In short

A mathematical proof is a chain of logical steps that starts from accepted truths (axioms, definitions, or previously proved results) and ends at the statement you want to establish. A direct proof is the simplest kind: you assume the hypothesis, apply definitions and known facts, and walk forward until you reach the conclusion. Every step must follow from the previous ones — no gaps, no leaps of faith, no appeals to intuition.

You already know that 2 + 3 = 5. You can check it by counting on your fingers, by adding on paper, or by punching it into a calculator. But how do you know that the sum of any two even numbers is always even — not for one specific pair, but for every pair of even numbers that could ever exist? You cannot check infinitely many cases by hand. You need a different kind of argument, one that covers all cases at once.

That argument is called a proof. A proof is the tool that takes a mathematical statement from "probably true" to "certainly true." It is the reason mathematics is the only subject where a result, once proved, is settled forever — not revised, not overturned by new data, not subject to opinion. If the proof is correct, the conclusion holds.

This article covers the anatomy of a proof and the most natural proof technique: the direct proof. Direct proof is the first technique you should reach for, and it handles a surprisingly large fraction of the results you will meet in school mathematics and beyond. The other techniques — proof by contradiction and proof by contrapositive — exist for situations where the direct path is blocked, but the direct path is the one to try first.

What a proof is made of

A proof has three ingredients:

  1. Hypotheses — the starting conditions you are allowed to assume. These come from the statement you are trying to prove. "If n is even, then n^2 is even" has one hypothesis: n is even.

  2. Logical steps — each step uses a definition, an axiom, an algebraic rule, or a previously proved theorem to derive something new from what you already have. Every step must be justified; the justification is what separates a proof from a guess.

  3. Conclusion — the final statement you wanted to reach. When you get there, the proof is done.

The whole structure is an implication: if the hypotheses hold, then the conclusion follows. A proof is a bridge from the "if" side to the "then" side, built one plank at a time.

Structure of a direct proof as a bridge from hypothesis to conclusionA diagram showing a box labelled Hypothesis on the left connected by a series of three arrows labelled Step 1, Step 2, and Step 3 to a box labelled Conclusion on the right. Each arrow represents one logical step in the proof.HypothesisStep 1···Step 2···Step 3Conclusion
The skeleton of a direct proof. You start at the hypothesis, take a sequence of justified steps, and arrive at the conclusion. Each arrow is one logical move — a definition applied, an algebraic manipulation performed, or a known theorem invoked. The proof is the entire chain.

Definitions: the raw material of proofs

Before you can prove anything about even numbers, you need to know precisely what "even" means. Not a vague sense — a formal definition that you can use as a tool.

Even and odd integers

An integer n is even if there exists an integer k such that n = 2k. An integer n is odd if there exists an integer k such that n = 2k + 1.

This definition is the key that unlocks every proof about even and odd numbers. When you assume a number is even, you are allowed to write it as 2k for some integer k. When you want to show a number is even, you must exhibit it in the form 2m for some integer m. The definition is both the hypothesis-unpacker and the conclusion-checker.

Similarly, for divisibility:

Divisibility

An integer a divides an integer b — written a \mid b — if there exists an integer q such that b = aq. When a \mid b, you say "a is a divisor of b" or "b is a multiple of a."

So 3 \mid 12 because 12 = 3 \times 4 (take q = 4). And 3 \nmid 14 because there is no integer q with 14 = 3q. Notice the direction: "a divides b" means b is the bigger number that is a clean multiple of a, not the other way around. The notation a \mid b reads left-to-right as "a divides b."

These definitions are your building blocks. A direct proof works by unpacking hypotheses through definitions, performing algebraic manipulations, and then re-packing the result into the form the conclusion demands.

The direct proof technique

Here is the recipe:

To prove "If P, then Q" directly:

  1. Assume P is true.
  2. Use definitions and known results to manipulate the assumption.
  3. Arrive at Q.

That is the whole method. The difficulty is never in the recipe — it is in finding the right algebraic path from P to Q. But the structure is always the same: assume the "if" part, work forward, reach the "then" part.

Here is the simplest possible direct proof, written out in full.

Claim. If n is an even integer, then n^2 is even.

Proof. Assume n is even. By definition, n = 2k for some integer k. Then:

n^2 = (2k)^2 = 4k^2 = 2(2k^2)

The expression 2k^2 is an integer (since k is an integer, k^2 is an integer, and 2k^2 is an integer). So n^2 = 2m where m = 2k^2 is an integer. By definition, n^2 is even. \square

Every line of that proof does one thing: it either applies the definition of "even" or performs an algebraic step. There are no leaps, no skipped steps, and no appeals to "you can see that." The \square symbol at the end (called a tombstone or Halmos symbol) is the standard way to mark that the proof is complete.

Flow diagram of the proof that n squared is even when n is evenA vertical chain of four boxes connected by downward arrows. The first box says assume n is even. The second box says n equals 2k by definition. The third box says n squared equals 4k squared equals 2 times 2k squared. The fourth box, highlighted in red, says n squared is even by definition.Assume n is evenhypothesisn = 2k for some integer kdefinitionn² = (2k)² = 4k² = 2·(2k²)algebran² is even (= 2m, m = 2k²)conclusion
The proof that $n^2$ is even whenever $n$ is even, laid out as a flow. Each box is one logical step; the right-hand labels name what kind of move each step is. The entire chain runs downward from hypothesis to conclusion with no gaps.

Proving algebraic identities

Direct proof is the natural technique for algebraic identities — statements that two expressions are always equal. The method is pure calculation: start with one side, apply algebraic rules, and transform it into the other side.

Claim. For all real numbers a and b,

(a + b)^2 = a^2 + 2ab + b^2

Proof. Start with the left side and expand using the distributive law.

(a + b)^2 = (a + b)(a + b)

Distribute the first factor across the second:

= a(a + b) + b(a + b)

Distribute again:

= a^2 + ab + ba + b^2

Since multiplication of real numbers is commutative, ba = ab:

= a^2 + ab + ab + b^2 = a^2 + 2ab + b^2

This matches the right side. \square

The identity (a + b)^2 = a^2 + 2ab + b^2 has a geometric interpretation too. Picture a square with side length a + b. Its total area is (a + b)^2. Now cut the square into four pieces with a horizontal and a vertical line at distance a from one corner. You get four rectangles: one a \times a square (area a^2), one b \times b square (area b^2), and two a \times b rectangles (each with area ab, total 2ab). The four pieces tile the big square exactly, so their areas add up to (a + b)^2.

Geometric proof that a plus b squared equals a squared plus 2ab plus b squaredA large square with side a plus b is divided into four regions by a horizontal and a vertical line. The top-left region is a square of side a with area a squared. The bottom-right region is a square of side b with area b squared. The top-right and bottom-left regions are rectangles of sides a and b, each with area ab. Together the four regions tile the large square.a²ababb²abab(a + b)² = a² + 2ab + b²
The identity $(a + b)^2 = a^2 + 2ab + b^2$ as a tiled square. The big square has side $a + b$. The four pieces inside it — two squares and two rectangles — account for every square unit of area. The algebraic proof and the geometric proof say the same thing in different languages.

Proving divisibility properties

Direct proof is also the right tool for divisibility claims. The pattern is the same: unpack the hypothesis using the definition of divisibility, perform algebra, and repack the result.

Claim. If a \mid b and a \mid c, then a \mid (b + c).

Proof. Assume a \mid b and a \mid c. By definition, b = aq for some integer q, and c = ar for some integer r. Then:

b + c = aq + ar = a(q + r)

Since q and r are integers, q + r is an integer. So b + c = a \cdot (q + r) where q + r is an integer, which means a \mid (b + c) by definition. \square

This proof is three lines long, but it is air-tight. Each line does one thing: unpack a definition, perform one algebraic step, or repack the result into a definition. If you can write proofs in this style — small, clean, with every step justified — you can prove anything that a direct proof can reach.

Here is a slightly more involved divisibility result.

Claim. If a \mid b, then a \mid bc for any integer c.

Proof. Assume a \mid b. By definition, b = aq for some integer q. Then:

bc = (aq)c = a(qc)

Since q and c are integers, qc is an integer. So bc = a \cdot (qc) and a \mid bc. \square

The same skeleton: unpack, algebra, repack. The technique carries you through every divisibility proof in the early chapters of number theory.

Knowing when direct proof works

Direct proof works whenever there is a clear algebraic path from hypothesis to conclusion. It is the right first attempt for:

Direct proof struggles when the conclusion is a negative statement — "there is no integer such that..." or "it is impossible that..." — because walking forward from the hypothesis tends to produce positive information, not negative information. For negative conclusions, proof by contradiction or proof by contrapositive is often the better tool.

The interactive figure below shows a direct proof in action. The claim is: "if n is odd, then n^2 is odd." Drag the slider to choose an odd value of n, and watch the figure compute n^2 and display its factorisation, confirming that n^2 is always of the form 2m + 1.

Interactive verification that odd squared is oddA number line from 0 to 10 with a draggable red point. Above the line, readouts show the current value of n, the value of n squared, and the expression 2 times k plus 1 that equals n squared, confirming n squared is odd whenever n is odd.0246810↔ drag to choose k (n = 2k+1)
Drag the red point to choose a value of $k$. The readout computes $n = 2k + 1$ (always odd) and then $n^2$. Because $n^2 = (2k+1)^2 = 4k^2 + 4k + 1 = 2(2k^2 + 2k) + 1$, the result always has the form $2m + 1$, confirming it is odd. The algebra in the proof handles every $k$ at once; the figure lets you verify individual cases.

Two worked examples

Example 1: Prove that the sum of any two even integers is even

Claim. If m and n are even integers, then m + n is even.

Step 1. Start by unpacking the hypothesis.

Since m is even, m = 2a for some integer a. Since n is even, n = 2b for some integer b.

Why: the definition of "even" is the entry point. You cannot do anything with "m is even" until you convert it into an equation you can manipulate. The equation m = 2a is the manipulable form.

Step 2. Compute m + n.

m + n = 2a + 2b

Why: this is pure substitution — replacing m and n with their definitions. No new idea is needed; the algebra does the work.

Step 3. Factor out the common factor of 2.

m + n = 2(a + b)

Why: factoring out 2 puts the expression into the form 2 \times (\text{integer}), which is exactly the form the definition of "even" demands. The distributive law is the algebraic rule doing the work here.

Step 4. Verify that a + b is an integer.

Since a and b are integers, their sum a + b is an integer. (The integers are closed under addition — see Operations and Properties.)

Why: the definition of "even" requires n = 2k where k is an integer. You must check that a + b lands in the integers, not just assume it. Closure of \mathbb{Z} under addition is the fact that makes this step work.

Step 5. Apply the definition of "even" to conclude.

Since m + n = 2(a + b) and a + b is an integer, m + n is even by definition. \square

Result. The sum of two even integers is always even.

Geometric picture of adding two even numbersTwo rows of paired dots representing even numbers m equals 2a and n equals 2b. The first row shows a pairs of dots for m. The second row shows b pairs for n. When combined into a single row, the dots remain in pairs, visually confirming the sum is even. The combined row has a plus b pairs, so m plus n equals 2 times a plus b.m = 2a:a = 3 pairsn = 2b:b = 2 pairsm + n:a + b = 5 pairsAll dots are paired → m + n = 2(a + b) → even
Even numbers come in pairs. With $m = 6$ ($a = 3$ pairs) and $n = 4$ ($b = 2$ pairs), the combined row has $a + b = 5$ pairs — still perfectly paired, so $m + n = 10$ is even. The picture works for any $a$ and $b$: concatenating two collections of pairs always gives a collection of pairs. The algebraic proof says the same thing in symbols: $2a + 2b = 2(a + b)$.

The proof has five steps, but the core move is just one idea: factor out the 2. The rest is definition-unpacking and definition-checking. Every clean direct proof has this shape — one key algebraic insight surrounded by careful bookkeeping.

Example 2: Prove that if $3 \mid n$ and $3 \mid m$, then $9 \mid nm$

Claim. If 3 divides n and 3 divides m, then 9 divides nm.

Step 1. Unpack the hypotheses using the definition of divisibility.

Since 3 \mid n, there exists an integer p such that n = 3p. Since 3 \mid m, there exists an integer q such that m = 3q.

Why: the definition of "3 divides n" is that n is a multiple of 3. Writing n = 3p is the algebraic form of that statement, and it gives you a handle to manipulate.

Step 2. Compute the product nm.

nm = (3p)(3q)

Why: substitute the definitions from Step 1. The product is now expressed entirely in terms of p and q, which are free integers.

Step 3. Simplify the product.

nm = 9pq

Why: 3 \times 3 = 9. The two factors of 3 — one from n and one from m — combine to give a factor of 9. This is the key insight of the proof.

Step 4. Conclude using the definition of divisibility.

Since p and q are integers, pq is an integer (closure of \mathbb{Z} under multiplication). So nm = 9 \cdot (pq) where pq is an integer, which means 9 \mid nm by definition. \square

Result. 9 \mid nm.

Area model showing that the product of two multiples of 3 is a multiple of 9A rectangle with height n and width m. The height is subdivided into 3 equal strips each of height p, and the width is subdivided into 3 equal strips each of width q. This creates a 3-by-3 grid of small rectangles, each with area pq. The total area is nm equals 9pq, showing that nm is a multiple of 9.pqpqpqpqpqpqpqpqpqm = 3qn = 3pqqqpppnm = (3p)(3q) = 9pq → 9 divides nm
Think of $n$ and $m$ as side lengths of a rectangle. Since $n = 3p$, the height splits into $3$ equal strips of height $p$. Since $m = 3q$, the width splits into $3$ equal strips of width $q$. The grid has $3 \times 3 = 9$ identical small rectangles, each of area $pq$. The total area is $9pq$ — a multiple of $9$. The picture is the proof: the two factors of $3$ combine into a $3 \times 3$ grid.

Notice how the area model makes the factor of 9 visually inevitable: a 3 \times 3 grid of identical blocks always has a total count that is a multiple of 9. The algebraic proof and the geometric picture are two views of the same fact.

Common confusions

Going deeper

If you came here to learn how to write a direct proof — what the structure is, how the steps connect, and how to handle even/odd and divisibility arguments — you have everything you need. The rest of this section is for readers who want to see how proofs connect to logic and where the formalism goes.

Proofs and logic

Every direct proof is, at its core, a chain of implications. The statement "if P, then Q" is an implication, written P \Rightarrow Q in symbolic logic. A direct proof of P \Rightarrow Q works by assuming P is true and deducing Q from it. The logical structure is:

P \Rightarrow R_1 \Rightarrow R_2 \Rightarrow \dots \Rightarrow Q

where R_1, R_2, \dots are intermediate results. Each arrow is one step of the proof. The chain is valid because implication is transitive: if P \Rightarrow R_1 and R_1 \Rightarrow R_2 and R_2 \Rightarrow Q, then P \Rightarrow Q.

This is the connection between proofs and Logic and Propositions. The proof is the concrete construction; the logic is the abstract framework that says why the construction works. You can write perfectly good proofs without ever mentioning the word "implication," but knowing the logical skeleton makes it easier to see where a proof is going and where it might be stuck.

Uniqueness proofs

A common variation of direct proof is the uniqueness proof: showing that something is the only object with a given property. The standard method is to assume two objects both have the property and then show they must be equal. For instance, to prove that the additive identity is unique: suppose 0 and 0' are both additive identities. Then 0 = 0 + 0' = 0', where the first equality uses the fact that 0' is an identity and the second uses the fact that 0 is an identity. So 0 = 0' — there is only one.

The role of definitions

A recurring theme in this article is that definitions are tools, not just labels. The definition of "even" is not a passive description — it is an active ingredient in every proof about evenness. The same is true of every definition in mathematics. When you learn a new definition, you are not just learning what a word means; you are gaining a new tool that can be plugged into the hypothesis or conclusion of a proof. The better you know your definitions, the more proofs you can write. Bhaskara II, one of the great Indian mathematicians, was famous for the precision and clarity of his definitions — and that precision was not decoration; it was the foundation of his proofs.

Where this leads next

Direct proof is the starting point of the proof universe. Every other technique is a variation on this one, used when the direct path is blocked.