Visual walkthrough — Vertical line test for functions

Step 1 — What is a point on the plane?

WHAT. Before any test, we need the stage: a flat sheet with two number lines crossing at right angles. The horizontal one we call the ==$x$-axis== (it holds the input). The vertical one is the ==$y$-axis== (it holds the output).

A single dot is written $(x,y)$. Read it as an address:

$(\underset{\text{how far right/left}}{\underbrace{x}},\underset{\text{how far up/down}}{\underbrace{y}})$

• The first slot $x$ = go right (positive) or left (negative) along the flat axis.
• The second slot $y$ = go up (positive) or down (negative).

WHY. Every graph we test is just a crowd of these dots. If we cannot read one dot, we cannot read a graph. So we start here.

PICTURE. The red dot sits at $(3,2)$: three steps right, two steps up. Notice the two dashed guide-lines — one vertical from the $x$-axis, one horizontal from the $y$-axis — meeting exactly at the dot.

Step 2 — What does "input → output" mean as a rule?

WHAT. A function is a machine with one law: feed it one $x$, it hands back exactly one $y$. We write that handing-back as $y=f(x)$, read "$y$ equals $f$ of $x$":

$\underset{\text{what comes out}}{\underbrace{y}}=\underset{\text{the machine}}{\underbrace{f}}\left(\underset{\text{what we put in}}{\underbrace{x}}\right)$

The word exactly is the whole game. Not "at least one." Not "some." Exactly one.

WHY. If pressing button $x=3$ could drop either $2$ or $5$, you'd never know what the machine will do. A rule you cannot predict is not a rule. So we demand: one $x$, one $y$. (This is the formal statement in 2.1.01-Definition-of-a-function.)

PICTURE. Left machine: input $3$ → single output $2$. Good. Right machine: input $3$ → two outputs $2$ and $5$ at once. Broken — the arrow splits.

Step 3 — Stacking dots that share the same $x$

WHAT. Take that broken case and draw both dots on our plane. Input $x=3$ produced $y=2$ and $y=5$, so we plot $(3,2)$ and $(3,5)$.

Look at their addresses side by side:

$(\underset{\text{same}}{\underbrace{3}},2)(\underset{\text{same}}{\underbrace{3}},5)$

The first slots are identical; only the second slots differ.

WHY. Two dots with the same $x$ but different $y$ are the fingerprint of a broken machine. We want a way to see that fingerprint instantly, without reading numbers. That is what the next step invents.

PICTURE. Two coral dots sit stacked vertically above the same spot $x=3$ on the axis. They line up perfectly one above the other — because they share the $x$-address.

Step 4 — Inventing the vertical line

WHAT. A vertical line at $x=a$ is the set of all dots whose first slot equals $a$:

$\{(a,y):y\text{ is any number}\}$

Term by term:

• $a$ is frozen — every dot on this line has the same $x$.
• $y$ is free — it slides through every possible height.

So one vertical line is a probe that asks a single question: "At the input $a$, which outputs does the graph offer?"

WHY. Why vertical and not slanted or horizontal? Because we are testing the input. Freezing $x=a$ and sweeping every $y$ is exactly the act of pressing one button and watching everything that falls out. A horizontal line would freeze the output instead — a different question (that's 2.2.05-Horizontal-line-test-for-injective-functions).

PICTURE. The lavender vertical line stands at $x=3$. An arrow slides up and down it, showing $y$ ranging freely while $x$ stays pinned at $3$.

Step 5 — Counting crossings = counting outputs

WHAT. Now drop the vertical line onto a graph and count how many dots it touches. Each touch is one point $(a,y)$ that lives on both the line and the graph — i.e. one output the machine gives for input $a$.

$\text{(crossings at }x=a\text{)}=\text{(number of outputs }y\text{ for input }a\text{)}$

WHY. This is the bridge. The line already collects "all dots with $x=a$." The graph already stores "all outputs of the machine." Where they overlap is precisely "outputs at $a$." Counting overlaps is counting outputs — no algebra needed.

PICTURE. Same lavender line at $x=3$, now laid over the broken graph. It punches through both coral dots — two crossings, glowing.

Step 6 — Reading the verdict

WHAT. Translate the count straight into a ruling:

A relation is a function when every vertical line gives $0$ or $1$ crossing — never more.

WHY. Look back at Step 2: the law was exactly one output. "$0$ or $1$" says "at most one," which allows the machine to simply be silent at some inputs (undefined there) while never being greedy (two answers). Silence is allowed; greed is not.

PICTURE. Three vertical lines side by side over one graph: one misses entirely (0), one grazes a single dot (1, green tick), one skewers two dots (2, red cross).

Step 7 — The edge cases we must not skip

WHAT. Four situations trip people up. We show each so no scenario surprises you.

1. A gap in the graph (like $y=\frac{1}{x}$ at $x=0$): the line finds zero crossings there → allowed.
2. A wavy graph (like $y=\sin x$): it wiggles a lot, but every vertical line still hits it once → passes.
3. A vertical line drawn ON the graph (like the relation $x=3$ itself): the line overlaps infinitely → fails hard.
4. A single stacked pair (our circle at just one $x$): even one bad vertical line is enough to fail the whole graph.

WHY. The rule says every vertical line must behave. A graph can look tame for a thousand lines and still fail on the one line you didn't check. And it can look wild (waves, gaps) yet pass everywhere. Shape is a distraction; crossing-count is the only judge.

PICTURE. Four mini-panels: (a) gap → 0, (b) wave → 1 everywhere, (c) the line $x=3$ → infinite, (d) circle → 2 at one spot.

The one-picture summary

Everything above, compressed: the vertical line is a probe that freezes $x$, sweeps $y$, and counts outputs by counting crossings. Zero or one → function. Two or more → not.

Connections

• 2.1.01-Definition-of-a-function — the "exactly one output" law this test draws
• 2.2.01-Graph-of-a-function — what the crowd of dots means
• 2.2.05-Horizontal-line-test-for-injective-functions — the mirror-image test on outputs
• 2.3.02-Domain-and-rangefrom-graphs — reading where a function is defined
• 3.1.01-Inverse-functions — needs both line tests to pass

#flashcards/maths