This deck covers the ordered-field axioms and shows why the rationals form an ordered field that nevertheless has gaps. It introduces the supremum and infimum and the Least Upper Bound axiom, which pins down the real numbers as the unique complete ordered field, then derives the Archimedean property and the density of the rationals from completeness and constructs the reals by Dedekind cuts. It targets the traps of confusing a maximum with a supremum, assuming that the supremum lies in the set, and believing that the rationals are already complete.
Subject: Foundations of Higher Mathematics · 103 slides · symbolic lesson
Open the interactive version of this deck · Homework for this lesson
Objectives
By the end of this deck you will be able to:
1. State the ordered-field axioms and check that the rationals satisfy them.
2. Define upper bound, supremum, and infimum, and prove a given number is the supremum of a set.
3. State the Least Upper Bound axiom and explain why it fails for the rationals but characterizes the reals.
4. Derive the Archimedean property and the density of the rationals from completeness.
Warm-up
Discussion prompt
Before we open Ordered Fields & the Completeness Axiom: without looking back, what was the main idea of Invertibility, Isomorphism & Duality, and what could you do by the end of it that you could not do before?
Hint: One sentence for the idea, one for the skill. If the second one is blank, that is the part to revisit.
Answer:
That deck presents the Invertible Matrix Theorem as a single phenomenon with many faces, then covers isomorphism of vector spaces and their classification by dimension, the dual space and the dual basis, and the double dual with its natural isomorphism. It targets the traps of supposing that a non-square matrix could be invertible, that isomorphic spaces must share a common set, that the determinant is unrelated to injectivity, and that identifying a space with its dual is canonical.
Concept
A field is a set with two operations, addition and multiplication, in which you can add, subtract, multiply, and divide by anything nonzero, and the usual algebraic laws hold.
\( (F, +, \cdot) \)
Both operations are commutative and associative, multiplication distributes over addition, and there are identities and inverses.
field — A commutative ring in which every nonzero element has a multiplicative inverse. The rationals, the reals, and the complex numbers are all fields.
Counterexample
Discussion prompt
A field is a set with two operations, addition and multiplication, in which you can add, subtract, multiply, and divide by anything nonzero, and the usual algebraic laws hold.
That is stated as though it always holds. Do one of two things: produce a case where it fails, or say precisely what rules such a case out. "It just does" is not on the menu.
Hint: Hunt at the extremes first — zero, one, negative, empty, equal. If every extreme survives, the reason they survive is the proof.
Answer:
Both operations are commutative and associative, multiplication distributes over addition, and there are identities and inverses.
Concept
A total order is a way to compare any two elements so that exactly one of three cases holds: the first is less, they are equal, or the first is greater.
\[ a < b \quad \text{or} \quad a = b \quad \text{or} \quad a > b \]
This is called trichotomy. The order must also be transitive.
\( a < b \ \text{and}\ b < c \ \Rightarrow\ a < c \)
Analogy
Discussion prompt
Explain Next, a total order by analogy to something with no Foundations of Higher Mathematics in it at all — a queue, a recipe, a map, a bank balance, whatever fits. Then say where your analogy breaks.
Hint: An analogy that never breaks is not an analogy, it is the same idea wearing a hat. Find the seam — that is the part that is actually new.
Answer:
A total order is a way to compare any two elements so that exactly one of three cases holds: the first is less, they are equal, or the first is greater.
Concept
An ordered field is a field carrying a total order that is compatible with both operations. Two compatibility axioms tie the order to the arithmetic.
First, addition preserves order: you may add the same thing to both sides.
\( a < b \ \Rightarrow\ a + c < b + c \)
Second, multiplying by anything positive preserves order.
\( a < b \ \text{and}\ 0 < c \ \Rightarrow\ ac < bc \)
ordered field — A field together with a total order such that order is preserved by addition and by multiplication by positive elements. The rationals and the reals are ordered fields; the complex numbers cannot be one.
Definition probe
Sort into buckets
Every line below is part of the definition of field or of ordered field — one or the other, never both. Put each where it belongs.
Picture it
Figure (svg): A horizontal number line with zero marked, negatives to the left and positives to the right.
Discussion prompt
Read the picture before the words. What is this showing, and what is the one thing it is built to make obvious? Commit to an answer, then read on.
Hint: Name the parts, then say what changes between them — and if nothing changes, say what is being held still.
Answer:
An ordered field is exactly a field you can lay out on a straight line, with smaller elements to the left and larger to the right.
Intuition
An ordered field is exactly a field you can lay out on a straight line, with smaller elements to the left and larger to the right.
Figure (svg): A horizontal number line with zero marked, negatives to the left and positives to the right.
The two compatibility axioms are just the statements that sliding the whole line and stretching it by a positive factor both keep the left-to-right order intact.
Explain it
Discussion prompt
Explain Picture it as a number line to a student a year behind you. No notation, no jargon they have not met — and it still has to be true.
Hint: If your explanation needs a symbol they have never seen, you are describing the notation rather than the idea.
Answer:
An ordered field is exactly a field you can lay out on a straight line, with smaller elements to the left and larger to the right.
Concept
Every element is exactly one of: positive, zero, or negative. The positive elements form a set closed under addition and multiplication.
\[ a > 0,\quad a = 0,\quad \text{or}\quad a < 0 \]
This positive part is called the positive cone, and it alone determines the order: a is less than b precisely when the difference is positive.
\( a < b \iff b - a > 0 \)
Socratic
Discussion prompt
Every element is exactly one of: positive, zero, or negative. The positive elements form a set closed under addition and multiplication.
Suppose that were not true. What is the first thing in Ordered Fields & the Completeness Axiom that would stop working?
Hint: Follow it one step downstream. The answer is whatever was quietly relying on it.
Ranking
Put in order
Put the moves of Worked example: every nonzero square is positive into the order they have to happen.
Why: These are the moves of the worked example in the order it makes them, and each one is set up by the one before it. By trichotomy a nonzero element is either positive or negative, so two cases exhaust all possibilities.
Worked example
Claim: in any ordered field, the square of a nonzero element is strictly positive.
\( a \neq 0 \ \Rightarrow\ a^2 > 0 \)
Split into the two possible signs of a
Why: By trichotomy a nonzero element is either positive or negative, so two cases exhaust all possibilities.
If a is positive, multiply the inequality by a
Why: From a greater than 0 and a positive, the order-times-positive axiom gives a times a greater than 0.
\( a > 0 \ \Rightarrow\ a \cdot a > 0 \)
If a is negative, work with its negative
Why: Then the negative of a is positive, and the square of the negative of a equals the square of a.
\( (-a)^2 = a^2 > 0 \)
Conclude the square is positive in both cases
Why: Both branches give the same conclusion, so the claim holds for every nonzero a. A special case: since 1 equals 1 squared, 1 is positive.
\( 1 = 1^2 > 0 \)
Verify on a concrete value
Why: Take a equal to negative three in the rationals: its square is nine, which is indeed positive. The claim checks out.
\( (-3)^2 = 9 > 0 \)
Picture it
Animation
Shows: Each line of the worked example "every nonzero square is positive", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: Take a equal to negative three in the rationals: its square is nine, which is indeed positive. The claim checks out.
Concept
The rational numbers, with the usual addition, multiplication, and less-than, satisfy every ordered-field axiom.
\[ \mathbb{Q} = \left\{ \tfrac{p}{q} : p, q \in \mathbb{Z},\ q \neq 0 \right\} \]
So the rationals look like a complete algebraic and order-theoretic world. The surprise of this deck is that something crucial is still missing.
Concept
There is no rational number whose square is two. The point where such a number ought to sit is a gap in the rational line.
\[ \text{there is no } r \in \mathbb{Q} \text{ with } r^2 = 2 \]
The rationals crowd right up to that spot from both sides but never land on it. This missing point is the whole reason we need a bigger field.
Step zero
Discussion prompt
Worked example: no rational squares to two — before any calculation: what is the plan? Name the moves in order, in plain English, without doing the arithmetic.
Hint: It starts with: Assume a rational solution exists in lowest terms
Answer:
Worked example
Prove by contradiction that no rational number has square equal to two.
Assume a rational solution exists in lowest terms
Why: Write it as p over q with p and q sharing no common factor. Any rational can be reduced this way.
\( \left(\tfrac{p}{q}\right)^2 = 2,\quad \gcd(p,q) = 1 \)
Clear the denominator
Why: Multiplying both sides by q squared turns the equation into a statement about integers.
\( p^2 = 2q^2 \)
Deduce that p is even
Why: The right side is twice an integer, so p squared is even, which forces p itself to be even. Write p as twice k.
\( p = 2k \ \Rightarrow\ 4k^2 = 2q^2 \)
Deduce that q is even too
Why: Dividing by two gives q squared equals twice k squared, so q squared is even and hence q is even.
\( q^2 = 2k^2 \)
Verify the contradiction
Why: Both p and q are even, so they share the factor two, contradicting that the fraction was in lowest terms. No rational solution can exist.
Intuition
Mark on the rational line every number whose square is below two on the left, and every number whose square is above two on the right. The two families press together at a single missing point.
Figure (svg): A number line split into a left region of rationals with square below two and a right region with square above two, meeting at an empty circle where the square root of two would be.
In the rationals that meeting point is a genuine hole. In the reals it will be filled by a single new number. Completeness is the axiom that fills every such hole.
Concept
Fix a set of numbers. A number is an upper bound for the set if it sits at or above every element of the set.
\[ M \text{ is an upper bound for } A \iff (\forall x \in A)\ x \le M \]
A set can have many upper bounds. If one exists at all, so do infinitely many larger ones.
Socratic
Discussion prompt
A set can have many upper bounds. If one exists at all, so do infinitely many larger ones.
Suppose that were not true. What is the first thing in Ordered Fields & the Completeness Axiom that would stop working?
Hint: Follow it one step downstream. The answer is whatever was quietly relying on it.
Concept
A set is bounded above if it has at least one upper bound, and bounded below if it has at least one lower bound.
\[ (\forall x \in A)\ m \le x \le M \]
A set squeezed between a lower and an upper bound is simply called bounded. The lower-bound story mirrors the upper-bound story exactly.
Intuition
Think of an upper bound as a ceiling the set never pokes through. There are many possible ceilings, some far above the set and some pressed right down onto it.
The interesting question is the lowest ceiling that still works. That single tightest ceiling is the star of this deck.
Concept
The maximum of a set is an element of the set that is also an upper bound for it. It is the largest member, and it must actually belong to the set.
\[ \max A = M \iff M \in A \ \text{and}\ (\forall x \in A)\ x \le M \]
The catch: many perfectly nice bounded sets have no maximum at all, because the natural top edge is not itself a member.
Trap
Tempting claim: any set bounded above must have a largest element.
Consider all numbers strictly below one.
\( A = \{ x : 0 \le x < 1 \} \)
There is no largest element: whatever candidate you name below one, its average with one is larger and still below one.
\( x < \tfrac{x+1}{2} < 1 \)
The set is bounded above by one, but one is not a member, so it is not the maximum.
There is no maximum, yet there is clearly a tightest ceiling, namely one. That tightest ceiling is what we will call the supremum.
\( \text{no max, but a least upper bound of } 1 \)
Concept
The supremum of a set is its least upper bound: an upper bound that is smaller than or equal to every other upper bound.
It is fixed by two conditions. First, it really is an upper bound.
\( (\forall x \in A)\ x \le s \)
Second, nothing smaller works: any number below it is undercut by some element of the set.
\( (\forall \varepsilon > 0)(\exists x \in A)\ x > s - \varepsilon \)
supremum — The least upper bound of a set: the smallest number that is greater than or equal to every element of the set. It may or may not belong to the set.
Socratic
Discussion prompt
The supremum of a set is its least upper bound: an upper bound that is smaller than or equal to every other upper bound.
Suppose that were not true. What is the first thing in Ordered Fields & the Completeness Axiom that would stop working?
Hint: Follow it one step downstream. The answer is whatever was quietly relying on it.
Answer:
supremum: The least upper bound of a set: the smallest number that is greater than or equal to every element of the set. It may or may not belong to the set.
Concept
The infimum of a set is its greatest lower bound: the largest number sitting at or below every element.
\[ (\forall x \in A)\ x \ge i \quad \text{and} \quad (\forall \varepsilon > 0)(\exists x \in A)\ x < i + \varepsilon \]
Everything proved about suprema has a mirror statement about infima, obtained by flipping every inequality.
Intuition
Imagine lowering a ceiling as far as it will go without cutting into the set. The moment before it touches a point of the set is the supremum.
The second condition is the important one: it says the ceiling is pressed so low that any gap beneath it would already be occupied by a member of the set.
Pattern
1. Show your candidate is an upper bound
Why: Prove every element of the set is at or below the candidate. This is the easy half and rules out anything smaller being missed.
2. Show nothing smaller is an upper bound
Why: Given any positive gap below the candidate, produce an actual element of the set inside that gap. This forces the candidate to be the least upper bound.
3. State whether it is attained
Why: Check separately whether the supremum belongs to the set. If it does, it is also the maximum; if not, there is a supremum but no maximum.
Real world
Discussion prompt
Outside this lesson: where does Ordered Fields & the Completeness Axiom actually turn up? Name one concrete situation — a job, a piece of software someone ships, a decision somebody has to make — and say which part of Recipe: proving a number is the supremum is doing the work in it.
Hint: Vague is the failure mode here. "Engineering" is not a situation; "deciding whether this build is fast enough to ship" is.
Answer:
The ordered-field axioms, why the rationals form an ordered field with gaps, supremum and infimum, and the Least Upper Bound axiom that pins down the real numbers as the unique complete ordered field. Derives the Archimedean property and density of the rationals from completeness and constructs the reals via Dedekind cuts.
Estimation
Predict first
Find and prove the supremum and infimum of the set of all values one minus one over n as n ranges over the positive integers.
Commit before you compute: what does Worked example: the supremum of one minus one over n come out to? A rough magnitude and the right form is enough — the point is to have something concrete to be wrong about.
Correct: Verify the supremum is one and not attained
Why: A prediction you can defend turns the computation into a check rather than a leap of faith — and an answer that contradicts it is caught on the spot. One is an upper bound and no smaller number is, so the supremum is one.
Worked example
Find and prove the supremum and infimum of the set of all values one minus one over n as n ranges over the positive integers.
\( A = \left\{ 1 - \tfrac{1}{n} : n \in \mathbb{N},\ n \ge 1 \right\} \)
List the first few members
Why: Seeing the pattern makes the candidates obvious: the values start at zero and creep upward toward one without reaching it.
\( 0,\ \tfrac{1}{2},\ \tfrac{2}{3},\ \tfrac{3}{4},\ \ldots \)
Claim the infimum is zero and it is attained
Why: At n equal to one the value is zero, and every value is at least zero, so zero is a lower bound that belongs to the set. It is the minimum.
\( \inf A = 0 \in A \)
Show one is an upper bound
Why: Since one over n is strictly positive, one minus it is strictly below one for every n. So one caps the set.
\( 1 - \tfrac{1}{n} < 1 \)
Show nothing below one is an upper bound
Why: Given any positive gap, pick n larger than one over that gap. Then the value lands inside the gap, so no number below one can be a ceiling.
\( \tfrac{1}{n} < \varepsilon \ \Rightarrow\ 1 - \tfrac{1}{n} > 1 - \varepsilon \)
Verify the supremum is one and not attained
Why: One is an upper bound and no smaller number is, so the supremum is one. No n gives exactly one, so it is a supremum with no maximum. Check: at n equal to one hundred the value is 0.99, still below one.
\( \sup A = 1 \notin A \)
Picture it
Animation
Shows: Each line of the worked example "the supremum of one minus one over n", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: One is an upper bound and no smaller number is, so the supremum is one. No n gives exactly one, so it is a supremum with no maximum. Check: at n equal to one hundred the value is 0.99, still below one.
Check
Consider the set of all numbers one minus one over n for positive integers n. Read the definition of supremum carefully before answering.
Check your understanding
What is the supremum of this set?
Answer: A
Why: One is an upper bound since one minus one over n is always below one, and nothing smaller is an upper bound because the values approach one arbitrarily closely. So the supremum equals one even though it is never attained.
Concept
Here is the axiom that makes the real numbers what they are. It is called completeness, or the Least Upper Bound property.
Every nonempty set of reals that is bounded above has a supremum that is itself a real number.
\[ A \neq \varnothing,\ A \text{ bounded above} \ \Rightarrow\ \sup A \in \mathbb{R} \]
completeness axiom — The Least Upper Bound property: every nonempty subset of the reals that has an upper bound has a least upper bound within the reals. This single axiom distinguishes the reals from the rationals.
Intuition
Every place where a set presses up against a ceiling, completeness guarantees an actual number sitting exactly at that ceiling. There is nowhere for a hole to hide.
The rationals fail this: the set of rationals whose square is below two presses against a ceiling, but no rational ceiling is the least one. That failure is the gap we saw earlier.
Concept
Completeness is not automatic. It is a genuine extra axiom, and the rationals are a clean example of an ordered field where it fails.
\[ S = \{ x \in \mathbb{Q} : x > 0,\ x^2 < 2 \} \]
This set is nonempty and bounded above inside the rationals, yet it has no least upper bound among the rationals. We prove that next.
Socratic
Discussion prompt
Completeness is not automatic. It is a genuine extra axiom, and the rationals are a clean example of an ordered field where it fails.
Suppose that were not true. What is the first thing in Ordered Fields & the Completeness Axiom that would stop working?
Hint: Follow it one step downstream. The answer is whatever was quietly relying on it.
Step zero
Discussion prompt
Worked example: no rational supremum for the square-below-two set — before any calculation: what is the plan? Name the moves in order, in plain English, without doing the arithmetic.
Hint: It starts with: Suppose a rational supremum c exists
Answer:
Worked example
Prove the set of positive rationals with square below two has no supremum in the rationals.
\( S = \{ x \in \mathbb{Q} : x > 0,\ x^2 < 2 \} \)
Suppose a rational supremum c exists
Why: Aim for a contradiction. Since no rational squares to two, trichotomy leaves only two cases: c squared below two, or c squared above two.
\( c \in \mathbb{Q},\ c > 0,\ c^2 \neq 2 \)
Build a nudged candidate from c
Why: The map below produces a new rational whose square moves toward two, controlled by a clean algebraic identity.
\( c' = \frac{2c + 2}{c + 2},\qquad (c')^2 - 2 = \frac{2\,(c^2 - 2)}{(c+2)^2} \)
Case c squared below two: c is not even an upper bound
Why: Then the numerator is negative so the new square is still below two, and the new value is larger than c, so it is a member of S above c.
\( c^2 < 2 \ \Rightarrow\ c' > c \ \text{and}\ (c')^2 < 2 \)
Case c squared above two: c is not the least upper bound
Why: Then the new value is smaller than c but still has square above two, so it is a strictly smaller upper bound, beating c.
\( c^2 > 2 \ \Rightarrow\ c' < c \ \text{and}\ (c')^2 > 2 \)
Verify both cases contradict c being the supremum
Why: Either c fails to be an upper bound or fails to be the least one, so no rational c works. Check: for c equal to seven fifths the nudge gives forty-eight over thirty-four, whose square is about 1.993, larger than c and still below two.
Concept
Richard Dedekind's idea: instead of hunting for a missing number, define it as the split it would make in the rational line.
Figure (svg): The rational line partitioned into a lower set L on the left and an upper set U on the right, with the cut point between them representing a real number.
A cut is a partition of the rationals into a lower set with no greatest element and an upper set, where everything in the lower set is below everything in the upper set. Each cut is a real number.
Trap
Tempting claim: the rationals are so densely packed that they have no gaps, so they must be complete.
Density and completeness are different properties. The rationals are dense in themselves yet still miss suprema.
\( S = \{ x \in \mathbb{Q} : x^2 < 2 \} \ \text{has no rational } \sup \)
Being dense means between any two rationals there is another rational. That says nothing about limits of bounded sets.
Completeness is a separate, stronger demand about suprema, and the square-below-two set shows the rationals fail it. Completeness is a real axiom, not a free consequence of density.
Concept
The real numbers can be defined outright as the set of all Dedekind cuts of the rationals, ordered by inclusion of their lower sets.
\[ \mathbb{R} := \{\, \text{Dedekind cuts of } \mathbb{Q} \,\} \]
Addition and multiplication of cuts are defined so that the rationals embed faithfully, and one proves the result is an ordered field satisfying completeness.
Intuition
You never need to see the number itself. Knowing exactly which rationals fall below it pins it down completely.
The square root of two is simply the cut whose lower set holds every rational with square below two. The completeness of the reals is then a theorem about cuts, not a mystery.
Concept
There is a second standard construction. Take sequences of rationals whose terms eventually huddle arbitrarily close together, and call two of them equal when their difference shrinks to zero.
A real number becomes an equivalence class of such sequences.
\( \mathbb{R} := \{\text{Cauchy sequences in } \mathbb{Q}\} / \sim \)
This is the same completion idea used to fill in any metric space, and it yields an ordered field order-isomorphic to the cut construction.
Socratic
Discussion prompt
A real number becomes an equivalence class of such sequences.
Suppose that were not true. What is the first thing in Ordered Fields & the Completeness Axiom that would stop working?
Hint: Follow it one step downstream. The answer is whatever was quietly relying on it.
Concept
Both constructions land in the same place. Up to a unique order-preserving field isomorphism, there is exactly one complete ordered field.
\[ \mathbb{R} \ \text{is the unique complete ordered field (up to isomorphism)} \]
So completeness does not just add something to the rationals: it characterizes the reals entirely. Any complete ordered field is a relabeled copy of the reals.
Elimination
Eliminate the wrong options
Which property distinguishes the reals from the rationals?
3 of these 4 are wrong. Strike them one at a time, and say what rules each one out before you strike the next. The survivor is the answer.
Survives elimination: A
Why: The Least Upper Bound property is exactly completeness, which holds in the reals and fails in the rationals as the square-below-two set shows. The other three are ordered-field axioms already satisfied by the rationals.
Check
Both the rationals and the reals are ordered fields. Exactly one listed property holds in the reals but fails in the rationals.
Check your understanding
Which property distinguishes the reals from the rationals?
Answer: A
Why: The Least Upper Bound property is exactly completeness, which holds in the reals and fails in the rationals as the square-below-two set shows. The other three are ordered-field axioms already satisfied by the rationals.
Concept
Whenever a supremum exists, it is the only one. This is why we may speak of the least upper bound.
\[ s = \sup A \ \text{and}\ s' = \sup A \ \Rightarrow\ s = s' \]
The proof is a short two-line squeeze using the two defining properties, and it works in any ordered set.
Hypothesis
Predict first
Worked example: uniqueness of the supremum is about to be worked. State your hypothesis first: which rule or definition decides this one, and what is the first move it forces? Then watch whether the example agrees with you.
Correct: Assume two suprema s and s prime
Why: Suppose both are least upper bounds of the same set. We show they must coincide.
A hypothesis you wrote down is falsifiable; a vague sense of how it will go is not. If the example opens somewhere else, that gap is the thing worth chasing.
Worked example
Prove that a set can have at most one supremum.
Assume two suprema s and s prime
Why: Suppose both are least upper bounds of the same set. We show they must coincide.
Use that s prime is least against the upper bound s
Why: Since s is an upper bound and s prime is the least upper bound, s prime is at or below s.
\( s' \le s \)
Swap the roles
Why: By the same reasoning with the names exchanged, s is at or below s prime.
\( s \le s' \)
Verify the two are equal
Why: Two numbers each at or below the other must be equal by antisymmetry of the order. So the supremum is unique.
\( s' \le s \ \text{and}\ s \le s' \ \Rightarrow\ s = s' \)
Picture it
Animation
Shows: Each line of the worked example "uniqueness of the supremum", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: Two numbers each at or below the other must be equal by antisymmetry of the order. So the supremum is unique.
Concept
The reals contain no infinitely large numbers. The Archimedean property says the natural numbers are unbounded above inside the reals.
\[ (\forall x \in \mathbb{R})(\exists n \in \mathbb{N})\ n > x \]
Equivalently, for any positive real, no matter how tiny, some multiple of it exceeds any target you name. This is a consequence of completeness, not a separate assumption.
Intuition
The Archimedean property rules out two exotic possibilities at once: a number bigger than every whole number, and a positive number smaller than every fraction one over n.
The two are the same statement seen from opposite ends. If some positive number were below every one over n, its reciprocal would be above every n, so the naturals would be bounded.
Socratic
Discussion prompt
The Archimedean property rules out two exotic possibilities at once: a number bigger than every whole number, and a positive number smaller than every fraction one over n.
Suppose that were not true. What is the first thing in Ordered Fields & the Completeness Axiom that would stop working?
Hint: Follow it one step downstream. The answer is whatever was quietly relying on it.
Ranking
Put in order
Put the moves of Worked example: Archimedean property from completeness into the order they have to happen.
Why: These are the moves of the worked example in the order it makes them, and each one is set up by the one before it. For contradiction, suppose some real caps every natural number.
Worked example
Prove from the Least Upper Bound axiom that the natural numbers are unbounded above in the reals.
Assume the naturals are bounded above
Why: For contradiction, suppose some real caps every natural number. Then the set of naturals is nonempty and bounded above.
\( \mathbb{N} \text{ bounded above in } \mathbb{R} \)
Invoke completeness to get a supremum
Why: By the Least Upper Bound axiom the naturals then have a real supremum s. This is where completeness does the work.
\( s = \sup \mathbb{N} \)
Push just below the supremum
Why: Since s is the least upper bound, s minus one is not an upper bound, so some natural n exceeds it.
\( \exists n \in \mathbb{N},\ n > s - 1 \)
Step up by one to break the bound
Why: Then n plus one is also a natural number and exceeds s, contradicting that s bounds the naturals.
\( n + 1 > s \)
Verify the contradiction settles it
Why: The supremum cannot be exceeded by a member of the set, so no bound exists. Check: with x equal to one thousand, the natural n equal to one thousand and one already beats it.
Picture it
Animation
Shows: Each line of the worked example "Archimedean property from completeness", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: The supremum cannot be exceeded by a member of the set, so no bound exists. Check: with x equal to one thousand, the natural n equal to one thousand and one already beats it.
Trap
Tempting move: once you find the supremum, treat it as an element of the set and plug it back in.
For the open interval below one, the supremum is one, but one is not in the set.
\( \sup \{x : 0 \le x < 1\} = 1 \notin \{x : 0 \le x < 1\} \)
The supremum is a least upper bound, an outside boundary. It belongs to the set only when the set actually attains its maximum.
Always check membership separately. If the supremum is a member it equals the maximum; if not, the set has a supremum but no largest element.
\( \sup A \in A \iff \max A \text{ exists and equals } \sup A \)
Break the constraint
Discussion prompt
The rule this trap just fixed:
The supremum is a least upper bound, an outside boundary. It belongs to the set only when the set actually attains its maximum.
Now break it on purpose. Build a case that violates it and follow the consequences until something visibly fails. Where does the failure first show up — and would you have noticed it if you had not been looking?
Hint: The dangerous rules are the ones whose violation still produces an answer. If yours fails loudly, try to find one that fails quietly.
Concept
Between any two distinct reals, no matter how close, there is a rational number. This is the density of the rationals in the reals.
\[ a < b \ \Rightarrow\ (\exists\, q \in \mathbb{Q})\ a < q < b \]
It is another payoff of the Archimedean property: you can always find a denominator fine enough that some multiple lands strictly between the two.
Intuition
No matter how tightly you zoom in on the real line, you still see infinitely many rational numbers inside the window.
This is why decimal approximations work: every real is surrounded arbitrarily closely by rationals you can actually write down. The irrationals fill the gaps, but the rationals are never far away.
Step zero
Discussion prompt
Worked example: a rational between any two reals — before any calculation: what is the plan? Name the moves in order, in plain English, without doing the arithmetic.
Hint: It starts with: Choose a fine enough denominator
Answer:
Worked example
Given reals with the first strictly below the second, construct a rational strictly between them.
\( a < b \)
Choose a fine enough denominator
Why: By the Archimedean property pick a natural q large enough that one over q is smaller than the gap, so q times the gap exceeds one.
\( q \in \mathbb{N},\quad q\,(b - a) > 1 \)
Pick the first integer above q times a
Why: Let p be the least integer strictly greater than q times a. Such an integer exists by the Archimedean property applied to the naturals.
\( p = \lfloor q a \rfloor + 1,\qquad p > q a \)
Show p stays below q times b
Why: Since p is at most q times a plus one and q times b exceeds q times a plus one, p is below q times b.
\( p \le q a + 1 < q b \)
Divide through by q
Why: Dividing the double inequality by the positive q lands the rational p over q strictly between a and b.
\( a < \frac{p}{q} < b \)
Verify on a concrete window
Why: Take a equal to 0.1 and b equal to 0.2. Then q equal to eleven works, p equals two, and two over eleven is about 0.182, which sits inside the window.
\( 0.1 < \tfrac{2}{11} < 0.2 \)
Picture it
Animation
Shows: Each line of the worked example "a rational between any two reals", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: Take a equal to 0.1 and b equal to 0.2. Then q equal to eleven works, p equals two, and two over eleven is about 0.182, which sits inside the window.
Prediction
Predict first
How many rational numbers lie strictly between 0.333 and 0.334?
Answer it in your own words, now, with nothing to choose from. The options are on the next slide — and picking the right one off a list is an easier skill than producing it.
Correct: Infinitely many
Why: Density says between any two distinct reals there is a rational, and you can then repeat the argument inside each resulting subinterval forever, producing infinitely many rationals in any open interval.
Check
Consider the open interval between 0.333 and 0.334 on the real line.
Check your understanding
How many rational numbers lie strictly between 0.333 and 0.334?
Answer: A
Why: Density says between any two distinct reals there is a rational, and you can then repeat the argument inside each resulting subinterval forever, producing infinitely many rationals in any open interval.
Concept
Take a chain of closed intervals, each sitting inside the previous one, whose lengths shrink toward zero. Completeness guarantees they share exactly one common point.
\[ I_1 \supseteq I_2 \supseteq I_3 \supseteq \cdots,\qquad \bigcap_{n=1}^{\infty} I_n = \{ c \} \]
This nested interval property is another equivalent face of completeness, and it is the engine behind bisection search and binary refinement.
Intuition
Picture two markers, one crawling right along the left endpoints and one crawling left along the right endpoints. They can never cross, and the gap between them vanishes.
Completeness is exactly what promises a real number waiting at the meeting spot. In the rationals the markers could squeeze down onto a hole and meet nothing.
Socratic
Discussion prompt
Picture two markers, one crawling right along the left endpoints and one crawling left along the right endpoints. They can never cross, and the gap between them vanishes.
Suppose that were not true. What is the first thing in Ordered Fields & the Completeness Axiom that would stop working?
Hint: Follow it one step downstream. The answer is whatever was quietly relying on it.
Answer:
Completeness is exactly what promises a real number waiting at the meeting spot. In the rationals the markers could squeeze down onto a hole and meet nothing.
Estimation
Predict first
Show a nested sequence of closed intervals with lengths tending to zero has exactly one common point.
Commit before you compute: what does Worked example: nested intervals pin a single point come out to? A rough magnitude and the right form is enough — the point is to have something concrete to be wrong about.
Correct: Verify on a concrete chain
Why: A prediction you can defend turns the computation into a check rather than a leap of faith — and an answer that contradicts it is caught on the spot. For the intervals from zero to one over n, the only common point is zero: it lies in every interval, and any positive number is excluded once one over n drops below it.
Worked example
Show a nested sequence of closed intervals with lengths tending to zero has exactly one common point.
\( I_n = [a_n, b_n],\quad a_n \le a_{n+1} \le b_{n+1} \le b_n \)
Bound the left endpoints above
Why: Every right endpoint bounds every left endpoint because the intervals are nested. So the set of left endpoints is bounded above by b one.
\( a_n \le b_1 \ \text{for all } n \)
Take the supremum of the left endpoints
Why: By completeness the left endpoints have a real supremum c, which lies at or above every left endpoint and at or below every right endpoint.
\( c = \sup_n a_n,\qquad a_n \le c \le b_n \)
Conclude c is in every interval
Why: Since c is trapped between the two endpoints of each interval, it belongs to all of them at once.
\( c \in \bigcap_{n} I_n \)
Use shrinking lengths for uniqueness
Why: If two points were common, the interval lengths could never drop below their distance, contradicting lengths going to zero.
Verify on a concrete chain
Why: For the intervals from zero to one over n, the only common point is zero: it lies in every interval, and any positive number is excluded once one over n drops below it.
\( \bigcap_{n} \left[0, \tfrac{1}{n}\right] = \{0\} \)
Picture it
Animation
Shows: Each line of the worked example "nested intervals pin a single point", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: For the intervals from zero to one over n, the only common point is zero: it lies in every interval, and any positive number is excluded once one over n drops below it.
Concept
Now the payoff. The very supremum the rationals could not supply, the reals do, and it is exactly the number whose square is two.
\[ s = \sup \{ x \ge 0 : x^2 < 2 \} \ \Rightarrow\ s^2 = 2 \]
So the square root of two is not conjured from nowhere. It is the least upper bound that completeness guarantees, and we can prove its square is exactly two.
Step zero
Discussion prompt
Worked example: the supremum squares to two — before any calculation: what is the plan? Name the moves in order, in plain English, without doing the arithmetic.
Hint: It starts with: Rule out s squared below two
Answer:
Worked example
Let s be the supremum of the nonnegative reals with square below two. Prove the square of s is two by eliminating the other two cases.
\( s = \sup \{ x \ge 0 : x^2 < 2 \} \)
Rule out s squared below two
Why: Pick a small positive h below one with h under the quantity two minus s squared over two s plus one. Then s plus h still has square below two, so it belongs to the set above s, contradicting that s is an upper bound.
\( (s+h)^2 < s^2 + h(2s+1) < 2 \)
Rule out s squared above two
Why: Take h equal to s squared minus two over two s. Then s minus h has square above two, so it is a smaller upper bound, contradicting that s is the least one.
\( (s-h)^2 > s^2 - 2sh = 2 \)
Conclude the square equals two
Why: Trichotomy leaves only the middle case standing, so the square of s must be exactly two.
\( s^2 = 2 \)
Verify the numerical value
Why: The supremum is the positive square root of two, roughly 1.41421, and its square 1.99998 rounds to two while it clearly sits between one and two.
\( 1 < s < 2,\qquad s = \sqrt{2} \)
Picture it
Animation
Shows: Each line of the worked example "the supremum squares to two", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: The supremum is the positive square root of two, roughly 1.41421, and its square 1.99998 rounds to two while it clearly sits between one and two.
Trap
Tempting shortcut: write maximum wherever you mean the top of a set, as if the two words were interchangeable.
A bounded nonempty set of reals always has a supremum, but it need not have a maximum, so a proof that quietly uses the maximum can be appealing to something that does not exist.
\( \{ x : 0 \le x < 1 \}:\ \sup = 1,\ \max \text{ undefined} \)
Reach for the supremum, which completeness always provides for bounded nonempty sets, and only call it a maximum after checking it is attained.
When the top edge is a member the two agree; when it is not, only the supremum exists. Using the supremum keeps every argument valid.
\( \max A \text{ exists} \ \Rightarrow\ \max A = \sup A \)
Two truths and a lie
Sort into buckets
Some of these hold up and some are the exact mistakes this lesson is built to prevent. Sort them.
Concept
Suprema interact cleanly with addition of sets. Form the set of all sums of one element from each of two bounded sets; its supremum is the sum of the two suprema.
\[ \sup (A + B) = \sup A + \sup B \]
Here the sum of two sets means the collection of every pairwise sum. This identity is a favourite exam problem and a clean use of the two-part definition.
Ranking
Put in order
Put the moves of Worked example: the supremum of a sum of sets into the order they have to happen.
Why: These are the moves of the worked example in the order it makes them, and each one is set up by the one before it. Any element of the sumset is x plus y with x at or below the supremum of A and y at or below the supremum of B, so their sum is capped by the total.
Worked example
Prove the supremum of the sumset equals the sum of the suprema, for bounded nonempty sets.
\( A + B = \{ x + y : x \in A,\ y \in B \} \)
Show the sum of suprema is an upper bound
Why: Any element of the sumset is x plus y with x at or below the supremum of A and y at or below the supremum of B, so their sum is capped by the total.
\( x + y \le \sup A + \sup B \)
This gives one inequality
Why: Being an upper bound, the sum of suprema is at least the least upper bound of the sumset.
\( \sup(A+B) \le \sup A + \sup B \)
Approach each supremum from below
Why: Given a positive gap, choose x within half the gap of the supremum of A and y within half the gap of the supremum of B. Their sum beats the total minus the gap.
\( x + y > \sup A + \sup B - \varepsilon \)
This gives the reverse inequality
Why: Since the sumset gets within any gap of the total, its supremum is at least the sum of the suprema.
\( \sup(A+B) \ge \sup A + \sup B \)
Verify with concrete intervals
Why: Take A and B as the half-open intervals up to two and up to three. Their sumset reaches up to five, and indeed two plus three equals five, matching the identity.
\( \sup([0,2) + [0,3)) = 5 = 2 + 3 \)
Picture it
Animation
Shows: Each line of the worked example "the supremum of a sum of sets", appearing one at a time.
The same working the example does, in the order a tutor would write it.
Takeaway: Take A and B as the half-open intervals up to two and up to three. Their sumset reaches up to five, and indeed two plus three equals five, matching the identity.
Check
Let A be the half-open interval from zero up to but not including two, and B the half-open interval from zero up to but not including three. Form the set of all sums of one point from each.
Check your understanding
What is the supremum of the sumset A plus B?
Answer: A
Why: The supremum of A is two and the supremum of B is three, and suprema add, so the supremum of the sumset is two plus three, which is five, even though five is never attained.
Concept
A computer's floating point numbers are a finite set. Like the rationals, they are riddled with gaps, and unlike the reals they are not even dense.
Between two adjacent floating point values there is nothing representable at all, so most reals get rounded to a nearby machine number.
This is why a stored value squared can fail to equal two exactly, and why numerical code must reason about rounding rather than assume true real arithmetic.
Intuition
Think of floating point as a ruler with only finitely many tick marks, spaced ever wider as the numbers grow. Anything between two ticks simply snaps to the nearest one.
Completeness is precisely the property this ruler lacks. Understanding the axiom tells you exactly where computation and true arithmetic part ways.
Concept
Exact-real libraries dodge rounding by representing a number as a program that, on demand, delivers a rational approximation good to any requested precision.
A real is thus modelled by its stream of ever-tighter rational bounds, which is the Cauchy-sequence picture turned into code.
Only countably many reals can be described by such programs. The completeness of the reals guarantees that the rest, the uncomputable ones, are still there.
Socratic
Discussion prompt
Exact-real libraries dodge rounding by representing a number as a program that, on demand, delivers a rational approximation good to any requested precision.
Suppose that were not true. What is the first thing in Ordered Fields & the Completeness Axiom that would stop working?
Hint: Follow it one step downstream. The answer is whatever was quietly relying on it.
Answer:
A real is thus modelled by its stream of ever-tighter rational bounds, which is the Cauchy-sequence picture turned into code.
Concept
Every headline theorem of calculus leans on completeness. Without it, bounded increasing sequences need not converge and continuous functions can skip over values.
A continuous function crossing from negative to positive must hit zero precisely because the reals have no gap for it to jump through.
So completeness is not an afterthought. It is the structural bedrock that makes limits, continuity, and the whole of analysis behave.
Explain it
Discussion prompt
Explain Why analysis needs completeness to a student a year behind you. No notation, no jargon they have not met — and it still has to be true.
Hint: If your explanation needs a symbol they have never seen, you are describing the notation rather than the idea.
Answer:
Every headline theorem of calculus leans on completeness. Without it, bounded increasing sequences need not converge and continuous functions can skip over values.
Check
Consider the set of all real numbers that are at least zero and strictly less than one.
Check your understanding
Which statement about this set is correct?
Answer: A
Why: One is the least upper bound because the set climbs arbitrarily close to it, but one is excluded from the set, so there is no largest element and hence no maximum.
Concept
The single completeness axiom can be stated in several logically equivalent ways, each convenient for different theorems.
Every nonempty bounded-above set has a least upper bound; every bounded monotone sequence converges; every nested chain of shrinking closed intervals has a common point.
Choosing the right face for the problem in front of you is a core skill of real analysis. They all say the same thing: the real line has no holes.
Analogy
Discussion prompt
Explain Completeness wears three equivalent faces by analogy to something with no Foundations of Higher Mathematics in it at all — a queue, a recipe, a map, a bank balance, whatever fits. Then say where your analogy breaks.
Hint: An analogy that never breaks is not an analogy, it is the same idea wearing a hat. Find the seam — that is the part that is actually new.
Answer:
The single completeness axiom can be stated in several logically equivalent ways, each convenient for different theorems.
Elimination
Eliminate the wrong options
Is there a positive real number smaller than one over n for every positive integer n?
3 of these 4 are wrong. Strike them one at a time, and say what rules each one out before you strike the next. The survivor is the answer.
Survives elimination: A
Why: If such a positive number existed, its reciprocal would exceed every natural number, contradicting the Archimedean property that the naturals are unbounded, which follows from completeness.
Check
Think about whether the reals can contain a positive number smaller than every fraction of the form one over n.
Check your understanding
Is there a positive real number smaller than one over n for every positive integer n?
Answer: A
Why: If such a positive number existed, its reciprocal would exceed every natural number, contradicting the Archimedean property that the naturals are unbounded, which follows from completeness.
Concept
The ordered-field axioms give arithmetic and order; the rationals satisfy them yet leave gaps; the completeness axiom fills every gap and forces a unique field, the reals.
From that one axiom flow the Archimedean property, density of the rationals, nested intervals, and the existence of roots. This is the foundation the next decks on sequences, series, and continuity are built on.
Counterexample
Discussion prompt
The ordered-field axioms give arithmetic and order; the rationals satisfy them yet leave gaps; the completeness axiom fills every gap and forces a unique field, the reals.
That is stated as though it always holds. Do one of two things: produce a case where it fails, or say precisely what rules such a case out. "It just does" is not on the menu.
Hint: Hunt at the extremes first — zero, one, negative, empty, equal. If every extreme survives, the reason they survive is the proof.
Connect it up
Draw it
One page, no notation unless you need it: draw how these connect — Recipe: proving a number is the supremum · First, recall what a field is · Next, a total order · An ordered field marries the two · Picture it as a number line. Put an arrow wherever one of them is what makes another possible, and label the arrow with why.
Recap
You can state the ordered-field axioms, recognize that the rationals satisfy them but have gaps, and prove specific bounded sets of rationals have no rational supremum.
You can define supremum and infimum, prove a number is the least upper bound using the two-part recipe, and keep maximum and supremum carefully distinct.
You can state the completeness axiom, explain why it characterizes the reals as the unique complete ordered field, and derive the Archimedean property and density of the rationals from it.
| Idea | What it says |
|---|---|
| Ordered field | field plus a compatible total order |
| Supremum | least upper bound, may not be attained |
| Completeness | every bounded nonempty set has a supremum |
| Archimedean | the naturals are unbounded in the reals |
Want this taught 1-on-1? Alexander tutors Foundations of Higher Mathematics — $55/session, free consultation.