How one limit turns average change into slope, velocity, and sensitivity at a single point.
By Theory Commons Editors4 min readPublished Aug 22, 2026
An average rate compares two points. A derivative asks what the rate becomes when those points collapse into one. That small conceptual move gives mathematics a language for motion, growth, optimization, and sensitivity.
From secant to tangent
Between x and x+h, the average rate of change of f is
hf(x+h)−f(x).
The numerator measures output change; the denominator measures input change. To isolate the rate at x, we let the interval shrink:
For f(x)=x2,
h(x+h)2−x2=2x+h,
so f′(x)=2x. The slope depends on where we stand: at x=3, the local slope is 6.
A derivative has units
If s(t) is measured in meters and t in seconds, then s′(t) is measured in meters per second. If cost C(q) is measured in dollars and quantity q in items, then C′(q) is dollars per item.
Units reveal the derivative’s meaning: it always measures output change per unit input change.
Local linearity
Near x=a, a differentiable function satisfies
f(a+Δx)≈f(a)+f′(a)Δx.
This approximation predicts small changes without recalculating the entire function. It is also why derivatives power error estimates and numerical methods.
Differentiability is stronger than continuity
A function must be continuous wherever it is differentiable, but continuity alone is not enough. The function f(x)=∣x∣ is continuous at zero, yet its left slope is −1 and its right slope is 1. The corner prevents a single derivative from existing.
Derive and interpret
For f(x)=1/x, the difference quotient simplifies to −1/[x(x+h)], so f′(x)=−1/x2. Its negative sign agrees with the graph’s decreasing behavior on each domain interval.
Check your understanding
Does f′(a)=0 guarantee an extremum?
Show the reasoningNo. $f(x)=x^3$ has $f'(0)=0$ but increases through zero.