Bounded Quantifiers and Restricted Domains

Real proofs almost never say 'for every $x$' — they say 'for every positive integer $x$,' or 'for every continuous function $f$ on $[0,1]$.' That little phrase that pins the variable to a set is the bounded quantifier, and it is how working mathematicians actually use $\forall$ and $\exists$. This section translates between the abstract unbounded quantifier you met in §2.1 and the bounded form that every analysis, algebra, and number-theory proof relies on. It also fixes the single rule about where the domain goes when you negate the statement — a rule that, once internalised, removes most quantifier mistakes from later chapters.

The bounded forms

For a set $A$ and a predicate $P$, the bounded universal and bounded existential are written:

$(\forall x \in A), P(x)$ — "every element of $A$ satisfies $P$."

$(\exists x \in A), P(x)$ — "some element of $A$ satisfies $P$."

You also see the shorthand $\forall x \in A,; P(x)$ and $\exists x \in A,; P(x)$ in textbooks; they mean the same thing. The set $A$ is called the restricted domain of the quantifier. In Velleman's notation the parentheses around the quantifier make the scope explicit, which matters once you start stacking multiple bounded quantifiers.

Rewriting bounded as unbounded

The bounded forms are defined as syntactic sugar for unbounded quantifiers plus a domain guard. The two rewrite rules are:

Notice the asymmetry: the universal uses $\Rightarrow$ (an implication), while the existential uses $\wedge$ (a conjunction). This is not an arbitrary stylistic choice — it is forced by what each statement says.

Try the universal first. "Every prime greater than $2$ is odd" says: whenever $x$ is a prime greater than $2$, $x$ is odd. As a formula:

If $x$ is not in $P_{>2}$, the implication is vacuously true and the statement makes no claim about $x$. That is exactly what we want: a universal claim about elements of $A$ should ignore elements outside $A$.

Now the existential. "Some prime is even" says: there is an $x$ that is both prime AND even. As a formula:

The conjunction is essential: the witness has to live in $A$ AND satisfy $P$. Using $\Rightarrow$ here would be wrong — it would be satisfied vacuously by any $x \notin A$, which is not what "there exists an element of $A$" means.

Negating bounded quantifiers

The negation rules for bounded quantifiers follow from the unbounded rules in §2.2, but the practical pattern is worth memorising directly:

The domain $A$ stays put — only the inside flips. To see why, expand to the unbounded form, push the negation in, and contract back:

The middle step uses $\neg(P \Rightarrow Q) \equiv P \wedge \neg Q$ — the same rule you saw at the end of §2.2. The domain $A$ is preserved through the whole manipulation.

Analysis preview: the $\varepsilon$–$\delta$ pattern

The most important place bounded quantifiers appear is the definition of a limit, which you will meet in calculus and again in §5.5 (continuity, treated informally). The official definition is:

The shorthand $\forall \varepsilon > 0$ is itself a bounded quantifier — it really means $\forall \varepsilon \in \mathbb{R},;(\varepsilon > 0 \Rightarrow \cdots)$. Negating this entire formula to get the definition of "$\lim_{x \to a} f(x) \neq L$" is a classic exercise:

Every quantifier flipped, the implication collapsed into a conjunction, and the inner inequality flipped — but $\varepsilon > 0$ and $\delta > 0$ stayed put as bounded conditions. That stability is exactly what the negation rules for bounded quantifiers buy you.

Pause and think: Why does the bounded universal expand with $\Rightarrow$ but the bounded existential expand with $\wedge$? (Hint: a universal claim about $A$ should make no commitment about elements outside $A$ — implication handles this. An existential claim about $A$ demands a witness that lives in $A$ — conjunction enforces this.)

Try it

• Before writing: predict the unbounded-quantifier expansion of "Every nonzero rational has a reciprocal." Then write the bounded form $(\forall q \in \mathbb{Q} \setminus {0}),\cdots$ and confirm your expansion matches the rule.
• Predict first: which connective does the existential bounded form expand with? Translate "Some integer between $4$ and $9$ is prime" into bounded form, then into unbounded form, and check.
• Negate "Every continuous function on $[0,1]$ attains a maximum value." Push the negation through the bounded quantifier without losing the domain ${f : f \text{ continuous on } [0,1]}$.
• The statement "Every prime greater than 2 is odd" is true. The statement "Every prime is odd" is false. Rewrite both as bounded quantifiers and identify exactly which set in the second statement contains a counterexample.
• Negate the $\varepsilon$–$\delta$ formula for "$\lim_{x \to 0} f(x) = 1$" step by step. Identify the witness object: what would a counterexample function look like?

A trap to watch for

The single most common mistake is using $\Rightarrow$ instead of $\wedge$ in the existential expansion (or vice versa). Writing $(\exists x \in A),P(x) \equiv \exists x,(x \in A \Rightarrow P(x))$ is wrong: that expanded form is trivially satisfied by any $x \notin A$, because $x \in A \Rightarrow P(x)$ is vacuously true when $x \notin A$. So the formula would claim "some witness exists," but the witness might be outside $A$ — defeating the whole point of the bounded quantifier. The rule is: universal pairs with implication, existential pairs with conjunction. The mnemonic is to test on an empty domain: the bounded universal over $\emptyset$ should be vacuously TRUE (no elements to falsify) and the bounded existential should be FALSE (no possible witness). The implication-vs-conjunction split is the only choice that produces both behaviours.

What you now know

You can read, write, and negate bounded quantifiers, and you know the implication-vs-conjunction asymmetry between bounded universals and bounded existentials. This is the working vocabulary of every later proof in the book: from "every Cauchy sequence converges" to "some root of $p(x)$ lies in $[0,1]$," the bounded form is what real mathematicians write. Chapter 2 closes here; Chapter 3 launches the structured-proof techniques that turn these statements into theorems.

References

• Velleman, D. J. (2019). How to Prove It: A Structured Approach (3rd ed.). Cambridge University Press, ch. 2.
• Enderton, H. B. (2001). A Mathematical Introduction to Logic (2nd ed.). Academic Press, ch. 2.
• Mendelson, E. (2015). Introduction to Mathematical Logic (6th ed.). Chapman and Hall/CRC, ch. 2.
• Halmos, P. R. (1960). Naive Set Theory. Van Nostrand, §3.