The question a derivative answers
Before any rules — power rule, product rule, chain rule — a derivative answers one question: how fast is something changing, right now, at this exact instant?
"Right now" is the hard part. Average speed over an hour is easy: divide distance by time. Speed at exactly 3:00:00 PM is harder, because at a single instant, no time has passed and no distance has been covered — dividing zero by zero gives nothing. The derivative is the tool that resolves this, by asking what the average rate of change approaches as the time interval shrinks toward zero.
From average rate to instantaneous rate
Take a function f(x). The average rate of change between two points x and x + h is:
[f(x + h) - f(x)] / h
This is the slope of the secant line connecting those two points on the graph. As h shrinks — the second point creeping closer to the first — that secant line rotates toward the tangent line at x. The derivative is defined as the limit of this slope as h approaches 0:
f'(x) = lim(h→0) [f(x + h) - f(x)] / h
Every derivative rule you'll ever learn is a shortcut for evaluating this limit for a particular kind of function, so it doesn't have to be computed from scratch every time. But the limit is what a derivative is — the rules are just faster ways to get the same answer.
Reading the notation
f'(x) reads as "f prime of x" — the derivative of f with respect to x. Leibniz notation, dy/dx, reads as "the derivative of y with respect to x" and is useful for making the units explicit: if y is measured in meters and x in seconds, dy/dx is in meters per second. Both notations mean the same thing; which one shows up depends on which is more convenient for a given problem — Leibniz notation tends to win when multiple variables are involved.
What "rate of change" means beyond slope
A derivative is a slope, but slope is a stand-in for something broader: how sensitive an output is to a small change in input. Velocity is the derivative of position with respect to time. Acceleration is the derivative of velocity with respect to time — a derivative of a derivative. Marginal cost in economics is the derivative of total cost with respect to quantity produced. None of these are literally about the slope of a line on paper; they're all instances of the same underlying idea, that a derivative measures sensitivity to change.
Where a derivative doesn't exist
A function can fail to have a derivative at a point for a few distinct reasons, and recognizing them saves a lot of confusion:
- A sharp corner (like
|x|atx = 0) — the slope approaching from the left doesn't match the slope approaching from the right, so there's no single tangent line. - A vertical tangent — the slope is technically infinite, which isn't a real number, so the derivative is undefined there.
- A discontinuity — if the function jumps, there's no well-defined tangent line at the jump.
A function has to be continuous at a point to be differentiable there, but continuity alone isn't enough — |x| is continuous at 0 but still has no derivative there, because of the corner.
Why this matters beyond the exam
Once derivatives click, a lot of calculus stops being separate topics and starts being one idea applied in different directions: optimization problems are "find where the derivative is zero," related rates problems are "differentiate both sides with respect to time," and curve sketching is "read the sign of the derivative to know where a function rises or falls." The limit definition is the foundation everything else sits on — worth understanding once, carefully, rather than skipping straight to the shortcuts.