Bounds and Extrema of Sets: Subsets of an Ordered Field and Boundedness
Definitions
A field $F$ with an order relation $<$ satisfying trichotomy, transitivity, order-addition compatibility, and order-multiplication compatibility is called an ordered field. The principal examples are $\mathbb{R}$ (the reals) and $\mathbb{Q}$ (the rationals).
Let $F$ be an ordered field and $E \subseteq F$, $E \neq \emptyset$. An element $b \in F$ is an upper bound of $E$ if $$x \leq b \quad \text{for every } x \in E.$$ If such a $b$ exists, $E$ is bounded above.
An element $a \in F$ is a lower bound of $E$ if $$a \leq x \quad \text{for every } x \in E.$$ $E$ is bounded if bounded above and below — equivalently, $\exists\, M > 0$ such that $\lvert x \rvert \leq M$ for all $x \in E$.
$\alpha \in F$ is the supremum of $E$, written $\alpha = \sup E$, if:
(i) $x \leq \alpha$ for all $x \in E$, and
(ii) if $\gamma < \alpha$ then $\exists\, x \in E$ with $\gamma < x$.
$\varepsilon$-characterisation: $\alpha = \sup E$ iff $\alpha$ is an upper bound and for every $\varepsilon > 0$ there exists $x \in E$ with $\alpha - \varepsilon < x \leq \alpha$.
$\beta \in F$ is the infimum of $E$, written $\beta = \inf E$, if:
(i) $\beta \leq x$ for all $x \in E$, and
(ii) if $\gamma > \beta$ then $\gamma$ is not a lower bound of $E$.
Key relation: $\inf E = -\sup(-E)$ where $-E = \{-x : x \in E\}$.
Rudin, Principles of Mathematical Analysis, 3rd Ed., Ch. 1, §1.7–§1.8
Bartle & Sherbert, Introduction to Real Analysis, 4th Ed., Ch. 2, §2.3
→ Intervals and the Real Line — bounded intervals are the simplest examples of bounded sets
→ Order Relations and Ordered Sets — bounds live in an ordered structure
→ Algebraic Structures and Field Axioms — the field framework is required
→ Absolute Value and the Real Line — used in the criterion $\lvert x\rvert \leq M$
Intuition & Historical Background
Think of $E$ as a collection of students' exam scores. An upper bound is any score that no student exceeded — 100, 150, or 200 all qualify. The supremum is the tightest such ceiling: the smallest number that is still a valid upper bound. Crucially, the supremum may or may not actually be achieved by any element of $E$. When it is, it coincides with the maximum; when it isn't, the set "approaches but never reaches" its ceiling.
"The greatest rigor is achievable only when one first comprehends the very nature of the infinite."
— Richard Dedekind
Origin Story. Richard Dedekind (1831–1916) formalised the concept of bounds in his 1872 monograph Stetigkeit und irrationale Zahlen. Working in Göttingen, he observed that calculus relied on intuitive notions of "continuity" without a rigorous foundation and set out to find the precise property distinguishing $\mathbb{R}$ from $\mathbb{Q}$.
The Problem It Solved. The set $S = \{x \in \mathbb{Q} : x > 0,\, x^2 < 2\}$ is bounded above in $\mathbb{Q}$ (e.g.\ by 2), yet it has no supremum within $\mathbb{Q}$ — the missing least upper bound is $\sqrt{2}$, which is irrational. This gap in $\mathbb{Q}$ cannot be patched by algebra alone; it requires the Completeness Axiom.
Interesting Facts.
• Cauchy proved convergence theorems for sequences tacitly assuming the supremum property, decades before Dedekind stated it explicitly.
• The supremum of $\left\{\tfrac{n}{n+1} : n \in \mathbb{N}\right\}$ is $1$, yet $1$ is never attained — an infinite set can "asymptotically approach" its supremum forever.
Modern Relevance. Supremum and infimum underpin the Riemann integral (Darboux sums), operator norms in functional analysis, and optimisation across mathematics, physics, and computing — and appear in nearly every CSIR NET real analysis question on convergence and continuity.
A bounded set $E \subset \mathbb{R}$: every lower bound lies at or left of $\inf E$; every upper bound lies at or right of $\sup E$.
मान लीजिए $E$ परीक्षा के अंकों का एक समुच्चय है। एक ऊपरी परिबंध (upper bound) वह संख्या है जिससे कोई भी अंक बड़ा नहीं है। Supremum वह सबसे छोटा upper bound है — चाहे वह $E$ में हो या न हो। इसी तरह, Infimum सबसे बड़ा lower bound है। Dedekind ने 1872 में यह अवधारणा इसलिए दी क्योंकि $\mathbb{Q}$ में कुछ परिबद्ध समुच्चयों का supremum नहीं था — यही $\mathbb{R}$ और $\mathbb{Q}$ का सबसे बड़ा अंतर है।
Solved Examples
1Very Easy — Direct Identification
Let $E = [2, 5]$. Find $\sup E$ and $\inf E$. Is $E$ bounded?
• Every $x \in E$ satisfies $x \leq 5$, so $5$ is an upper bound. Since $5 \in E$, no number smaller than $5$ is an upper bound. Hence $\sup E = \max E = 5$.
• Every $x \in E$ satisfies $x \geq 2$, so $\inf E = \min E = 2$.
• $E$ is bounded: $\lvert x\rvert \leq 5$ for all $x \in [2,5]$.
2Easy–Medium — Supremum Not Attained
Let $E = (0, 3)$. Find $\sup E$ and $\inf E$.
• For all $x \in (0,3)$, $x < 3$, so $3$ is an upper bound.
• Let $\gamma < 3$. Set $x_0 = \frac{\gamma + 3}{2}$; then $\gamma < x_0 < 3$, so $x_0 \in E$ yet $x_0 > \gamma$. Hence $\gamma$ is not an upper bound. Therefore $\sup E = 3$.
• Similarly $\inf E = 0$. Neither $0$ nor $3$ belongs to $E$, so $E$ has no maximum and no minimum.
3Medium–Hard — Infimum Not Attained
Let $E = \left\{\dfrac{1}{n} : n \in \mathbb{N}\right\}$. Find $\sup E$ and $\inf E$, and state whether each is attained.
• Supremum: $\frac{1}{1} = 1 \in E$ and $\frac{1}{n} \leq 1$ for all $n$. Hence $\sup E = \max E = 1$ (attained).
• Infimum: All terms are positive, so $0$ is a lower bound. Let $\varepsilon > 0$. By the Archimedean Property, $\exists\, n_0 \in \mathbb{N}$ with $\frac{1}{n_0} < \varepsilon$. Since $\frac{1}{n_0} \in E$ and $\frac{1}{n_0} < \varepsilon$, no positive number is a lower bound. Hence $\inf E = 0$ (not attained, since $0 \notin E$).
$\{1/n : n\in\mathbb{N}\}$: $\sup = 1$ (attained at $n=1$), $\inf = 0$ (not attained).
4Hard — CSIR NET / IIT JAM Level
Let $A = \left\{\dfrac{m}{n} : m, n \in \mathbb{N},\, m < n\right\}$. Determine $\sup A$ and $\inf A$, and state whether each is attained.
Step 1. Since $m,n \geq 1$ and $m < n$, every element satisfies $0 < m/n < 1$. So $A$ is the set of all positive rationals strictly less than $1$.
Step 2 — Supremum. $1$ is an upper bound. For $\gamma < 1$, by density of $\mathbb{Q}$ choose $m/n \in A$ with $\gamma < m/n < 1$. So $\gamma$ is not an upper bound; hence $\sup A = 1$. Since $m < n$ strictly, $1 \notin A$: the supremum is not attained.
Step 3 — Infimum. $0$ is a lower bound. For $\varepsilon > 0$, pick $n_0 > 1/\varepsilon$ (Archimedean Property); then $1/n_0 \in A$ and $1/n_0 < \varepsilon$. So $\inf A = 0$, not attained.
Quick Revision Cards
A — Key Definitions
Upper bound: $x \leq b\ \forall x\in E$
Lower bound: $a \leq x\ \forall x\in E$
Bounded: $\exists M>0$, $\lvert x\rvert\leq M\ \forall x\in E$
$\sup E$ = least upper bound
$\inf E$ = greatest lower bound
$\inf E = -\sup(-E)$
B — Conditions & Edge Cases
• $\sup E\in E \Leftrightarrow \max E$ exists
• $\inf E\in E \Leftrightarrow \min E$ exists
• Finite sets always have max & min
• In $\mathbb{Q}$: $\{x:x^2<2\}$ has no sup
• $\varepsilon$-char: $\forall\varepsilon>0,\exists x\in E: x>\sup E-\varepsilon$
C — Exam Tips
🔵 CSIR NET: $\varepsilon$-proof of sup/inf
🟢 GATE: bounded $\Leftrightarrow$ $|x|\leq M$
🟠 IIT JAM: check if sup/inf attained
🔴 Raj B.Sc.: distinguish sup from max
Common Mistakes
1. Confusing $\sup E$ with $\max E$. The supremum always exists in $\mathbb{R}$ for non-empty bounded-above sets (Completeness Axiom); the maximum exists only if $\sup E \in E$. The open interval $(0,1)$ has $\sup = 1$ but no maximum.
2. Assuming sup exists in every ordered field. In $\mathbb{Q}$, the set $\{x \in \mathbb{Q}: x^2 < 2\}$ is bounded above but has no supremum within $\mathbb{Q}$. The Completeness Axiom is the extra property of $\mathbb{R}$.
3. Weak inequality in the $\varepsilon$-argument. To show $\alpha = \sup E$, you must exhibit $x \in E$ with $x > \alpha - \varepsilon$ (strict inequality). The weak $x \geq \alpha - \varepsilon$ is insufficient.
4. Sign error with $-E$. $\sup(-E) = -\inf E$ and $\inf(-E) = -\sup E$. Forgetting the negation is a frequent error in exam proofs.
Real-World Applications
Every numerical solver seeks the $\sup$ or $\inf$ of an objective function. The Extreme Value Theorem guarantees a continuous function on $[a,b]$ attains its supremum.
Darboux upper and lower sums are suprema and infima of $f$ over subintervals. The Riemann integral is defined via $\inf_P U(f,P) = \sup_P L(f,P)$.
The operator norm $\|T\| = \sup\{\|Tx\| : \|x\|=1\}$ is a supremum construction central to Banach space theory.
IEEE 754 floating-point overflow triggers when a value exceeds the supremum of representable numbers. Complexity bounds use infima of running times.
Summary Table
| Concept | Definition / Statement | Condition | Reference |
|---|---|---|---|
| Upper bound | $b\in F$: $x\leq b\ \forall x\in E$ | $E\neq\emptyset$ | Rudin §1.7 |
| Lower bound | $a\in F$: $a\leq x\ \forall x\in E$ | $E\neq\emptyset$ | Rudin §1.7 |
| Bounded | $\exists M>0$: $\lvert x\rvert\leq M\ \forall x\in E$ | Both bounds exist | Bartle §2.3 |
| $\sup E$ | Least upper bound | $E$ bounded above | Rudin §1.8 |
| $\inf E$ | Greatest lower bound | $E$ bounded below | Rudin §1.8 |
| $\sup(-E)$ | $= -\inf E$ | $E$ bounded below | Standard |
| $\varepsilon$-char. | $\forall\varepsilon>0,\exists x\in E: x>\sup E-\varepsilon$ | $E$ bounded above | Bartle §2.3 |
Cross-References
→ Intervals and the Real Line — bounded intervals as concrete examples
→ Order Relations and Ordered Sets — the ordered structure where bounds are defined
→ Algebraic Structures and Field Axioms — field axioms underpinning ordered fields
→ Absolute Value and the Real Line — the $\lvert x\rvert \leq M$ boundedness criterion
→ Foundations of Real Numbers — why $\mathbb{R}$ is needed at all
⟶ Next topic: The Completeness Axiom and the Archimedean Property
How did you find this post?
Tap a reaction — counts are saved permanently on our server.
Have a question, doubt, or thought about this post? Choose how you want to join the discussion below.
💬 Comment on Telegram — No account needed