JOVANA
Explore Library Glossary Getting Started Three Levels Fields How it works Mission
Join the mission
All guides

If-Then: How a Proof Actually Works

You already know the undefined words, the postulates, and the angles you can measure. Now we wire them together. A proof is just an honest chain of if-then steps that drags a fact you doubt back to facts you have already accepted.

The shape of every claim: if-then

In the last three guides you collected the raw materials: the undefined words point, line, and plane; the postulates and common notions we agreed to accept without proof; and the angles, segments and rays you learned to measure. A proof is what you build out of those materials. But before you can build, you need to see the one shape that almost every geometric claim secretly has.

That shape is the conditional statement, written "if P, then Q". The part after "if" is the hypothesis (P) — what you are allowed to assume. The part after "then" is the conclusion (Q) — what you must reach. "If two angles form a linear pair, then they are supplementary" hands you P (a linear pair) for free and asks you to earn Q (their measures add to 180 degrees). A conditional statement is a one-way promise: it says nothing at all about what happens when P is false.

The same words, rearranged — and the trap

Take any conditional "if P, then Q" and you can shuffle P and Q into three relatives. The converse swaps them: "if Q, then P". The inverse negates both: "if not P, then not Q". The contrapositive does both — swap and negate: "if not Q, then not P". These four are not interchangeable, and confusing a statement with its converse is the single most common beginner error in all of geometry.

Hold these four side by side. The original is "if P, then Q". The converse is "if Q, then P", and it is not guaranteed to follow. The inverse is "if not P, then not Q", also not guaranteed. Only the contrapositive, "if not Q, then not P", always shares the original's truth value — when one is true, so is the other; when one is false, so is the other.

Here is the trap made concrete. "If a number is a multiple of 4, then it is even" is true. Its converse, "if a number is even, then it is a multiple of 4", is false — 6 is even but not a multiple of 4. The contrapositive, "if a number is not even, then it is not a multiple of 4", is true, and it must be, because a statement and its contrapositive always rise or fall together. That fact is not a coincidence; it is the engine behind a whole proof style we will meet in the final guide.

When the converse also happens to be true, P and Q are equivalent, and we fuse the two arrows into one. "M is the midpoint of AB if and only if |AM| = |MB| and M is between A and B" is a biconditional — "if and only if" in both directions. Every careful geometric definition is secretly a biconditional, which is exactly why a definition can be used forwards or backwards inside a proof, a freedom an ordinary theorem does not always grant.

One counterexample is fatal

Geometry is brutally asymmetric about evidence. To kill an "if P, then Q" claim, you do not need an argument at all — you need a single case where P holds but Q fails. That single case is a counterexample, and one is enough. The number 6 above demolished the false converse all by itself. No amount of examples that fit the pattern can ever rescue a claim that has even one counterexample.

But notice the flip side, because it is the whole reason proof exists. To establish "if P, then Q", finding a thousand confirming examples proves nothing — there could always be a counterexample you have not drawn yet. Measuring three pairs of vertical angles and seeing them equal is encouraging, not conclusive. Confirmation never closes the gap; only a proof, an argument that covers every possible case at once, does. This asymmetry — one counterexample disproves, but no finite pile of examples proves — is precisely why the axiomatic method was invented.

What a deductive step is allowed to be

A proof is a finite chain of statements running from the hypothesis to the conclusion, where every link is justified by something you have already earned the right to use. And there are only four kinds of justification you may ever cite, which is liberating once you see it — the menu is short.

  1. A given — something the problem hands you as part of the hypothesis P. You may simply write it down.
  2. A definition — and because every definition is a biconditional, you may use it forwards or backwards (e.g. "midpoint" lets you conclude |AM| = |MB|, and equal halves let you conclude "midpoint").
  3. A postulate or common notion — a rule accepted without proof, such as the angle addition postulate or the segment addition postulate you met last guide.
  4. A theorem already proven — once "vertical angles are equal" is proved, it joins your toolbox and any later proof may cite it by name.

This is exactly the definition-versus-theorem distinction from guide 2 doing real work. A definition is a free naming move you may invoke anytime; a theorem is a claim that had to be earned by proof before it could be reused. What you may never do is justify a step by "it looks true in my drawing". The picture is a guide for your hand, never a reason for your conclusion — the diagram might be drawn slightly wrong, and a proof must hold even for the cases you cannot see.

Why two columns — and a tiny proof in full

Now the famous format makes sense. A two-column proof puts each statement in the left column and, directly beside it, its justification in the right. The left column is the chain of facts; the right column answers, for every single line, "by what right?" Nothing may appear on the left that the right cannot defend with a given, a definition, a postulate, or a prior theorem. The layout is not bureaucracy — it is a promise that no step has been smuggled in unjustified.

Given:  ray BD lies between ray BA and ray BC, and m(angle ABD) = m(angle DBC).
Prove:  ray BD bisects angle ABC.

  Statement                              Reason
  1. m(angle ABD) = m(angle DBC)         1. Given
  2. m(angle ABD) + m(angle DBC)         2. Angle addition postulate
        = m(angle ABC)                        (ray BD is between the others)
  3. m(angle ABD) + m(angle ABD)         3. Substitute line 1 into line 2
        = m(angle ABC)
  4. 2 * m(angle ABD) = m(angle ABC)     4. Combine like terms
  5. ray BD bisects angle ABC            5. Definition of angle bisector
Every left-column fact is defended on the right by one of the four allowed reasons.

Read that proof slowly and watch the machine run. Line 1 is a given. Line 2 cites a postulate. Lines 3 and 4 are pure algebra — and notice that the right to do algebra on angle measures came from guide 3, where the measures became real numbers you can add and substitute. Line 5 reads the definition of angle bisector backwards, which a biconditional permits. The doubt you started with — does BD really bisect the angle? — has been dragged, link by honest link, back to facts you already accepted.

The two-column form is a beginner's scaffold, not the only honest format. The very same argument written as flowing prose is a paragraph proof, and as you grow you will lean on the words and drop the grid. But the discipline underneath never changes: every claim earns its place, or it does not belong. In the final guide of this rung you will hold the pen yourself and write one from a blank page.