Order Relations and Ordered Sets
Order Relations and Ordered Sets
We say $3 < 5$ so naturally that we rarely pause to ask: what exactly does the symbol $<$ mean, and what rules must it satisfy? This post formalises the notion of an order relation, climbs from partial to total orders, adds the order structure to a field to get an ordered field, and then verifies concretely that $\mathbb{R}$ — the real number system — is the prime example. Along the way we prove that $\mathbb{C}$ can never be ordered.
🇮🇳 हिंदी में पढ़ेंOrder Relations — Partial and Total
A partial order on a non-empty set $S$ is a relation $\leq$ satisfying for all $x,y,z \in S$:
(i) Reflexivity: $x \leq x$.
(ii) Antisymmetry: $x \leq y$ and $y \leq x \;\Rightarrow\; x = y$.
(iii) Transitivity: $x \leq y$ and $y \leq z \;\Rightarrow\; x \leq z$.
The pair $(S, \leq)$ is called a partially ordered set or poset.
Total order (linear order): A partial order that additionally satisfies Trichotomy — for all $x, y \in S$, exactly one of $x < y$, $x = y$, $y < x$ holds. Then $(S,\leq)$ is a totally ordered set (or chain).
Given a partial order $\leq$ on $S$, the strict order $<$ is defined by: $x < y \;\Leftrightarrow\; x \leq y$ and $x \neq y$. Conversely, given a strict (irreflexive, transitive, asymmetric) order $<$, recover $\leq$ by $x \leq y \;\Leftrightarrow\; x < y$ or $x = y$.
Key distinction: $\leq$ is non-strict (allows equality); $<$ is strict (excludes equality).
Partial Order
Reflexive
Antisymmetric
Transitive
Total Order
Partial order
+ Trichotomy
(comparability)
Ordered Field
Field
+ Total order
+ (O1) + (O2)
ℝ
Ordered field
+ LUB property
(Complete)
← Sets and Basic Notation — ordered sets are sets equipped with extra structure.
← Functions and Relations — an order is a special binary relation on a set.
← Algebraic Structures and Field Axioms — the nine field axioms that ℝ already satisfies.
← Consequences of Field Axioms — derived properties that carry into ordered fields.
Intuition — Why Order Needs Axioms
The symbol $<$ on the real number line feels obvious — but the moment you ask "can we order the complex numbers the same way?", the answer is a definitive no. The order axioms pin down exactly what properties any sensible ordering must have, and reveal which number systems admit them.
Think of the distinction between partial and total orders this way: in a total order, any two elements can be placed on a single "number line." In a partial order, some elements are simply incomparable — like subsets $\{1\}$ and $\{2\}$ under $\subseteq$, neither contains the other.
An ordered field is a field $(F, +, \cdot)$ with a total order $<$ satisfying two additional axioms:
(O1) — Order + Addition
Order is preserved when adding the same element to both sides.
(O2) — Positive Product
Product of two positives is positive.
Key consequences (Rudin, Prop. 1.18): In any ordered field: $x^2 \geq 0$ for all $x$; $1 > 0$; $x > 0 \Rightarrow x^{-1} > 0$; $x < 0 \Rightarrow x^{-1} < 0$; multiplying by a negative reverses the inequality.
Origin Story. The idea of ordering numbers goes back to antiquity, but the precise axiomatic treatment of ordered fields emerged from the work of Richard Dedekind (1872) and was codified by David Hilbert in his Grundlagen der Geometrie (1899). Dedekind's motivation was to construct $\mathbb{R}$ from $\mathbb{Q}$ rigorously — his famous Dedekind cuts gave the first proof that a complete ordered field exists. Georg Cantor independently constructed $\mathbb{R}$ via Cauchy sequences in the same year.
The Problem It Solved. Analysts since Cauchy (1820s) had been using the "number line" without a rigorous definition of what a real number actually is. Dedekind's order-based construction resolved this gap and showed precisely why $\mathbb{Q}$ — though an ordered field — is insufficient for analysis: it has "holes" (like $\sqrt{2}$).
Interesting Fact. The fact that $\mathbb{C}$ cannot be ordered is not a limitation of our imagination — it is a theorem. Any attempt to put a total order on $\mathbb{C}$ compatible with its field operations immediately leads to $-1 > 0$ (since $i^2 = -1$ and squares must be non-negative), a contradiction. This is why complex analysis has no "positive" complex numbers.
Why It Matters Today. The ordered field structure of $\mathbb{R}$ — together with completeness — is the foundation on which all of real analysis, calculus, differential equations, and optimisation theory is built; and the non-orderability of $\mathbb{C}$ is a standard CSIR NET / GATE question.
आंशिक क्रम (Partial Order): तीन गुण — स्वतुल्यता, प्रतिसममिति, संक्रामकता। पूर्ण क्रम (Total Order): आंशिक क्रम + तुलनीयता (Trichotomy)। क्रमित क्षेत्र (Ordered Field): Field + (O1) + (O2)। $\mathbb{R}$ एक क्रमित क्षेत्र है क्योंकि $y < z \Rightarrow x+y < x+z$ और धनात्मक $\times$ धनात्मक $= $ धनात्मक। $\mathbb{C}$ को क्रमित नहीं किया जा सकता — $i^2 = -1$ से $-1 > 0$ आता है जो विरोधाभास है।
Solved Examples
(a) $(\mathcal{P}(\{1,2\}),\subseteq)$: $\mathcal{P}(\{1,2\}) = \{\emptyset, \{1\}, \{2\}, \{1,2\}\}$. Reflexive ($A \subseteq A$) $\checkmark$; antisymmetric ($A \subseteq B$ and $B \subseteq A \Rightarrow A=B$) $\checkmark$; transitive $\checkmark$. But $\{1\}$ and $\{2\}$ are incomparable. Partial order, not total.
(b) $(\mathbb{R}, \leq)$: All three partial order properties hold, and for any $x,y \in \mathbb{R}$ exactly one of $x < y$, $x=y$, $y < x$ holds. Total order.
(c) $(\mathbb{N}, \mid)$: $a \mid a$ (reflexive $\checkmark$); $a \mid b$ and $b \mid a \Rightarrow a=b$ for $a,b \in \mathbb{N}$ (antisymmetric $\checkmark$); transitive $\checkmark$. But $2$ and $3$ are incomparable under $\mid$. Partial order, not total. $\blacksquare$
Verify (O1): Take $x = 3$, $y = 1$, $z = 5$ in $\mathbb{R}$. We have $y < z$ (i.e. $1 < 5$). Then $x + y = 4$ and $x + z = 8$, and $4 < 8$. $\checkmark$
General (O1) proof in $\mathbb{R}$: If $y < z$, then $z - y > 0$. Then $(x+z)-(x+y) = z - y > 0$, so $x+y < x+z$. $\checkmark$
Verify (O2): Take $x = 2 > 0$ and $y = 3 > 0$. Then $xy = 6 > 0$. $\checkmark$
General (O2) in $\mathbb{R}$: By standard real arithmetic, the product of two positive reals is always positive. $\checkmark$ $\blacksquare$
Given: $x < 0$, so $-x > 0$. We want $x^{-1} < 0$.
Suppose for contradiction $x^{-1} > 0$. Then by (O2): $(-x) \cdot x^{-1} > 0$.
But $(-x) \cdot x^{-1} = -(x \cdot x^{-1}) = -(1) = -1$ (by field consequences).
So $-1 > 0$, meaning $1 < 0$. But we proved $1 = 1^2 > 0$ in every ordered field — contradiction.
Since $x^{-1} \neq 0$ (it is a multiplicative inverse), we conclude $x^{-1} < 0$. $\blacksquare$
Part 1 — $\mathbb{C}$ is not orderable.
Suppose $<$ makes $(\mathbb{C}, +, \cdot, <)$ an ordered field. By consequence (iv) of Prop. 1.18: $x^2 \geq 0$ for all $x$. Apply to $x = 1$: $1 > 0$. Apply to $x = i$: $i^2 \geq 0$, i.e. $-1 \geq 0$. But $1 > 0$ implies $-1 < 0$ (by O1 with $x = -1, z = 0, y+(-1) < z+(-1)$). So $-1 \geq 0$ and $-1 < 0$ simultaneously — contradiction. $\blacksquare$
Part 2 — Properties of $P = \{x \in F : x > 0\}$.
Closed under addition: If $a, b \in P$ then $a > 0$ and $b > 0$. By (O1): $a + b > 0 + b = b > 0$. So $a + b \in P$. $\checkmark$
Closed under multiplication: Directly from (O2): $a, b \in P \Rightarrow ab \in P$. $\checkmark$
Trichotomy for $P$: For any $x \in F$, trichotomy gives exactly one of: $x > 0$ (i.e. $x \in P$), $x = 0$, $x < 0$ (i.e. $-x \in P$). $\blacksquare$
Quick Revision Cards
📊 A — Key Definitions & Results
- Partial order: Reflexive + Antisymm. + Trans.
- Total order: Partial + Trichotomy
- Strict: $x < y \Leftrightarrow x \leq y$, $x \neq y$
- (O1): $y < z \Rightarrow x+y < x+z$
- (O2): $x>0$ and $y>0 \Rightarrow xy>0$
- $x^2 \geq 0$; $1 > 0$ in any ordered field
- $\mathbb{R}$: unique complete ordered field
⚙️ B — Edge Cases & Conditions
- $(\mathcal{P}(S), \subseteq)$: partial, NOT total (for $|S| \geq 2$)
- $\mathbb{C}$: field but NOT orderable (proof by $i^2 = -1$)
- $\mathbb{Q}$: ordered field, but NOT complete
- $\mathbb{Z}$: totally ordered, but NOT a field
- Multiplying by negative: reverses $<$ (theorem, not axiom)
- Every total order is a partial order; not vice versa
🎯 C — Exam Tips
- 🔵 CSIR NET: "$\mathbb{C}$ ordered field?" — always NO; one-line proof via $i^2=-1$
- 🟢 GATE: Disprove total order: find one incomparable pair
- 🟠 IIT JAM: Know all five consequences of Prop. 1.18
- 🔴 B.Sc. Raj.: Verify (O1) and (O2) separately with explicit steps for $\mathbb{R}$
Common Mistakes
❌ Errors to Avoid
Wrong: "Subsets can always be compared using $\subseteq$." | Correct: $\{1\}$ and $\{2\}$ are incomparable under $\subseteq$ — neither is a subset of the other. So $\subseteq$ is only a partial order when $|S| \geq 2$.
Wrong: "We can order $\mathbb{C}$ by comparing moduli $|z|$." | Correct: The modulus gives a total order on $\mathbb{R}_{\geq 0}$, but it is not compatible with the field operations of $\mathbb{C}$. The ordering axioms (O1), (O2) are violated: e.g. $i > 0$ and $i > 0$ would give $i \cdot i = -1 > 0$ — contradiction.
Wrong: Citing "multiplying by a negative reverses the inequality" as axiom (O3). | Correct: There are only two order axioms (O1) and (O2). Sign-reversal is a derived theorem: if $x < 0$ and $y < z$, then $(-x) > 0$, so $(-x)(z-y) > 0$ by (O2), giving $xz < xy$ by (O1).
Wrong: "An order is partial if it's not strict." | Correct: "Partial" refers to the possibility of incomparable elements. "Strict" ($<$) vs "non-strict" ($\leq$) is a separate distinction — both can be partial or total.
Real-World Applications
Sorting Algorithms
Every comparison-based sort (quicksort, mergesort) requires a total order on the data. Topological sort in build systems and compilers uses a partial order (dependency DAG).
Optimisation & Economics
Consumer preference is modelled as a partial order (indifference is allowed). Linear programming over $\mathbb{R}$ requires the ordered field structure to compare objective values and determine feasibility.
Real Analysis
The ordered field structure of $\mathbb{R}$ (plus completeness) is what makes limits, derivatives, and integrals well-defined. $\mathbb{Q}$ is an ordered field but lacks the completeness that analysis requires.
Database Theory
B-tree indices require a total order on keys. Git commit history is a poset (partial order) — commits have a partial, not total, temporal relationship when branches diverge.
Summary Table & Key Result
| Structure | Key Requirement | Example | Non-example |
|---|---|---|---|
| Partial order | Reflexive, antisymm., transitive | $(\mathcal{P}(S),\subseteq)$ | $(\mathbb{Z},\mid)$ on all of $\mathbb{Z}$ |
| Total order | Partial order + trichotomy | $(\mathbb{R},\leq)$, $(\mathbb{Q},\leq)$ | $(\mathcal{P}(\{1,2\}),\subseteq)$ |
| Ordered field | Field + (O1) + (O2) | $\mathbb{Q}$, $\mathbb{R}$ | $\mathbb{C}$, $\mathbb{F}_p$ |
| Complete ord. field | Ordered field + LUB property | $\mathbb{R}$ (unique) | $\mathbb{Q}$ |
The Hierarchy of Number Systems with Order:
$$(\mathbb{Z},\leq)\ \text{total order, not a field} \quad\subset\quad (\mathbb{Q},\leq)\ \text{ordered field, not complete} \quad\subset\quad (\mathbb{R},\leq)\ \text{complete ordered field}$$
In every ordered field: $x^2 \geq 0$, $\;1 > 0$, $\;x > 0 \Leftrightarrow x^{-1} > 0$.
$\mathbb{C}$ is a field but admits no compatible total order.
Cross-References & Related Posts
← Prerequisites: Sets and Basic Notation — sets are the carrier sets of ordered structures | Functions and Relations — orders are special relations | Algebraic Structures and Field Axioms — the field axioms combined with order axioms here | Consequences of Field Axioms — derived properties used in ordered field proofs.
→ Next Topic: Bounded Sets, Supremum and Infimum — the Least Upper Bound property of $\mathbb{R}$ in depth; Archimedean property; density of $\mathbb{Q}$ in $\mathbb{R}$.
📖 Further Reading: Rudin, Ch. 1, §§1.5–1.21; Apostol, Ch. 1, §§1.7–1.10; Bartle & Sherbert, Ch. 2, §2.3.
How did you find this post?
Tap a reaction — counts update in real time across all devices.
Have a question, doubt, or thought about this post? Choose how you want to join the discussion below.
💬 Comment on Telegram — No account neededRequires a free GitHub account — takes 30 seconds to create with just an email address.