How to Read a Theorem

A Theorem Makes a Claim Under Precise Conditions

In How to Read a Mathematical Definition, we unpacked definitions into their objects, domains, and conditions. A theorem uses such definitions, together with axioms and earlier results, to make a claim that has been established by proof. When you read one, your first task is not to accept a familiar-looking conclusion or begin calculating. First identify exactly what the theorem says.

Many theorems have the form “if $P$, then $Q$.” Here $P$ is the hypothesis, the condition under which the result is asserted, and $Q$ is the conclusion the theorem guarantees in that circumstance. The theorem does not claim that $Q$ always holds, nor does it claim that $P$ always holds. It says that whenever the permitted inputs satisfy $P$, they also satisfy $Q$.

Reading rule. Record the domain of every variable, the full hypothesis, and the full conclusion. When applying the theorem, check that the particular objects lie in the stated domain and satisfy every part of the hypothesis.

For instance, compare “for every real number $x$, if $x>0$, then $x+2>2$” with “for every real number $x$, $x+2>2$.” The first statement has a condition and is true; the second makes a stronger claim and is false, since $x=-3$ gives $x+2=-1$. A phrase such as “if $x>0$” is not decoration: it restricts the circumstances in which the conclusion is guaranteed.

Find the Domain, Hypothesis, and Conclusion

A theorem often begins with words such as “let $x,y\in\mathbb R$” or “for every $n\in\mathbb Z$.” These identify the domain: the objects from which the variables may be chosen. The rest of the statement tells you which conditions are assumed and what must follow from them. A compact symbolic statement may make the logical form easier to see.

$$ \forall x\in D,\quad P(x)\Longrightarrow Q(x). $$

This says that for every permitted $x$ in $D$, if $P(x)$ holds, then $Q(x)$ holds. The theorem is not asserting $P(x)$ by itself. Nor does it say that $Q(x)$ must hold for inputs outside the specified domain. If a theorem names two variables, check the domain of each and whether the conclusion concerns both of them.

A theorem with no “if” phrase may still have conditions in its opening sentence. For example, “Let $A$ be a nonempty set bounded above. Then ...” includes both the nonemptiness and bounded-above requirements as hypotheses. The word “then” separates those requirements from the conclusion. Every stated hypothesis matters, even when it appears before the main sentence.

Mark the domain.
Identify the allowed objects for every variable, such as real numbers, integers, or elements of a stated set.

Collect all hypotheses.
Include conditions in the opening phrase as well as conditions following “if,” “whenever,” or “suppose.”

Locate the conclusion.
Find what the theorem promises once its hypotheses hold.

Match an application to the statement.
Substitute the intended objects and verify each hypothesis before using the conclusion.

Keep the direction unchanged.
A theorem from $P$ to $Q$ does not automatically permit reasoning from $Q$ back to $P$.

Applying a Theorem Means Checking Its Hypotheses

To apply a theorem, choose the objects that will play the roles of its variables. Then verify that they are in the stated domains and meet the assumptions. Only after that can you infer the conclusion. This is an instance of using an established result: the theorem supplies a valid implication, while the current argument must show that its hypothesis is satisfied.

A theorem quantified over every real number applies to integer inputs viewed as real numbers, provided its hypotheses hold. Integers are real numbers, so they fall within the real-number domain. Conversely, a theorem restricted to positive inputs cannot automatically be used for negative inputs. The domain and the hypotheses work together: belonging to the domain alone is not enough if the conditions are unmet.

Worked Example: Substituting an Integer into a Real-Number Theorem

Consider the theorem: for every $x\in\mathbb R$, if $x>0$, then $x+2>2$. We want to use it with the integer $x=5$. First, $5\in\mathbb R$, so the domain condition is met. Second, $5>0$, so the hypothesis is met. Substituting $5$ into the conclusion gives $5+2>2$, or $7>2$. This is a valid application because both the domain and the hypothesis have been checked.

Now try $x=-5$. This is still a real number, but the hypothesis $-5>0$ is false. Therefore the theorem gives no conclusion for this input. In fact, its conclusion would be $-3>2$, which is false. The theorem remains true: its conditional claim does not promise anything for inputs that fail the hypothesis.

The distinction between a theorem and its converse is another important part of reading. The converse of “if $P$, then $Q$” is “if $Q$, then $P$.” The original theorem does not establish its converse. The two statements have different directions, and the converse requires its own proof or a separate established result.

Statement form	What it says	What it does not establish by itself
If $P$, then $Q$	Whenever $P$ holds, $Q$ follows.	That $P$ holds, or that $Q$ implies $P$.
If $Q$, then $P$	Whenever $Q$ holds, $P$ follows.	That $P$ implies $Q$.
$P$ if and only if $Q$	Both implications hold.	Neither direction can be omitted when establishing the equivalence.

Worked Example: The Converse Is a Different Claim

We will prove below that if $0\leq x<y$, then $x^2<y^2$. Its converse would say: if $x^2<y^2$, then $0\leq x<y$. Take $x=-1$ and $y=2$. Then $x^2=1$ and $y^2=4$, so $x^2<y^2$ is true. But $0\leq x$ is false because $-1<0$. Thus the converse is false, even though the original theorem is true. This example shows why the direction of an implication must be preserved.

A single instance like this is enough to disprove the converse, because the converse claims the implication for all real $x,y$. It does not disprove the original theorem: for this same pair, the original hypothesis $0\leq x<y$ is false, so the theorem makes no promise for it.

Reading a Theorem Together with Its Proof

Theorem statements and proofs have complementary roles. The statement tells you what follows from which assumptions; the proof must justify that implication. In Introduction to Mathematical Proof, a conditional claim was approached by assuming its hypothesis and deriving its conclusion. When reading a proof, keep the theorem’s hypotheses visible and track how each is used.

Here is a new elementary result written in theorem form. It makes explicit a condition on the number added to both sides of an inequality.

Theorem. For all real numbers $x,y,c$, if $x<y$, then $x+c<y+c$.

Proof. Let $x,y,c\in\mathbb R$, and suppose $x<y$. Adding the same real number $c$ to both sides of a strict inequality preserves the inequality. Thus $$ x+c<y+c. $$ This is the conclusion for arbitrary real $x,y,c$ satisfying the hypothesis, so the theorem is proved.

The theorem’s scope includes every real $c$, not just positive $c$. For example, if $x=1$, $y=4$, and $c=-7$, its hypothesis $1<4$ holds and its conclusion is $-6<-3$. The value of $c$ does not need to be positive because adding the same number preserves order regardless of its sign. Reading the quantifiers prevents us from adding an unstated restriction.

Now consider a theorem whose hypothesis contains two conditions. It is not enough to check only one. The following result is a useful example: nonnegativity is needed along with the strict inequality in order to compare the squares.

Theorem. For all real numbers $x,y$, if $0\leq x<y$, then $x^2<y^2$.

Proof. Let $x,y\in\mathbb R$, and suppose $0\leq x<y$. From $x<y$, we have $y-x>0$. Since $x\geq0$, adding $x$ to both sides of $x<y$ gives $2x<x+y$; and since $y>x\geq0$, we have $x+y>0$. More directly, $x+y\geq y>0$ because $x\geq0$. Therefore both $y-x$ and $x+y$ are positive, so their product is positive. Factoring the difference of squares gives $$ y^2-x^2=(y-x)(y+x)>0. $$ Hence $x^2<y^2$. The argument applies to arbitrary real $x,y$ satisfying both parts of the hypothesis, which proves the theorem.

Worked Example: Checking Both Conditions Before Comparing Squares

Take $x=2$ and $y=5$. The theorem requires $0\leq x<y$. Here $0\leq2$ and $2<5$, so both conditions hold. The conclusion is $2^2<5^2$, and indeed $4<25$.

By contrast, $x=-3$ and $y=-1$ satisfy $x<y$, but they do not satisfy $0\leq x$. The theorem therefore cannot be used to compare their squares. In this case $(-3)^2=9$ and $(-1)^2=1$, so the conclusion $9<1$ would be false. This does not contradict the theorem, because the nonnegativity hypothesis was not met.

A Careful Reading Prevents Common Errors

Several tempting shortcuts lead to invalid uses of theorems. One is to remember only the conclusion and forget the hypotheses. Another is to silently broaden a domain, such as applying a statement about positive real numbers to a negative number. A third is to reverse an implication because the conclusion seems to suggest its hypothesis. Each mistake changes the claim.

There is also a distinction between using a theorem and proving a new statement. If all its hypotheses are known, a theorem permits its conclusion to be recorded. If even one hypothesis has not been established, the conclusion cannot yet be claimed from that theorem. A proof may need to establish the missing condition, use another theorem, or take a different approach. A theorem is not a general license to assert a useful-looking result.

When a theorem is stated with “for every,” do not test just one convenient input and treat that as proof of the whole theorem. A universal theorem covers every permitted input. To prove such a claim directly, the standard method is to take an arbitrary object in the stated domain, assume the hypotheses, and derive the conclusion. By contrast, to refute a universal claim, one counterexample in the domain that satisfies its hypotheses but violates its conclusion is enough.

Keep the implication intact. A theorem licenses the conclusion only for objects in its domain that satisfy all of its hypotheses. It neither asserts that the hypotheses always hold nor reverses its own direction.

Check Your Understanding

For each question, identify the relevant domain, hypotheses, and conclusion rather than relying on a familiar pattern.

In a theorem stated “for every $x\in\mathbb R$, if $P(x)$, then $Q(x)$,” what must be checked before concluding $Q(a)$ for a particular $a$?
A theorem applies to every real number $x$. May it be applied to an integer $n$, regarded as a real number? What additional check may still be needed?
State the converse of “if $P$, then $Q$.” Does the original implication establish the converse?
For the theorem “if $0\leq x<y$, then $x^2<y^2$,” list both parts of the hypothesis.
The pair $x=-1,\ y=2$ satisfies $x^2<y^2$, but does not satisfy $0\leq x<y$. What does this show about the converse, and why does it not contradict the theorem?
When applying a theorem to a specific pair of objects, why is checking only the conclusion—or only one of several hypotheses—not enough?

Statement form	What it says	What it does not establish by itself
If \(P\), then \(Q\)	Whenever \(P\) holds, \(Q\) follows.	That \(P\) holds, or that \(Q\) implies \(P\).
If \(Q\), then \(P\)	Whenever \(Q\) holds, \(P\) follows.	That \(P\) implies \(Q\).
\(P\) if and only if \(Q\)	Both implications hold.	Neither direction can be omitted when establishing the equivalence.

What You'll Learn