Derivatives Explained Simply
Written by Viktor Jonas · Reading time about 7 minutes
A derivative answers a single question: how fast is something changing right now? Not on average over an hour, not over a kilometre, but at exactly one instant. On a graph, that means: how steep is the curve at this very point? This is precisely what differential calculus was invented for, and today it sits inside every cruise control, every weather forecast and every curve-sketching exercise in upper secondary school.
Difference quotient — the average rate of change between two
points, that is, the slope of the secant.
Derivative (differential quotient) — the limit of the difference
quotient, the instantaneous rate of change.
Secant — a straight line that cuts the graph at two points.
Tangent — a straight line that touches the graph at one point and has
the same slope as the graph there.
Extremum — a point with a horizontal tangent, that is,
f′(x) = 0. A maximum or a minimum.
Inflection point — the curvature changes there, which shows up as
f″(x) = 0.
From secant to tangent
Finding the slope of a straight line is easy: the difference in height divided by the horizontal distance. With a curve this does not work directly, because the slope is different at every point. The trick is to be content with an approximation at first. You take the point P you want to examine and a second point Q a little further to the right. Through both you draw a straight line, the secant. Its slope is easy to compute, but it only describes the average behaviour between P and Q.
Now you slide Q closer and closer to P. The gap, usually called h, gets smaller and smaller. The secant tips over as it does so and approaches one particular position: the tangent at P. Its slope is the instantaneous rate of change of the curve that we were looking for. The limit as h goes to zero is the precise mathematical way of saying “slide closer and closer”. You cannot simply set h equal to zero, because that would leave a zero in the denominator — hence the detour through the limit.
As Q slides towards P, the secant tips into the tangent. Its slope is the derivative at P.
In practice nobody evaluates this limit every time. It was worked out once for the basic functions, and from that came rules that you can learn by heart. The most important one is the power rule: the exponent moves to the front as a factor and then goes down by one.
f(x) = x³ → f′(x) = 3x²
f(x) = 5x² → f′(x) = 10x
f(x) = 7 → f′(x) = 0
Combined: f(x) = x³ − 3x + 4
f′(x) = 3x² − 3
Slope at x = 2:
f′(2) = 3 · 4 − 3 = 9
So the curve rises nine times as steeply there as the line y = x.
A constant has derivative zero — that is not a trick question, it is easy to picture: the graph of f(x) = 7 is a horizontal line, and a horizontal line does not rise anywhere.
| Rule | Formula | Example |
|---|---|---|
| Power rule | xn → n·xn−1 | x⁴ → 4x³ |
| Constant multiple rule | c·f(x) → c·f′(x) | 6x² → 12x |
| Sum rule | f + g → f′ + g′ | x² + x → 2x + 1 |
| Product rule | f·g → f′g + fg′ | x·sin x → sin x + x·cos x |
| Chain rule | f(g(x)) → f′(g(x))·g′(x) | (2x+1)³ → 3(2x+1)²·2 |
With these five rules you can differentiate almost everything that turns up at school. The chain rule is the only one that really needs practice: you differentiate the outer function, put the inner function in unchanged, and finally multiply by the derivative of the inner function. That last multiplication is the step people forget most often.
What it is used for
The most obvious use is curve sketching: where f′(x) = 0 there is a horizontal tangent, so there may be a maximum or a minimum. Whether it really is one, and which kind, is revealed by the second derivative — if f″ is negative there, the curve bends downwards and you have a maximum; if it is positive, you have a minimum. This is the core of every optimisation problem: the largest area for a given length of fence, the least material for a given volume.
The physical reading is just as important. If you differentiate the distance travelled with respect to time, you get the velocity. Differentiate the velocity once more and you get the acceleration. The speedometer in a car shows nothing other than the first derivative of the distance function, and the feeling of being pressed into your seat is the second.
💬 Mia asks Grandpa Theo
Why can't we just set h equal to zero? The calculation would be so much shorter.
Because then you would have zero divided by zero, and that is not defined. The limit is the detour that avoids exactly this problem: you look at where the fraction is heading without ever arriving.
And the second derivative — is that the slope of the slope?
Exactly. The first one tells you how fast something changes. The second tells you how fast that change is changing. When you drive: speed and gas pedal.
Common mistakes
- Forgetting the inner derivative in the chain rule
- (2x+1)³ becomes 3(2x+1)² times 2, not just 3(2x+1)².
- Declaring every f′(x) = 0 an extremum straight away
- For f(x) = x³ we have f′(0) = 0, but the point at 0 is a saddle point, not an extremum.
- Carrying the constant along
- x² + 5 becomes 2x, not 2x + 5.
- Confusing the derivative with the function value
- f′(2) is the slope at x = 2, not the height of the graph there.
Let f(x) = x³ − 6x² + 9x. Find all extremum points and state whether each one is a maximum or a minimum.
Show solution
First derivative: f′(x) = 3x² − 12x + 9
Set it to zero: 3x² − 12x + 9 = 0, divided by 3 this gives x² − 4x + 3 = 0
With the quadratic formula: x = (4 ± √(16 − 12)) / 2 = (4 ± 2) / 2
So x₁ = 1 and x₂ = 3.
Second derivative: f″(x) = 6x − 12
f″(1) = −6, which is negative → maximum at x = 1, function value f(1) = 4
f″(3) = 6, which is positive → minimum at x = 3, function value f(3) = 0
So the curve rises up to x = 1, falls until x = 3 and then rises again.
Whoever has understood the derivative also understands its reverse: the integral gathers up again what the derivative has taken apart. Together they form the fundamental theorem of calculus — and the reason why you can work back from a velocity to the distance travelled.
Exponential growth · Sine, cosine and tangent · Falling speed