The Derivative as a Microscope

Derivative as slope of the tangent line — f(x) = x²

a = 1.002

Slope of tangent at x = a equals f′(a) = 2a

Zoom in on a smooth curve. Keep zooming. At some point, the curve stops looking like a curve and starts looking like a straight line.

This is not an illusion — it is what differentiability means. A function is differentiable at a point if and only if, when you zoom in far enough, the graph looks like a straight line. That line is the tangent line. Its slope is the derivative.

What "zoom in" actually means

If $f$ is differentiable at $x = a$ , then near $a$ :

$f(a + h) \approx f(a) + f'(a) \cdot h$

This is the linear approximation. The error in this approximation shrinks faster than $h$ itself — meaning the approximation becomes proportionally more accurate as you zoom in.

This is what the definition of derivative formalises:

$f'(a) = \lim_{h \to 0} \frac{f(a+h) - f(a)}{h}$

The fraction is the slope of the secant line through $(a, f(a))$ and $(a+h, f(a+h))$ . As $h \to 0$ , this secant slope converges to the tangent slope — if and only if the function is smooth enough to have a tangent line there.

Why some functions don't have derivatives

Not every function looks like a straight line when you zoom in. The classic example: $f(x) = |x|$ at $x = 0$ .

Zoom in on the origin from the right: you see a line with slope $+1$ . Zoom in from the left: you see a line with slope $-1$ . The view depends on which direction you approach from — there's no single tangent line, and therefore no derivative.

More exotic functions can be continuous everywhere but differentiable nowhere — like the Weierstrass function, which looks jagged at every scale no matter how far you zoom. These functions were considered mathematical pathologies in the 19th century; they're now understood as ubiquitous (most continuous functions are nowhere differentiable, in a precise technical sense).

Differentiability implies continuity — but not the reverse

If $f$ is differentiable at $a$ , it must be continuous at $a$ . The contrapositive: if $f$ is discontinuous at $a$ , it can't have a derivative there.

But continuity does not imply differentiability. The absolute value function $|x|$ is continuous everywhere and differentiable everywhere except $x = 0$ . Continuity is a weaker condition.

The microscope interpretation in practice

The linear approximation $f(a + h) \approx f(a) + f'(a) \cdot h$ is enormously useful:

Estimation: $\sqrt{4.02} \approx \sqrt{4} + \frac{1}{2\sqrt{4}}(0.02) = 2 + 0.005 = 2.005$ (true value: $2.00499\ldots$ )
Physics: Newton's laws use infinitesimal changes in position to define velocity. The derivative is the formalization of "instantaneous" speed.
Engineering: almost every numerical method for solving equations, optimising systems, or simulating physics is based on locally approximating complicated functions by their tangent lines.

Why eˣ is its own tangent line

A remarkable fact: the derivative of $e^x$ is $e^x$ itself. This means $e^x$ looks like $e^x$ at every scale — the tangent line at any point has slope equal to the function's value there.

This is equivalent to saying $e^x$ solves the simplest possible differential equation: $\frac{dy}{dx} = y$ . Growth proportional to current size. Every exponential growth or decay model in science — population, radioactive decay, compound interest — is secretly just the statement that the derivative equals the function.

The bigger picture

The derivative replaces the study of curves with the study of straight lines. Since we understand straight lines completely, this is extraordinarily powerful. When we write $f'(a) = 3$ , we are saying: near $a$ , $f$ behaves like the function $3x + c$ . The whole apparatus of derivatives — product rule, chain rule, implicit differentiation — is just the machinery for extracting these local linear approximations from composite and implicit functions.

Calculus is, in this sense, the science of local linearisation. The derivative is the instrument.