Functions

⚡ Key Takeaways

A function is a rule that assigns to each input exactly one output. Written $f: A \to B$ , it maps every element of a domain $A$ to a single element of a codomain $B$ . Functions appear everywhere in mathematics, science, economics, and computing. The modern definition crystallized in the 19th century through the work of Dirichlet, Riemann, and others.

💡 The Core Insight

The defining feature of a function is single-valuedness: every input has one and only one output. The same output may come from many inputs, and not every element of the codomain must be used — but no input may point to two outputs. This is what separates a function from a general relation.

DefinitionFunction

A function $f$ from a set $A$ (the domain) to a set $B$ (the codomain) is a rule assigning to each $x \in A$ exactly one element $f(x) \in B$ . We write:

1

\quad f: A \to B, \qquad x \mapsto f(x)

Derivation hint: This is notation, not derived — read " $f$ maps $A$ into $B$ , sending $x$ to $f(x)$ ." The defining condition (single-valuedness) is:

2

\quad x_1 = x_2 \;\implies\; f(x_1) = f(x_2)

Derivation hint: Equal inputs must produce equal outputs; this is the contrapositive of "two outputs from one input." Here:

$x$ is the independent variable (input, argument).
$f(x)$ is the value or image of $x$ (output).
The range is the set of all actual outputs: $\{f(x) : x \in A\} \subseteq B$ .

DefinitionFunction as a Set of Ordered Pairs

Formally, a function is a set of ordered pairs $f \subseteq A \times B$ such that:

3

\quad \forall x \in A,\ \exists!\, y \in B : (x,y) \in f

Derivation hint: " $\exists!$ " means "there exists exactly one." Each $x$ appears in precisely one pair. This set-theoretic version makes the definition rigorous without referring to any formula or process.

TheoremVertical Line Test

A curve in the Cartesian plane is the graph of a function $y = f(x)$ if and only if every vertical line $x = c$ intersects the curve in at most one point.

4

\quad \text{graph}(f) = \{(x, f(x)) : x \in A\}

Derivation hint: Two intersection points on one vertical line would mean one input $x = c$ gives two outputs, violating $[2]$ .

ProofWhy the Vertical Line Test Works

We prove the Vertical Line Test follows directly from the definition. 1. Suppose $C$ is a curve that is the graph of a function $f$ . By definition $[3]$ , each $x = c$ in the domain has exactly one value $f(c)$ , so the only point of $C$ with first coordinate $c$ is $(c, f(c))$ . 2. Therefore the vertical line $x = c$ meets $C$ in exactly one point (or none, if $c \notin A$ ) — never more than one. 3. Conversely, suppose every vertical line meets $C$ in at most one point. Define $A$ as the set of $x$ -values appearing in $C$ . For each $x \in A$ , there is exactly one $y$ with $(x,y) \in C$ ; call it $f(x)$ . 4. This assignment satisfies $[3]$ , so $C$ is the graph of the function $f$ . $\blacksquare$

∎

mathematician · Peter Gustav Lejeune Dirichlet | 1805–1859

Dirichlet gave one of the first fully modern definitions of a function in 1837, freeing the concept from any requirement of a formula. He proposed that $y$ is a function of $x$ if to each value of $x$ there corresponds a unique value of $y$ — by any rule whatsoever. His famous example, the Dirichlet function (equal to 1 on rationals and 0 on irrationals), showed that functions need not be drawable or computable by a single expression. This abstraction reshaped analysis.

🧠 Intuition

Think of a function as a machine: you feed in an input, the machine processes it, and out comes exactly one output. Press the same button twice and you get the same result — that is single-valuedness. Different buttons may give the same result (that's allowed), but one button can never give two different results. Other useful mental pictures:

A mapping of arrows from a left-hand bag of inputs to a right-hand bag of outputs: every left dot has exactly one arrow leaving it.
A table of values: each row's input column entry is unique to that row's output.
A dependency: "distance depends on time," "price depends on quantity" — the output is determined by the input.

Figure

A function as a mapping between sets: every input arrow leaves exactly one dot

ExampleEvaluating and Finding the Domain

Given: $f(x) = \dfrac{\sqrt{x-2}}{x-5}$ .

Find: (a) $f(6)$ , (b) the natural domain.

Steps:

(a) Substitute $x = 6$ :

5

\quad f(6) = \frac{\sqrt{6-2}}{6-5} = \frac{\sqrt{4}}{1} = \frac{2}{1} = 2

Derivation hint: Compute inside the radical first, then the denominator, then divide.

(b) Two restrictions: the radicand must be $\ge 0$ , so $x - 2 \ge 0 \Rightarrow x \ge 2$ . The denominator must be nonzero, so $x \ne 5$ .

Combining:

6

\quad \text{Domain} = [2, 5) \cup (5, \infty)

Derivation hint: Intersect $x \ge 2$ with $x \ne 5$ ; the excluded point 5 splits the interval.

Answer: $\boxed{f(6) = 2, \quad \text{Domain} = [2,5)\cup(5,\infty)}$

ExampleComposition of Functions

Given: $f(x) = 2x + 3$ and $g(x) = x^2 - 1$ .

Find: $(f \circ g)(x)$ and $(g \circ f)(2)$ .

Steps:

The composition $(f \circ g)(x) = f(g(x))$ . Substitute $g(x)$ into $f$ :

7

\quad (f \circ g)(x) = f(x^2 - 1) = 2(x^2 - 1) + 3 = 2x^2 + 1

Derivation hint: Replace the argument of $f$ with the whole expression $g(x)$ , then simplify.

For $(g \circ f)(2)$ , first compute $f(2) = 2(2)+3 = 7$ , then $g(7) = 7^2 - 1 = 48$ :

8

\quad (g \circ f)(2) = g(f(2)) = g(7) = 48

Derivation hint: Work from the inside out — apply $f$ first, then $g$ .

Answer: $\boxed{(f\circ g)(x) = 2x^2 + 1, \quad (g\circ f)(2) = 48}$ Note that $f \circ g \ne g \circ f$ in general — composition is not commutative.

⚠️ Common mistake

"A function must have a formula." False. A function is any single-valued assignment. It can be defined by a table, a graph, a piecewise rule, or words like " $f(n) =$ the $n$ -th prime." The formula is just one convenient way of specifying the rule.

" $f(x)$ means $f$ times $x$ ." No — $f(x)$ denotes the output of $f$ at input $x$ , not multiplication. Confusing this with the distributive law leads to errors like writing $f(a+b) = f(a) + f(b)$ , which is generally false.

CounterexampleA Relation That Is Not a Function

Consider the unit circle $x^2 + y^2 = 1$ . Solving for $y$ gives $y = \pm\sqrt{1 - x^2}$ . At $x = 0$ we get two outputs, $y = 1$ and $y = -1$ . Since one input yields two outputs, the circle is not a function of $x$ — it fails the Vertical Line Test. (We can split it into two functions: the upper and lower semicircles.)

CounterexampleLinearity Fails for Most Functions

Take $f(x) = x^2$ . Then $f(1+2) = f(3) = 9$ , but $f(1) + f(2) = 1 + 4 = 5$ . Since $9 \ne 5$ , we see $f(a+b) \ne f(a) + f(b)$ . Only special functions (linear ones, $f(x) = mx$ ) satisfy that additive property.

connection

Functions are the connective tissue of mathematics. They generalize the idea of an Equation by emphasizing dependence rather than a single solution set. Special families — Linear Functions, Quadratic Functions, Polynomial Functions, Exponential Functions, and Trigonometric Functions — are studied individually but all obey the single-valuedness rule. The study of how outputs change with inputs leads directly to Derivatives in calculus, while accumulation leads to Integrals. Functions that "undo" each other are Inverse Functions, and combining functions yields Function Composition. The graphical view connects to Coordinate Geometry.

🔬 Where this appears

⏳ History

- c. 1350 — Nicole Oresme graphs varying quantities, an early intuition of functional dependence.

- 1637 — René Descartes links algebra and geometry in La Géométrie, enabling curves as relations between variables.

- 1673 — Gottfried Leibniz coins the term function (Latin functio) for quantities related to a curve.

- 1748 — Leonhard Euler, in Introductio in analysin infinitorum, defines a function as an analytic expression and introduces the notation

f(x)

.

- 1822 — Joseph Fourier represents arbitrary functions as trigonometric series, stretching the notion of "function."

- 1837 — Dirichlet gives the modern, formula-free definition: a unique output for each input by any rule.

- 1854 — Bernhard Riemann refines the concept while developing integration theory.

- 1872 — Karl Weierstrass exhibits a continuous-but-nowhere-differentiable function, challenging intuitions.

- 1900s — Set theory (Cantor, Zermelo) recasts functions as sets of ordered pairs, giving the rigorous modern foundation.

open-problem

The modern set-theoretic definition treats a function as a completed set of pairs, but this raises deep questions explored in advanced mathematics. The Axiom of Choice is needed to guarantee certain functions exist (e.g., choosing one element from each set in an infinite collection), and its status was debated for decades. In computability theory, not every function $\mathbb{N} \to \mathbb{N}$ can be computed by an algorithm — there are uncountably many functions but only countably many programs, so most functions are non-computable. Whether certain specific functions (e.g., the Busy Beaver function) are even well-defined for large inputs touches on the limits of what mathematics can decide.

Formal Statement

Why It Holds / Derivation

Building Intuition

Worked Examples

Common Misconceptions

Connections

Real-World Applications

Historical Development

Subtleties & Limits

See Also