Chapter 5 — Key Takeaways

The One-Sentence Summary

The derivative is a single number that measures the instantaneous rate of change of a function at a point; it is defined as the limit of average rates of change (difference quotients) over shrinking intervals, and it is the central object of all of differential calculus.

From Average to Instantaneous

  1. Average rate of change. Over an interval $[a, b]$ with $a \neq b$, $$\frac{f(b) - f(a)}{b - a} = \frac{\Delta y}{\Delta x}.$$ This is a single number — the slope of the secant line through $(a, f(a))$ and $(b, f(b))$. It summarizes the whole interval and is blind to what happens in between (Section 5.1).

  2. The difference quotient. The average rate over the small interval $[a, a+h]$ is $$\frac{f(a + h) - f(a)}{h}.$$ The width $h$ is the variable we will shrink. This is the slope of the secant from $(a, f(a))$ to $(a+h, f(a+h))$ (Section 5.2).

  3. The instantaneous rate is a limit. You cannot reach the instant (setting $h = 0$ gives the meaningless $0/0$). Instead you approach it. The limit of the difference quotient as $h \to 0$ is the instantaneous rate of change — and that is the derivative (Section 5.2).

The Definition of the Derivative

  1. The derivative at a point $a$ (when the limit exists): $$f'(a) = \lim_{h \to 0} \frac{f(a + h) - f(a)}{h} \;=\; \lim_{x \to a} \frac{f(x) - f(a)}{x - a}.$$ The two forms are identical in content (substitute $x = a + h$); the first is usually easier to compute, the second is sometimes cleaner. When this limit exists, $f$ is differentiable at $a$ (Section 5.2).

  2. The derivative as a function. Run the same limit at every input to get a new function $$f'(x) = \lim_{h \to 0} \frac{f(x + h) - f(x)}{h},$$ whose value at each $x$ is the slope of $f$ there. This shift from "a number $f'(a)$" to "a function $f'$" is the bridge into Chapter 6 (Section 5.5).

  3. Higher derivatives. Since $f'$ is itself a function, differentiate again: $f''(x)$ is the rate of change of the slope, $f'''(x)$ the next, and so on (Section 5.5).

Three Interpretations — One Object

  1. Geometric. $f'(a)$ is the slope of the tangent line to $y = f(x)$ at $(a, f(a))$. The tangent is the limiting position of secant lines, and its equation is $$y - f(a) = f'(a)\,(x - a). \qquad (\text{Section 5.3})$$

  2. Kinematic. If $s(t)$ is position, then $s'(t)$ is velocity and $s''(t)$ is acceleration. Speed is $|v(t)|$ (Section 5.4).

  3. Marginal. In economics, $C'(q)$ is marginal cost, $R'(q)$ is marginal revenue — the cost or revenue of one more unit (Section 5.8). The same limit also reads as engineering sensitivity and as the local linear approximation $f(a+h) \approx f(a) + f'(a)h$ (the seed of Chapter 11).

All of these are the same number $f'(a)$; only the story around it changes.

Notation (Section 5.6)

Notation Name Typical use
$f'(x)$ Lagrange this book's default
$\dfrac{df}{dx},\ \dfrac{dy}{dx}$ Leibniz chain rule, integrals, differentials
$\dot{x},\ \ddot{x}$ Newton dot time derivatives in physics
$Df,\ D_x f$ Operator differential equations

All name the same object; read them all without friction.

When the Derivative Fails (Section 5.7)

A function fails to be differentiable at $a$ in exactly four recognizable ways:

Failure Example at $0$ What goes wrong
Corner $|x|$ one-sided slopes exist but disagree ($-1 \neq +1$)
Cusp $x^{2/3}$ one-sided slopes run to $+\infty$ and $-\infty$
Vertical tangent $x^{1/3}$ slope $\to +\infty$; tangent is vertical (no slope)
Discontinuity any jump/hole not continuous $\Rightarrow$ cannot be differentiable

Key theorem. Differentiability implies continuity — but not conversely. A differentiable function is automatically continuous; a continuous function can still have corners, cusps, or vertical tangents (the Weierstrass function is continuous everywhere yet differentiable nowhere). Smoothness is strictly stronger than connectedness.

A Few Derivatives, Straight From the Definition (Section 5.9)

$$\frac{d}{dx}(c) = 0, \qquad \frac{d}{dx}(mx + b) = m, \qquad \frac{d}{dx}(x^2) = 2x, \qquad \frac{d}{dx}(x^3) = 3x^2.$$

The last two hint at the power rule $\frac{d}{dx}(x^n) = n\,x^{n-1}$ — but the general rule, along with the product, quotient, and chain rules, is proved in Chapter 7. In this chapter we differentiate the slow, honest way; Chapter 7 makes it fast.

Common Errors to Avoid

  • Confusing average value with average rate of change. The average rate is $\frac{f(b)-f(a)}{b-a}$ (a ratio of differences); the average value, $\frac{1}{b-a}\int_a^b f$, is a different object with different units (Section 5.1; average value defined in Chapter 14).
  • Plugging $h = 0$ too early. You must simplify the difference quotient — cancel the $h$ — before taking the limit. The cancellation is legal only because $h \neq 0$ throughout the limit process.
  • "A tangent touches the curve at one point." False in general; the tangent is the limit of secants, defined by slope, not by intersection count (Section 5.3).
  • Assuming continuity gives differentiability. It does not — corners and cusps are continuous but not differentiable (Section 5.7).
  • Forgetting units and signs. A rate has units of output per unit input, and its sign encodes direction (negative = decreasing).

Connections Forward and Back

  • Built on: limits (Chapter 3) and continuity (Chapter 4) — the derivative cannot exist without the limit.
  • Leads to: the derivative as a function and the gradient-descent anchor (Chapter 6); the differentiation rules (Chapter 7); applications to extrema and concavity (Chapter 9); optimization (Chapter 10); linear approximation (Chapter 11).
  • The deep payoff: in Chapter 14 the derivative meets the integral in the Fundamental Theorem of Calculus, which reveals differentiation and accumulation as inverse operations.

Skills You Should Now Have

  • Compute an average rate of change and identify it as a secant slope.
  • Compute simple derivatives directly from the limit definition.
  • Write the equation of a tangent line at a given point.
  • Recover velocity and acceleration from a position function.
  • Identify the four ways a function fails to be differentiable.
  • Interpret a derivative geometrically, kinematically, and marginally.

Part I Complete

This is the last chapter of Part I — Limits and Continuity. Across five chapters you have:

  • Chapter 1 — motivated calculus through the tangent and area problems;
  • Chapter 2 — built function fluency and introduced plotting;
  • Chapter 3 — defined the limit formally and computed limits, including indeterminate forms;
  • Chapter 4 — connected limits to continuity, with the IVT and bisection;
  • Chapter 5 — defined the derivative as a limit and interpreted it across fields.

What's next. Part I was about meaning; Part II is about power. Chapter 6 develops the derivative as a function in its own right and launches the gradient-descent anchor; Chapter 7 delivers the differentiation rules that turn today's slow limit computations into a near-instant skill. By the end of Part II, differentiation is automatic — freeing you to apply it. Press forward.