We now trap the height $f(c)$ between two inequalities. First from the left.

Because $c$ is the least upper bound, points of $S$ crowd right up against it: we can pick a sequence $x_1,x_2,\dots$ inside $S$ with $x_n\to c$ (arrows pointing at $c$ from the left). Each has $f(x_n)<0$.

WHAT: we showed the edge height is not above the ruler. WHY continuity here? Only continuity permits $f(\lim x_n)=\lim f(x_n)$; a jump would break it. PICTURE: blue dots marching in from the left toward $c$, their heights all below the ruler, closing onto $f(c)$.

Step 5 — Squeeze from the right: $f(c)\ge 0$

Now clamp from the other side. Any $x$ just to the right of $c$ is beyond the edge, so $x\notin S$, meaning $f(x)\ge 0$.

Pick $y_1,y_2,\dots$ with $y_n\to c$ from the right, each with $f(y_n)\ge 0$.

WHAT: the edge height is not below the ruler. WHY: right of the edge the curve has already stopped being negative — that is what "edge of $S$" means. PICTURE: pink dots marching in from the right, heights on or above the ruler, closing onto $f(c)$.

Step 6 — The pincer closes: $f(c)=0$

Two facts now collide: $f(c)\le 0\text{(Step 4)}\text{and}f(c)\ge 0\text{(Step 5)}.$

A number that is both $\le 0$ and $\ge 0$ can only be $0$.

And $c$ is genuinely inside: since $f(a)<0$ but $f(c)=0$, we have $c\ne a$; since $f(b)>0$ but $f(c)=0$, we have $c\ne b$. So $c\in (a,b)$. $■$

PICTURE: both squeeze-arrows meeting exactly on the ruler; the crossing dot glows.

Step 7 — Degenerate & edge cases (never leave a gap)

A careful walkthrough must survive every scenario the reader could throw at it.

The one-picture summary

Everything at once: the curve starts below the ruler ($f(a)<0$), ends above it ($f(b)>0$); the below-region $S$ is shaded; its right edge is $c=\sup S$; left-squeeze forces $f(c)\le 0$, right-squeeze forces $f(c)\ge 0$; they pin $f(c)=0$.

Active Recall

Connections

• Intermediate Value Theorem — the parent result these pictures derive.
• Bolzano's Theorem — the $N=0$ core case we actually proved.
• Completeness Axiom of Real Numbers — makes $c=\sup S$ exist (Step 3).
• Continuity — the load-bearing hypothesis (Steps 4–5, Case D).
• Bisection Method — turns this existence proof into a construction of $c$.
• Extreme Value Theorem — sibling closed-interval theorem.
• Rolle's Theorem and Mean Value Theorem — derivative-level cousins.
• Fixed Point Theorems — an application of the same crossing idea.