Writing a Geometric Proof

Build one complete argument

Two segments made from matching end pieces and the same middle piece have equal lengths. Take four distinct points in the order $A,B,C,D$ on a line. Suppose $AB=CD=18$ and $BC=26$, with every length in centimeters. Segment addition gives:

Both lengths are $44$ centimeters. The reason is the matching end pieces and the shared middle length, not how the picture was drawn.

Now remove the particular measurements. The givens are the facts supplied: the four distinct points have that order on one line, and $AB=CD$. The goal is the statement to establish: $\overline{AC}\cong\overline{BD}$. A segment is the object; its length is the number, so $AC=BD$ will establish the segment congruence.

When you draw the general case, label the points in the stated order and put matching tick marks on $AB$ and $CD$. Do not mark $AC$ and $BD$ as congruent yet. That is the goal, not a given. The numerical labels belong to the particular example above; the following proof uses no particular lengths.

This is a paragraph proof, a complete argument in connected sentences. The numbers $18$ and $26$ were not needed in its steps, so it applies to any positive end and middle lengths satisfying the givens.

Reasons you can use

The paragraph proof uses familiar algebra to handle geometric lengths. For example, $18+26=26+18$ is the commutative property of addition: the same two lengths have the same sum in either order. A reason states why a step follows from the facts already available.

Equality concerns numbers. For lengths or angle measures $a,b,c$, the following properties are available from Algebra I; they are not new conjectures based on drawings.

For instance, if two whole lengths are both $44$ and you remove the same length $26$ from each, the remainders are equal:

$44-26=18.$

In a general proof, $AC=BD$ similarly permits $AC-BC=BD-BC$. To identify these differences with particular segments, you still need betweenness. Algebra alone does not tell you which piece lies inside which whole.

Segment congruence means equality of lengths. Angle congruence means equality of the selected angle measures. Because their measures obey equality properties, congruence of segments and congruence of angles are also reflexive, symmetric, and transitive. For example, if $\angle U\cong\angle V$ and $\angle V\cong\angle W$, their measures are equal in succession, so $\angle U\cong\angle W$. Use unambiguous angle labels when a diagram has several angles at one vertex.

The same proof in three formats

The complete paragraph argument came first. Two other formats display its structure without changing what must be justified.

A two-column proof pairs each statement with its reason. Here the givens are still four distinct collinear points in order $A,B,C,D$, with $AB=CD$.

A flow proof places statements with their reasons in boxes. Arrows show which facts support the next conclusion. The first two boxes supply the substitution. Commutativity then reorders that sum. A separate segment-addition statement supplies the same sum for $BD$. Symmetry and transitivity join the two equalities, after which equal lengths give segment congruence. The arrows belong to the proof, rather than adding geometric givens.

Paragraph form keeps the reasoning connected. Two-column form makes a missing reason easy to spot. Flow form makes dependencies easy to see. You need to recognize the same argument in all three; learning three different mathematical arguments is unnecessary.

Choose a route before writing

In a direct proof, start with the givens and follow justified steps toward the goal. The overlapping-segment proof is direct. Its goal concerns two whole lengths, and segment addition expresses both as sums of parts.

Planning can run backward. Ask what would establish the goal, then ask which givens could supply it. For a segment-congruence goal, an equality of lengths would suffice. For an equality of whole lengths, equal sums of their parts might suffice. Write the finished proof forward from facts already known; planning backward does not permit assuming the goal.

Choose a route by the information supplied. A midpoint or bisector given suggests equal parts. A betweenness condition suggests segment addition. A dividing ray inside a selected angle suggests angle addition. An equality with expressions suggests substitution and algebra. These are possible starting points, not guarantees that a particular tool will finish every proof.

A positive goal may follow directly from an available definition or equation. A goal saying that something does not exist, cannot happen, or occurs at most once often suggests contradiction: its negation supplies a possible object or extra case to test against the givens. These are planning cues; more than one approach can give a valid proof.

Proof by contradiction

Sometimes the useful starting point is the opposite of the goal. Suppose two circles in one plane have the same center $O$, one of radius $11$ and one of radius $19$. They cannot share a point. If a point $P$ were on both, the same distance $OP$ would have to be both $11$ and $19$. Those unequal numbers cannot both be that one length.

This method is proof by contradiction, a form of indirect proof. Keep the givens, temporarily assume the negation of the goal, and deduce an impossibility. The negation is the statement that the goal is false. Since that assumption cannot coexist with the givens, reject it and conclude the original goal.

For a goal saying that no common point exists, its negation says that at least one common point exists. You must rule out that whole possibility, not just show that a poorly chosen point fails.

Another contradiction may conflict with a uniqueness postulate rather than with arithmetic. The same rule applies: name the exact given, definition, or accepted foundation that the temporary assumption violates. Failure to find a direct proof is not itself a contradiction.

Lesson 1.1 established that two distinct lines have at most one common point. Here is a self-contained proof by contradiction: suppose two distinct lines share distinct points $P$ and $Q$. The accepted unique-line postulate says that exactly one line passes through those two distinct points, so both proposed lines would be that same line. This contradicts their given distinctness. Therefore two distinct lines cannot share two distinct points.

Repair assumptions and track foundations

An argument may use correct algebra after an unjustified geometric step. Suppose you know only $AB=BC$ and claim that $B$ is the midpoint of $\overline{AC}$. The lengths match, but the midpoint definition also needs $B$ to lie on $\overline{AC}$. A bent configuration shows what the equality alone permits.

Repair the proof by citing the fact that $B$ lies on $\overline{AC}$, if the problem gives it, or by proving that location from other givens. If neither is available, the midpoint conclusion is not established. Drawing a straight configuration does not repair a missing condition in the original problem.

The status reference from 1.5 now gains the results established here. Each row records what supports the claim, rather than turning every convincing drawing into a theorem.

Common mistakes