Students often memorize $f'(x)=\lim_{h\to 0}\frac{f(x+h)-f(x)}{h}$ and still feel unsure what the number means. The derivative measures how sensitive the output is to a tiny nudge in the input — an instantaneous rate of change, and equivalently the slope of the best linear approximation at that point.
From average rate to instantaneous rate
The difference quotient $\frac{f(x+h)-f(x)}{h}$ is the average rate of change of $f$ on the interval between $x$ and $x+h$. Sending $h\to 0$ asks whether those averages settle on a single number. When they do, that number is $f'(x)$.
Example If $s(t)$ is position in meters and $t$ is time in seconds, then $s'(t)$ has units meters per second: velocity. If $C(q)$ is cost in dollars to produce $q$ items, $C'(q)$ is marginal cost in dollars per item. Units are not decoration; they are the story.
Linear approximation is the point
Differentiability at $a$ means there is a number $m$ such that the error $f(a+h)-f(a)-mh$ shrinks faster than $h$ as $h\to 0$. That $m$ is $f'(a)$. Locally, the graph looks like the line $y=f(a)+f'(a)(x-a)$. Optimization, Newton’s method, and error estimates all lean on this idea.
What the derivative is not
- It is not “the next value of the function.”
- It is not automatically the percentage change (that would involve dividing by $f(x)$ when $f(x)\neq 0$).
- A zero derivative does not guarantee a local max/min (consider $x^3$ at $0$).
- Existence of $f'(a)$ is stronger than continuity at $a$, but continuity of $f'$ is stronger still.
For the rigorous neighborhood language behind limits of difference quotients, see Epsilon-delta intuition without terror. For the wider map, see Calculus foundations.
Citations & further reading
- Strang, G. Calculus. Wellesley-Cambridge / MIT OCW companion materials. MIT 18.01SC
- OpenStax, Calculus Volume 1, derivatives chapters. openstax.org
- 3Blue1Brown, “The paradox of the derivative” (video intuition for instantaneous rate). YouTube
- Paul’s Online Math Notes, “The Definition of the Derivative.” tutorial.math.lamar.edu