1. The undefined terms
If you sit down to write a dictionary, eventually you hit a problem: every definition is made of words, and every one of those words needs its own definition. Either the chain goes on forever, or it loops back on itself, or it has to stop somewhere. Mathematics chooses to stop — deliberately, in the open — and the words it refuses to define are called primitive terms.
In Euclidean geometry there are three:
A location with no size — no length, no width, no thickness. We draw it as a dot only because we have to draw it as something; the dot is not the point, it's a stand-in. Points are labeled with capital letters: $A$, $B$, $P$.
A straight, one-dimensional object that extends without end in both directions. It has no thickness and no curvature. A line through points $A$ and $B$ is written $\overleftrightarrow{AB}$, or sometimes named with a single lowercase letter like $\ell$.
A perfectly flat, two-dimensional surface extending without end in every direction within itself. Like the line, it has no thickness. We name a plane with a single capital letter, often written in script ($\mathcal{P}$), or by three of its non-collinear points.
Notice that each "definition" above is really a description — it leans on intuitive words like "flat" and "straight" that we haven't pinned down. That's on purpose. You can't define everything in terms of something simpler, so we stop at three primitives, agree on what they intuitively are, and build from there.
2. Postulates and axioms
If we don't define the primitives, how do we say anything precise about them? By stating rules they obey — assumed up front, accepted without proof. These rules are called postulates (or, in more modern language, axioms). They're the things you agree to before the game starts.
A postulate is not a theorem we haven't proved yet. It's a starting point we have chosen — and everything that follows is true relative to that starting point.
Here are the postulates that get the most use in this topic:
| Postulate | What it says |
|---|---|
| Two-point | Through any two distinct points there is exactly one line. |
| Three-point | Through any three non-collinear points there is exactly one plane. |
| Line containment | If two points of a line lie in a plane, the entire line lies in that plane. |
| Plane intersection | If two distinct planes intersect, their intersection is a line. |
| Line intersection | If two distinct lines intersect, their intersection is exactly one point. |
The two-point postulate is so familiar it can feel like an observation — but it is genuinely an assumption. On a sphere, "lines" are great circles, and through two nearly-antipodal points there is still exactly one shortest great-circle arc; through two exactly-antipodal points, infinitely many great circles pass. Different starting assumptions, different geometry.
Two points always determine a unique line — but infinitely many planes contain that line, each rotated around it like the pages of a book around the spine. You need a third point, off the line, to lock the page into one specific position. That's why the three-point postulate quietly demands the points be non-collinear.
3. Collinear and coplanar
Two pieces of vocabulary that you'll use constantly:
Points that all lie on a single line. Any two points are trivially collinear (there is always a line through them); the word starts earning its keep at three or more points.
Points (or lines) that all lie in a single plane. Any three points are trivially coplanar — collinear or not, a plane can always be found that contains them. The word becomes interesting at four or more points, where it stops being automatic.
So the natural threshold is: collinearity is a real condition starting at three points; coplanarity is a real condition starting at four points. Below those thresholds the property is free; above them you have to check.
4. Segments, rays, and lines
Once you have two points $A$ and $B$, there are three different things you might mean by "the straight path between them" — and geometry uses three different notations to keep them straight.
| Object | Notation | Endpoints | Extent |
|---|---|---|---|
| Segment | $\overline{AB}$ | Two: $A$ and $B$ | Finite — stops at $A$ and $B$. |
| Ray | $\overrightarrow{AB}$ | One: $A$ (the endpoint) | Starts at $A$, passes through $B$, continues forever. |
| Line | $\overleftrightarrow{AB}$ | None | Extends without end in both directions. |
The order of the letters carries meaning only for the ray: $\overrightarrow{AB}$ starts at $A$ and goes through $B$, while $\overrightarrow{BA}$ starts at $B$ and goes through $A$ — these are different rays, even though they share infinitely many points. For segments and lines, $\overline{AB} = \overline{BA}$ and $\overleftrightarrow{AB} = \overleftrightarrow{BA}$.
A useful related word: length. Only the segment has a length — written $AB$ (no bar). It's a number, not a geometric object. Rays and lines are infinite, so length doesn't apply.
5. Intersections
"Intersection" means the set of points two objects share. With our three primitives, the possibilities are tidy:
| Objects | Generic intersection | Degenerate cases |
|---|---|---|
| Line ∩ Line | A single point | Empty (parallel) — or the whole line (same line) |
| Plane ∩ Plane | A line | Empty (parallel) — or the whole plane (same plane) |
| Line ∩ Plane | A single point | Empty (parallel) — or the whole line (line lies in plane) |
Notice the pattern: when two objects of the same kind intersect, the intersection is usually one dimension lower. Two planes (2D each) cross in a line (1D). Two lines (1D each) cross in a point (0D). The degenerate cases are when the objects are either too far apart to meet or so aligned that they overlap entirely.
"Parallel" is more than just "doesn't intersect." Two lines in 3-D space can fail to intersect without being parallel — they're called skew. Parallel lines are coplanar and don't meet; skew lines aren't coplanar at all.
6. Why proof matters here
Around 300 BCE, Euclid's Elements did something no earlier mathematical text had attempted at scale: he stated a handful of definitions and postulates up front, then derived everything else — hundreds of theorems — by careful logical steps from those starting points. Nothing was claimed to be true because it looked true. Each result had to be earned.
This is why geometry, more than arithmetic, is where most students first meet proof. The objects are simple enough that intuition gives you a strong guess about every theorem, which makes it all the more striking when the proof reveals exactly why the guess is right — or, occasionally, that the guess was subtly wrong.
Three things make this topic the right place to plant those seeds:
- The objects are minimal. Point, line, plane — three things, with a short list of rules. There's nowhere to hide complexity.
- The postulates are visible. You can list them on one page. Every later claim is traceable back to that list.
- The reasoning is portable. The habit of "what am I assuming, and what does it force?" transfers to algebra, number theory, analysis, and beyond.
You're going to spend the rest of geometry watching small assumptions push very far. That's the engine — and you've just seen its starter motor.
7. Common pitfalls
They don't — they define a line. You need three non-collinear points for a plane. Saying "two points define a plane" is one of the most common slips and it breaks the rest of the chapter.
The bar matters. A flat bar means segment (finite, both ends). One arrow means ray (one end, one infinite direction). Two arrows means line (no ends). Mixing them up will silently change what your sentence claims.
A diagram can suggest a result, but it can never prove one. Two lines that look parallel on paper might meet a thousand units off the page; three points that look collinear might miss by a hair too small to draw. Always justify with postulates and earlier theorems, not pixels.
The two-point postulate feels like a fact about the universe — but on a sphere it fails (antipodal points have infinitely many "lines" through them). Change a postulate and you get a different, still-consistent geometry. Postulates are assumed, not discovered.
8. Worked examples
Each problem here is about identification, notation, or vocabulary — the meat of this topic. Try answering before opening the solution.
Example 1 · Name the object: $\overrightarrow{PQ}$
The single arrow over $PQ$ means this is a ray.
- It has one endpoint: $P$.
- It passes through $Q$ and continues forever past $Q$.
- It is not the same as $\overrightarrow{QP}$, which would start at $Q$ and go the other way.
Example 2 · True or false: any three points are coplanar.
True.
If the three points are non-collinear, the three-point postulate gives exactly one plane through them. If the three points are collinear, they all lie on a single line — and many planes contain that line, so there is still at least one. Either way, you can always find a plane containing all three.
This is why coplanarity is only an interesting condition for four or more points.
Example 3 · Two distinct planes share point $X$. What else must they share?
By the plane-intersection postulate, two distinct planes that intersect do so in a line. So if they share point $X$, they must in fact share an entire line through $X$ — and infinitely many other points along that line.
It's impossible for two distinct planes to share only a single point. They either share a line or they don't intersect at all.
Example 4 · True or false: $\overline{AB} = \overline{BA}$ and $\overrightarrow{AB} = \overrightarrow{BA}$.
The first is true; the second is false.
A segment is just the set of points between its two endpoints — it doesn't care which endpoint you name first. So $\overline{AB}$ and $\overline{BA}$ describe the exact same set of points.
A ray, by contrast, has a direction baked in. $\overrightarrow{AB}$ starts at $A$ and heads toward $B$; $\overrightarrow{BA}$ starts at $B$ and heads toward $A$. They share the segment $\overline{AB}$ but extend in opposite directions, so they are different rays.
Example 5 · How many lines are determined by four points, no three of which are collinear?
Each pair of points determines exactly one line (two-point postulate). With four points, the number of pairs is
$$ \binom{4}{2} = \frac{4 \cdot 3}{2} = 6. $$So there are 6 distinct lines.
The "no three collinear" condition is what keeps the count honest. If three of the four points were collinear, those three pairs would all name the same line, and the count would drop to $6 - 3 + 1 = 4$.