← mapBedrock

Order & Inequality

⚗ Dr. Möbius, from the lab

Equations get all the glory, but inequalities run the real world — every "at least," "at most," "no more than," every optimization, every bound in every proof ever written. They come with rules, and one of those rules makes students lose their damn minds: multiply both sides by a negative and the sign flips. Today I prove to you exactly why, with the number line, so you never have to fear it or get it wrong again. This is one of those things I actually enjoyed deriving in the lab.

THE BIG IDEA

The number line is an ordered structure governed by trichotomy and a handful of provable manipulation rules — the only surprising one being that multiplying by a negative reverses the inequality, because negation flips the line.

Order is structure, not decoration

Back in What Is a Number? we said a<ba < b means you count through aa on the way to bb — equivalently, you can add a positive amount to aa to reach bb. That single idea, extended to all of Q\mathbb{Q}, makes the number line an ordered thing: every number sits in a definite spot, left-to-right.

We're going to nail down the rules of this order precisely, because in the Algebra stratum you'll solve inequalities constantly, and a wrong sign-flip there poisons everything downstream. I have graded enough work to know that "I thought the sign flip was optional" is a sentence that should not exist in the wild.

Trichotomy: exactly one of three

The foundation. Take any two numbers aa and bb. Then exactly one of these holds:

a<b,a=b,a>b.a < b, \qquad a = b, \qquad a > b.

This is the law of trichotomy ("three-cut"). Not "at least one" — exactly one. Two numbers can't be both less and equal; if neither is bigger, they're equal. It sounds trivial but it's absolutely not — it's the bedrock guarantee that the line has no ambiguous spots, no two numbers in a quantum superposition of order. Every number has a definite place relative to every other, and that's a hell of a thing to guarantee about an infinite set.

Drag the point and watch where it sits relative to the others — it's always strictly left, strictly right, or dead on:

number line
-10-8-6-4-202468105
x = 5−x = -5|x| = 5

Adding to both sides: order survives

First manipulation rule: you can add (or subtract) the same thing to both sides without disturbing the order.

If a<b, then a+c<b+c for any c.\text{If } a < b, \text{ then } a + c < b + c \text{ for any } c.

Why? From the first-principles definition, a<ba < b means bab - a is positive. Adding cc to both sides: (b+c)(a+c)=ba(b+c) - (a+c) = b - a, the same positive number. The gap between them is unchanged, so the order is unchanged. Picture it: adding cc slides both points the same distance the same direction — their relative position can't flip. 3<73 < 7, add 1010: 13<1713 < 17. Add 100-100: 97<93-97 < -93. Still true.

Multiplying by a positive: fine. By a negative: FLIP.

Second rule, and here's the famous one. Multiplying both sides by a number scales the gap — but the sign of that number decides everything.

Positive multiplier preserves order:

If a<b and c>0, then ac<bc.\text{If } a < b \text{ and } c > 0, \text{ then } ac < bc.

Why? bab - a is positive, and a positive times a positive is positive (from Zero & the Negatives), so c(ba)=cbcac(b-a) = cb - ca is positive, meaning ca<cbca < cb. Scaling by a positive stretches the gap but keeps its direction. 2<52 < 5, times 33: 6<156 < 15. Good.

Negative multiplier REVERSES order:

If a<b and c<0, then ac>bc.\text{If } a < b \text{ and } c < 0, \text{ then } ac > bc.

Why? Now cc is negative. bab - a is positive, but a negative times a positive is negative (again from Zero & the Negatives), so c(ba)=cbcac(b-a) = cb - ca is negative, which means cb<cacb < ca, i.e. ac>bcac > bc. The gap flipped sign, so the order flipped.

The picture is gorgeous — honestly I think about this in the shower: multiplying by 1-1 reflects the entire number line across 00. Whatever was on the left lands on the right. 2<52 < 5, but multiply by 1-1 and reflect: 2>5-2 > -5. (2-2 is indeed to the right of 5-5.) The flip isn't some weird arbitrary rule bolted on — it's what reflection does. Reflect a line and left/right swap. That's the whole story.

Chaining: transitivity

Third rule, the one that lets you build long arguments: order chains.

If a<b and b<c, then a<c.\text{If } a < b \text{ and } b < c, \text{ then } a < c.

This is transitivity. If aa is left of bb and bb is left of cc, then aa is left of cc — you can't get back to the right by going left twice. From first principles: ba>0b-a > 0 and cb>0c-b > 0, and the sum of two positives is positive, so ca=(cb)+(ba)>0c - a = (c-b)+(b-a) > 0, giving a<ca < c. This is what makes a chain like 2<5<9<1002 < 5 < 9 < 100 meaningful and lets you conclude 2<1002 < 100 at a glance.

Intervals: naming chunks of the line

When a condition picks out a whole stretch of the line, we name it with interval notation:

  • (a,b)(a, b) — all xx with a<x<ba < x < b. Round brackets = open = endpoints excluded.
  • [a,b][a, b] — all xx with axba \le x \le b. Square brackets = closed = endpoints included.
  • [a,b)[a, b) — half-open: includes aa, excludes bb.
  • (a,)(a, \infty) — a ray: all x>ax > a. Infinity always gets a round bracket — you never "reach" it, so it's never included.

So "xx is at least 33" is [3,)[3, \infty), and "1<x4-1 < x \le 4" is (1,4](-1, 4]. The bracket shape is the entire grammar: round excludes, square includes. Memorize that and interval notation is free.

The tiny inequality that runs mathematics

Last, and don't let its size fool you — this is one of the most-used facts in all of mathematics:

a20for every real number a.a^2 \ge 0 \quad \text{for every real number } a.

A square is never negative. Why? Trichotomy splits it into cases. If a>0a > 0, then a2=aaa^2 = a\cdot a is positive times positive = positive. If a<0a < 0, then a2a^2 is negative times negative = positive (the proved rule from Zero & the Negatives). If a=0a = 0, then a2=0a^2 = 0. In every case a20a^2 \ge 0, with equality only when a=0a = 0.

It looks like nothing. It's not nothing — this tiny bastard is the seed of the quadratic formula's discriminant, the proof that distances are nonnegative, the AM–GM inequality, and a hundred theorems you'll meet later. The smallest facts often carry the most weight. File it away — it detonates everywhere. Go run the gauntlet.

🔬 SPECIMENS (worked examples)

Worked example 1 — preserve the order

Given 4<9-4 < 9, what happens to the inequality when you add 77 to both sides? When you multiply both sides by 55?

Add 77 to both sides — adding preserves order (slides both points the same way):

4+7<9+7    3<16.-4 + 7 < 9 + 7 \;\Longrightarrow\; 3 < 16. \checkmark

Multiply both sides by 5555 is positive, so order is preserved (the gap stretches but doesn't flip):

45<95    20<45.-4 \cdot 5 < 9 \cdot 5 \;\Longrightarrow\; -20 < 45. \checkmark

Both stay "<<" because nothing here reflects the line — we only slid and stretched in the positive direction. No flip required.

Worked example 2 — the flip in action

Solve the inequality 3x12-3x \le 12 for xx, and write the solution in interval notation.

To isolate xx, divide both sides by 3-3. The divisor is negative, so the inequality reverses — this is the reflection of the line across 00:

3x12    x123    x4.-3x \le 12 \;\Longrightarrow\; x \ge \frac{12}{-3} \;\Longrightarrow\; x \ge -4.

(Notice \le became \ge.) Now translate to interval notation. "x4x \ge -4" is everything from 4-4 rightward, including 4-4 itself, so a square bracket on 4-4 and a round bracket on \infty:

[4,).[-4, \infty).

Quick check with x=0x = 0 (which is 4\ge -4): 3(0)=012-3(0) = 0 \le 12. True. And x=10x=-10 (not in the set): 3(10)=3012-3(-10) = 30 \le 12? False — correctly excluded. The flip was real.

Worked example 3 — the trap: squares and a fake counterexample

A student claims "a20a^2 \ge 0 can't always be true — what about a=5a = -5? That's negative!" Diagnose the error and confirm the inequality.

The student confused aa with a2a^2. The claim is about the square, not the input. Compute it:

(5)2=(5)(5)=250.(-5)^2 = (-5)(-5) = 25 \ge 0. \checkmark

A negative input produces a positive square, because negative times negative is positive (proved in Zero & the Negatives). The inequality a20a^2 \ge 0 never said aa was nonnegative — it said the result of squaring is.

Confirm by trichotomy across all cases:

  • a>0a > 0: a2=(pos)(pos)>0a^2 = (\text{pos})(\text{pos}) > 0.
  • a<0a < 0: a2=(neg)(neg)>0a^2 = (\text{neg})(\text{neg}) > 0.
  • a=0a = 0: a2=0a^2 = 0.

Every case lands in a20a^2 \ge 0, equality only at a=0a = 0. The "counterexample" was just a misread of what's being squared.

☠ KNOWN HAZARDS

  • Forgetting to flip when multiplying or dividing by a negative. 2x<6-2x < 6 becomes x>3x > -3, NOT x<3x < -3. Dividing both sides by 2-2 reflects the line — the sign must reverse. This is the number-one inequality error in existence; the reflection picture is your antidote. Burn it in.

  • Mixing up open and closed brackets. (2,5)(2,5) excludes 22 and 55; [2,5][2,5] includes them. "At most 55" needs the square bracket \le; "less than 55" needs the round one. The bracket shape IS the strict-vs-nonstrict distinction. Swapping them changes the answer.

  • Writing [a,][a,\infty]. Infinity is never reached, so it's never included — always (a,)(a,\infty) with a round bracket. A square bracket on \infty is a syntax crime and I will circle it in red.

  • Believing a20a^2 \ge 0 needs aa to be positive. It holds for every real aa — that's the whole damn point. Negative inputs square to positives. The only way to hit equality is a=0a=0.

TL;DR

  • Trichotomy: for any a,ba,b, exactly one of a<ba<b, a=ba=b, a>ba>b holds — the line has no ambiguous spots.

  • Adding the same amount to both sides preserves order (the gap bab-a is unchanged).

  • Multiply by a positive: order preserved. Multiply by a negative: order REVERSES — because ×(1)\times(-1) reflects the line across 00, swapping left and right.

  • Order is transitive (a<b<ca<ca<b<c \Rightarrow a<c), which is what makes inequality chains work. Intervals: ()() excludes endpoints, [][] includes; \infty always gets ()().

  • a20a^2 \ge 0 for every real aa, with equality only at a=0a=0 — tiny, provable by cases, and the most-used inequality in mathematics.

unlocks