Order, fields and completeness

Try

Start with the interval of numbers strictly between 1 and 3. Every member is below 3, and 3 is the smallest number with that property. Yet 3 is not a member. This separates three ideas: an upper bound, a least upper bound and a maximum. Before working with infinite processes, analysis needs language that does not assume a boundary is included merely because it can be approached.

Predict

Predict which familiar algebraic facts distinguish rational numbers from real numbers. Adding, multiplying and dividing nonzero rational numbers stays rational; inequalities also obey the usual rules. Both Q and R are ordered fields. The difference needed for analysis concerns arbitrary bounded sets, not a new arithmetic operation. An ordered field need not contain the boundary suggested by every bounded collection of its elements.

Experiment

Compare these examples.

Set	One upper bound	Least upper bound in R	Maximum?
(1,3)	4	3	No
[1,3]	4	3	Yes
Negative integers	0	−1	Yes

For each row, explain why every smaller proposed upper bound fails. Checking that a number is above the set proves only the upper-bound half. Leastness is an additional universal statement about all competing upper bounds. Try replacing the negative integers by all negative real numbers and identify which columns change.

Understand

The completeness axiom says every nonempty subset of R that is bounded above has a least upper bound in R. Nonempty and bounded above are essential restrictions. The empty set has no least real upper bound because every real number is an upper bound; the whole real line has no upper bound.

Here is a complete uniqueness argument. Suppose s and t are both least upper bounds of A. Since t is an upper bound and s is least, s≤t. Reversing their roles gives t≤s. The order law of antisymmetry gives s=t. Existence comes from completeness; uniqueness comes from order. Combining these into one unexplained phrase would conceal two different logical ingredients.

Apply

The rational gap is concrete. Suppose a rational p/q in lowest terms squared to 2. Then p²=2q², so p is even. Writing p=2k gives q²=2k², so q is even too, contradicting lowest terms. Thus the positive rationals whose squares are below 2 approach a boundary absent from Q. To construct the real square root without assuming density, instead let A be the nonnegative real numbers with square below 2. This set contains 1 and is bounded by 2, so s=sup A exists with 1≤s≤2. If s²<2, choose 0<h≤min(1,(2−s²)/(2(2s+1))). Then (s+h)²≤s²+(2s+1)h<2, contradicting the upper bound s. If s²>2, choose 0<h≤min(s/2,(s²−2)/(4s)). Then (s−h)²≥s²−2sh≥(s²+2)/2>2. Every nonnegative member of A is therefore below s−h, making s−h a smaller upper bound. Both alternatives fail, so s²=2. Positivity makes the root unique. Later rational density identifies this same root as the supremum of the rational subset.

Repair “R is complete because every decimal has infinitely many digits.” Decimal notation requires its own existence and equivalence theory. It is a representation of the real numbers, not an independent proof of their least-upper-bound property.

Continue

Optional depth studies equivalent formulations of completeness, including convergence of bounded monotone sequences and the Cauchy property. Later lessons will derive these from the least-upper-bound axiom, making the dependency explicit instead of assuming all formulations simultaneously.

Write your own proof that multiplying a positive number by a negative number produces a negative result, using distributivity and order compatibility. Then label which steps use field axioms and which use order. This small audit prepares the central habit of analysis: specify the structure available, identify the claim being quantified, and avoid using a future theorem as an unstated premise.