Foundations — Second derivative test — concavity, inflection points

Before you can use the second derivative test, every symbol on that page must feel obvious. This page builds them one at a time, from nothing, in the order they depend on each other.

1. A function $f$ and its graph — the "road"

• $x$ = how far right you have walked (horizontal axis).
• $f(x)$ = how high the road is at that spot (vertical axis).

WHY the topic needs it: everything — slope, bending, peaks, valleys — is a statement about the shape of this road. If you can't picture the road, nothing else lands.

2. Slope and the first derivative $f^{\prime }$ — "how steep, and which way"

Look at any tiny stretch of the road. Over a small step to the right of size $\Delta x$ (the symbol $\Delta$ just means "a little change in"), the height changes by $\Delta f$. The slope is the ratio

$\text{slope}=\frac{\Delta f}{\Delta x}=\frac{\text{rise}}{\text{run}}.$

The little tick mark ${}^{\prime }$ is just notation for "the derivative of". The tall symbol ${\lim }_{\Delta x\to 0}$ is the limit: the value the ratio approaches as the step gets infinitely small.

WHY the limit and not just a small step? A slope depends on which two points you pick unless the step is truly zero-sized. The limit is the tool that answers "what is the steepness at exactly this one point?" — a question a plain fraction cannot answer.

WHY the topic needs it: the second derivative test only fires at flat spots ($f^{\prime }=0$), and its rescue plan (the First derivative test) reads the sign of $f^{\prime }$ on each side. No $f^{\prime }$, no test.

3. Critical points — "where the road goes flat"

WHY only these? To be at the very top of a bump or very bottom of a dip, the road cannot still be tilting — if it tilted, you could walk a little to go higher or lower. So a peak/valley of a smooth road must have slope zero. See Critical points for the full argument.

WHY the topic needs it: the second derivative test is a sentence that begins "Let $f^{\prime }(c)=0$…". Critical points are its starting ingredient.

4. The second derivative $f^{\prime \prime }$ — "slope of the slope"

Here is the key move. We already turned a road $f$ into a new road $f^{\prime }$ whose height at each $x$ is the slope of the original road there. Since $f^{\prime }$ is itself just a function (another road), we can ask the same question about it: how steep is $f^{\prime }$?

Read the sign the same way you read any slope, but applied to $f^{\prime }$:

• $f^{\prime \prime }(x)>0$: the slope $f^{\prime }$ is increasing (getting more uphill / less downhill as you move right).
• $f^{\prime \prime }(x)<0$: the slope $f^{\prime }$ is decreasing.
• $f^{\prime \prime }(x)=0$: the slope is momentarily not changing.

WHY a second derivative and not something new? "Is the curve bending?" is literally the question "is the slope changing?", and change of any quantity is measured by its derivative. So the natural — and only — tool is to differentiate $f^{\prime }$ once more. No new machinery invented; we reuse Section 2's tool on Section 2's output.

5. Concavity — from "slope increasing" to "cup vs cap"

Now connect the sign of $f^{\prime \prime }$ to a picture of bending.

WHY the topic needs it: concavity is the whole point. At a flat spot, a $\cup$ means you're at the valley floor (minimum) and a $\cap$ means you're at the hill top (maximum). See Concavity and convex functions.

6. Inflection point — "where the bending switches"

The candidate spots are where $f^{\prime \prime }(c)=0$ or $f^{\prime \prime }(c)$ doesn't exist — but "candidate" is the operative word (a zero that doesn't flip sign, like $x^4$ at $0$, is not an inflection). Section 5's picture shows you exactly what a genuine flip looks like.

WHY "changes sign" and not just "$=0$"? Think of a smile deepening then flattening then flipping to a frown — only the flip is an inflection. A smile that merely flattens for an instant and stays a smile never changed which way it holds water.

7. Taylor's little sum — why the test is true

The parent proves the test using one line of a Taylor expansion. You only need the idea, not the machinery (full story: Taylor series).

WHY the topic needs it: this is the engine that turns "concave up at a flat point" into the guaranteed conclusion "local minimum". It is the bridge between Section 5's picture and the classification rule.

Prerequisite map

Equipment checklist

Give each a genuine attempt, then reveal.

Connections

• Second derivative test — concavity, inflection points
• First derivative test
• Critical points
• Concavity and convex functions
• Taylor series
• Curve sketching
• Maxima and minima — optimization