Derivatives: From Slopes to Rates of Change

The derivative of a function measures its instantaneous rate of change — the slope of the tangent line at a point. Defined as the limit of average rates of change (difference quotients) as the interval shrinks to zero, written $f^{'} (x)$ or $\frac{d y}{d x}$ . It applies everywhere a quantity varies: velocity from position, marginal cost from cost, growth rates in biology. Foundational work by Newton and Leibniz, 1660s–1680s.

The derivative of a function $f$ at a point $x$ is the limit of the average rate of change of $f$ over an interval as that interval shrinks to zero length. Geometrically it is the slope of the line tangent to the graph of $f$ at $x$ ; physically it is the instantaneous rate at which one quantity changes with respect to another. It matters because nearly every dynamic process in science — motion, heating, population growth, electrical current — is governed by how fast things change, and the derivative is the precise mathematical tool for capturing "how fast." A surprising fact: a function can be continuous everywhere yet differentiable nowhere (the Weierstrass function), meaning a curve can be unbroken yet have a tangent line at no single point.

The Core Insight

The slope of a straight line between two points is easy: rise over run. But a curve has a different steepness at every point. The genius move is to compute the slope between a fixed point and a second point that we slide closer and closer to the first. The secant slopes approach a limiting value — the tangent slope — and that limit IS the derivative. "Rate of change at an instant" and "slope of the tangent" are the same idea wearing two costumes.

Formal Statement

The Difference Quotient

For a function $f$ and a point $x$ , the average rate of change over the interval from $x$ to $x + h$ is the difference quotient:

\frac{Δ y}{Δ x} = \frac{f ( x + h ) - f ( x )}{h}

⟳rise

= f (x + h) - f (x)

divided by run

= (x + h) - x = h

; this is the slope of the secant line through

(x, f (x))

and

(x + h, f (x + h))

Here $h \neq = 0$ is the step size (the horizontal gap between the two points), $f (x + h) - f (x)$ is the change in output ( $Δ y$ ), and $h = Δ x$ is the change in input.

The Derivative

The derivative of $f$ at $x$ is the limit of the difference quotient as the step size $h$ approaches zero, provided this limit exists:

f^{'} (x) = h \to 0 lim \frac{f ( x + h ) - f ( x )}{h}

⟳take equation [1] and shrink the interval to zero by letting

h \to 0

; the secant slopes converge to the tangent slope.

Symbols: $f^{'} (x)$ (Lagrange notation, read "f prime of x") is the derivative; $lim_{h \to 0}$ means "the value approached as $h$ gets arbitrarily close to 0"; $h$ is the vanishing step. An equivalent form uses a second point $a$ approaching $x$ :

f^{'} (x) = a \to x lim \frac{f ( a ) - f ( x )}{a - x}

⟳substitute

a = x + h

into [2], so

a \to x

is the same as

h \to 0

, and

a - x = h

Notations for the same object: $f^{'} (x)$ , $\frac{d y}{d x}$ (Leibniz), $\overset{y}{˙}$ (Newton, usually for time derivatives), and $D f$ (operator notation).

Differentiability Implies Continuity

If $f$ is differentiable at $x$ , then $f$ is continuous at $x$ .

f^{'} (x) exists ⟹ h \to 0 lim f (x + h) = f (x)

⟳write

f (x + h) - f (x) = \frac{f ( x + h ) - f ( x )}{h} \cdot h

; as

h \to 0

the first factor tends to

f^{'} (x)

(finite) and the second to

0

, so the product tends to

0

The converse is FALSE: continuity does not guarantee differentiability (see the corner of $∣ x ∣$ ).

Why It Holds / Derivation

Deriving the Derivative of $f(x)=x^2$ from the Limit

We compute $f^{'} (x)$ for $f (x) = x^{2}$ directly from definition [2].

Step 1. Form the difference quotient:

\frac{f ( x + h ) - f ( x )}{h} = \frac{( x + h ) ^{2} - x ^{2}}{h}

⟳substitute

f (x + h) = (x + h)^{2}

and

f (x) = x^{2}

into [1].

Step 2. Expand the numerator:

(x + h)^{2} = x^{2} + 2 x h + h^{2}

, so

\frac{x ^{2} + 2 x h + h ^{2} - x ^{2}}{h} = \frac{2 x h + h ^{2}}{h}

⟳the

x^{2}

terms cancel; what remains contains a factor of

h

Step 3. Cancel

h

(legal because

h \neq = 0

before the limit):

\frac{h ( 2 x + h )}{h} = 2 x + h

⟳factor

h

from

2 x h + h^{2}

, then divide.

Step 4. Take the limit:

f^{'} (x) = h \to 0 lim (2 x + h) = 2 x

⟳as

h \to 0

the term

h

vanishes, leaving

2 x

The slope of $y = x^{2}$ at any point $x$ is $2 x$ . At $x = 3$ the tangent slope is $6$ . ∎

Isaac Newton | 1643–1727

Newton developed his "method of fluxions" around 1665–1666 during the plague years at Woolsthorpe. He treated variables as fluents flowing in time and their rates as fluxions (denoted $\overset{x}{˙}$ ), motivated directly by physics — instantaneous velocity and acceleration in his laws of motion. His Principia (1687) used geometric limiting arguments rather than explicit derivative notation. Newton thought of the derivative dynamically: as the velocity of a generated quantity.

Building Intuition

Imagine zooming in on a smooth curve with a microscope. The more you magnify, the straighter it looks — eventually it is indistinguishable from a straight line. The slope of THAT line is the derivative. This is local linearity: a differentiable function looks linear up close.

Now picture driving. Your odometer gives position $s (t)$ ; your speedometer gives the derivative $s^{'} (t)$ — instantaneous speed. The average speed over a trip is $\frac{Δ s}{Δ t}$ (a difference quotient). The speedometer reading at one instant is the limit as the time interval shrinks to nothing. You can't divide a distance by zero time directly — but you CAN ask what the ratio approaches, and that answer is the derivative.

Secant lines (slopes 5, 4) collapsing to the tangent line (slope 2) on the curve $y=x^2$ at $x=1$.

Worked Examples

Instantaneous Velocity of a Falling Ball

Given: A ball dropped from rest has position

s (t) = 4.9 t^{2}

metres after

t

seconds. Find its instantaneous velocity at

t = 2

s. Step 1. Velocity is the derivative

v (t) = s^{'} (t)

. Form the difference quotient at

t = 2

\frac{s ( 2 + h ) - s ( 2 )}{h} = \frac{4.9 ( 2 + h ) ^{2} - 4.9 ( 2 ) ^{2}}{h}

⟳substitute

t = 2

and step

h

into definition [2].

Step 2. Expand:

4.9 (2 + h)^{2} = 4.9 (4 + 4 h + h^{2}) = 19.6 + 19.6 h + 4.9 h^{2}

, and

4.9 (4) = 19.6

, so

\frac{19.6 + 19.6 h + 4.9 h ^{2} - 19.6}{h} = \frac{19.6 h + 4.9 h ^{2}}{h}

⟳the constant

19.6

cancels.

Step 3. Cancel

h

and take the limit:

v (2) = h \to 0 lim (19.6 + 4.9 h) = 19.6

⟳factor and divide by

h

, then let

h \to 0

v (2) = 19.6 m/s

Slope of a Tangent and the Tangent-Line Equation

Given: Find the equation of the tangent line to

f (x) = x^{2} - 4 x + 3

x = 1

. Step 1. Compute the derivative from the limit (or use the power-rule result

f^{'} (x) = 2 x - 4

derived the same way as [8]):

f^{'} (x) = h \to 0 lim \frac{[( x + h ) ^{2} - 4 ( x + h ) + 3 ] - [ x ^{2} - 4 x + 3 ]}{h} = 2 x - 4

⟳expand, cancel the

h

-free terms, factor

h

, take limit — exactly the method of the

x^{2}

proof plus the linear term contributing

- 4

Step 2. Evaluate the slope at

x = 1

f^{'} (1) = 2 (1) - 4 = - 2

. Step 3. Find the point:

f (1) = 1 - 4 + 3 = 0

, so the tangent passes through

(1, 0)

. Step 4. Use point-slope form

y - y_{1} = m (x - x_{1})

y - 0 = - 2 (x - 1)

⟳substitute

m = - 2

(x_{1}, y_{1}) = (1, 0)

y = - 2 x + 2

Common Misconceptions

The derivative is NOT $\frac{f ( x + h ) - f ( x )}{h}$ with $h = 0$ . You may never set

h = 0

inside the quotient — that gives the meaningless

\frac{0}{0}

. The derivative is the limit as

h

approaches zero, which is a fundamentally different (and well-defined) operation. The cancellation of

h

(steps like [7]) is what removes the division-by-zero problem before the limit is taken.

Also: $\frac{d y}{d x}$ is a single symbol denoting a limit, NOT a fraction $d y$ divided by $d x$ . It behaves like a fraction in many situations (chain rule, separation of variables), but that convenience is a theorem, not the definition.

Continuity Without Differentiability

The function $f (x) = ∣ x ∣$ is continuous everywhere, including at $x = 0$ , yet it has NO derivative at $x = 0$ .

h \to 0^{-} lim \frac{∣0 + h ∣ - ∣0∣}{h} = - 1, h \to 0^{+} lim \frac{∣ h ∣}{h} = + 1

⟳for

h > 0

∣ h ∣/ h = 1

; for

h < 0

∣ h ∣/ h = - 1

; the one-sided limits disagree.

Because the left-hand and right-hand limits of the difference quotient differ ( $- 1 \neq = + 1$ ), the two-sided limit does not exist, so $f^{'} (0)$ does not exist. Geometrically, the graph has a sharp corner — no single tangent line. This disproves the tempting belief that "smooth-looking and unbroken" implies "differentiable."

Connections

The derivative is one of the two pillars of calculus, intimately tied to its partner through the Fundamental Theorem of Calculus, which states that differentiation and Integration are inverse operations. The mechanics of computing derivatives without going back to the limit each time are packaged into the Differentiation Rules (power, product, quotient, chain). The derivative is built on the machinery of Limits and Continuity, and it generalizes upward to the Second Derivative (rate of change of the rate of change — acceleration, concavity) and outward to Partial Derivatives for functions of several variables. Its application to finding maxima and minima is the subject of Optimization with Derivatives.

Real-World Applications

Physics — kinematics: Velocity is the derivative of position; acceleration is the derivative of velocity. Newton's second law $F = ma$ depends on the second derivative of position.
Economics — marginal analysis: Marginal cost $C^{'} (q)$ and marginal revenue $R^{'} (q)$ are derivatives; profit is maximized where they are equal.
Medicine — pharmacokinetics: The rate at which drug concentration in the bloodstream rises or falls is a derivative, guiding dosing intervals.
Engineering — control systems: PID controllers use the Derivative term to anticipate error trends and stabilize systems like cruise control and thermostats.
Biology — population dynamics: Growth rate $\frac{d P}{d t}$ in logistic and exponential models predicts when a population peaks or crashes.
Machine learning — gradient descent: Training neural networks minimizes a loss function by following the negative derivative (gradient) downhill, the engine behind modern AI.
Geography/Hydrology — slope and flow: The derivative of terrain elevation gives slope steepness, determining water runoff and erosion paths.

Historical Development

c. 250 BCE: Archimedes uses the "method of exhaustion" and tangent-like constructions to study the spiral, foreshadowing limiting processes.
1629: Pierre de Fermat develops adequality, a method for finding maxima/minima and tangents that effectively computes derivatives via a vanishing increment.
1637: René Descartes, in La Géométrie, gives an algebraic method of normals and tangents to curves.
1665–1666: Isaac Newton invents the method of fluxions during the plague years, treating rates of change of flowing quantities.
1675–1684: Gottfried Wilhelm Leibniz independently develops differential calculus, introducing the $d y / d x$ and $\int$ notations and publishing in Acta Eruditorum (1684).
1690s–1716: The bitter Newton–Leibniz priority dispute divides British and Continental mathematics for over a century.
1734: Bishop George Berkeley, in The Analyst, attacks the logical foundations of infinitesimals, calling them "ghosts of departed quantities."
1821–1823: Augustin-Louis Cauchy gives the first rigorous definition of the derivative based on limits in his Cours d'analyse.
1872: Karl Weierstrass constructs a function that is continuous everywhere but differentiable nowhere, and provides the rigorous $ε$ – $δ$ foundation underpinning the modern definition.

Subtleties & Limits

The classical derivative requires a function to be smooth enough to have a tangent — but vast classes of important functions are not. Brownian motion (the random path of a pollen grain in water) is continuous yet nowhere differentiable, like the Weierstrass function. This forced mathematicians to invent broader notions: the weak derivative and distributional derivative (Sobolev, Schwartz, mid-20th century) let us differentiate sharp corners and even the Dirac delta "function." Fractional calculus asks what a "half-derivative" $\frac{d ^{1/2}}{d x ^{1/2}}$ could mean and finds applications in viscoelasticity and anomalous diffusion. At the boundary of physics and analysis, questions about the regularity (differentiability) of solutions to partial differential equations — most famously the Navier–Stokes existence and smoothness problem, a Clay Millennium Prize problem — remain unsolved: we do not know whether fluid-flow solutions stay smooth or develop infinite derivatives (singularities) in finite time.

Derivatives: From Slopes to Rates of Change

The Core Insight

Formal Statement

The Difference Quotient

The Derivative

Differentiability Implies Continuity

Why It Holds / Derivation

Deriving the Derivative of $f(x)=x^2$ from the Limit

Isaac Newton | 1643–1727

Building Intuition

Secant lines (slopes 5, 4) collapsing to the tangent line (slope 2) on the curve $y=x^2$ at $x=1$.

Worked Examples

Instantaneous Velocity of a Falling Ball

Slope of a Tangent and the Tangent-Line Equation

Common Misconceptions

Continuity Without Differentiability

Connections

Real-World Applications

Historical Development

Subtleties & Limits

See Also

Before this

Where this leads