The Families of Numbers

You probably met "numbers" as one thing: symbols you count, add, and multiply with. Mathematicians see something else — families. Each family was born for a reason. Something broke. An operation refused to give an answer, so people invented new numbers to fix it.

That's this whole phase. Not "memorize five definitions" but one thread: the number system kept growing because math kept needing it to. Once you see the thread, the names stop being trivia and become inevitable.

The driving idea: numbers grew to keep math working

Here's the mental model to carry through everything below.

Every family starts as "the numbers we already have." Then someone asks a fair question — a subtraction, a division, a length — and the honest answer is: there is no number here that works. A math where reasonable questions have no answer feels broken.

So people widened the system instead of accepting the gap. They invented new numbers so the question did have an answer. Each widening keeps everything you already had and adds what was missing. Nothing gets thrown away.

We'll do this four times. Watch the same move repeat.

Naturals (ℕ): the counting numbers

Start with the most ancient numbers — the ones you'd use to count sheep:

0, 1, 2, 3, 4, 5, ...

These are the natural numbers, written ℕ. (Some books start them at 1, some at 0. We'll include 0 here; it's a convention, not a deep truth.) They go up forever.

Addition is happy here: any two naturals add to another natural. Multiplication too. So far the system holds together.

Then someone asks: what is 3 − 5?

The answer should be "two below zero." But there is no natural number below zero. Within ℕ, 3 − 5 has no answer at all. Subtraction — an ordinary operation — is allowed to fail.

That gap is the reason for the next family.

Integers (ℤ): add the negatives

To make subtraction always work, we add mirror-image numbers below zero:

..., -3, -2, -1, 0, 1, 2, 3, ...

These are the integers, written ℤ (from the German Zahlen, "numbers"). Now 3 − 5 = -2 has a real answer. In fact any subtraction of integers gives an integer. The hole is patched.

We didn't lose the naturals. Every natural number is still here — 4 is both a natural and an integer. We only extended the system downward.

Then someone asks the next fair question: what is 1 ÷ 2?

No integer, doubled, gives 1. The nearest integers are 0 and 1, and neither works. Within ℤ, 1 ÷ 2 has no answer. Division — again, an ordinary operation — fails.

Same move, one more time.

Rationals (ℚ): ratios of integers

To make division work (almost always — we'll get to the catch), we allow one integer divided by another:

1/2,  -3/4,  7/1,  22/7,  0,  5

These are the rationals, written ℚ (for quotient). A rational is any number you can write as a fraction a/b where a and b are integers and b is not zero. That last rule matters: division by zero stays undefined even here. 1 ÷ 2 is now 1/2. The hole is patched again.

And once more, the old families survive. Every integer is a rational: 5 is 5/1. Every natural is too.

Rationals also have a decimal personality worth knowing. Write a rational as a decimal and it always does one of two things:

• It terminates: 1/2 = 0.5, 3/8 = 0.375.
• Or it repeats forever in a pattern: 1/3 = 0.333..., 1/7 = 0.142857142857...

Terminating or repeating — those are the only options for a rational. Hold onto that, because it's how we'll catch the numbers that aren't rational.

So have we got every number? It feels like fractions should cover everything. They don't, and the reason comes not from arithmetic but from geometry.

Irrationals: the numbers no fraction can reach

Draw a square one unit on each side. Its diagonal has a length. You can measure it, point at it, build it — it's a real distance. By the Pythagorean relationship that length is √2 (the number that, squared, gives 2).

Now the shock: √2 cannot be written as any fraction a/b. Not because we haven't found the right one yet — it's been proven that no such fraction exists. The ancient Greeks discovered this, and it genuinely unsettled them: fractions, their whole notion of number, had a hole.

The circle constant π is the same story. The ratio of a circle's circumference to its diameter is a perfectly real number, but it is provably not a fraction either.

Numbers like √2 and π are irrational: they cannot be written as a ratio of integers. As decimals, they never terminate and never fall into a repeating pattern — they run on forever without settling down:

√2 = 1.41421356237...
π  = 3.14159265358...

That non-terminating, non-repeating decimal is the fingerprint of an irrational. A rational always terminates or repeats; an irrational does neither.

Reals (ℝ): the whole number line

Put the rationals and the irrationals together and you get the real numbers, written ℝ. This is the number line you've always pictured — every point on it, no gaps. The rationals are dense (there's one between any two you pick), but they leave pinprick holes exactly where numbers like √2 and π live. The irrationals fill those holes. Together they form one unbroken line.

Almost any quantity you measure in the physical world — a length, a temperature, a weight — lives in ℝ.

Is this the end? For everyday math, yes. But the same pattern has one more famous chapter: ask "what is √−1?" and no real number works, because any real squared is zero or positive. To answer it, mathematicians invented the complex numbers (ℂ), which extend ℝ the way ℤ once extended ℕ. We won't use them here — notice the move never really stops.

They nest like Russian dolls

Step back and look at what we built. Each family contained the last one whole:

ℕ  ⊂  ℤ  ⊂  ℚ  ⊂  ℝ
naturals  integers  rationals  reals

That ⊂ means "is contained in." Read it as a chain of true statements:

• Every natural number is also an integer.
• Every integer is also a rational (5 = 5/1).
• Every rational is also a real.

The reverse is not true, and that's the interesting part: -2 is an integer but not a natural; 1/2 is rational but not an integer; √2 is real but not rational. Each family is strictly bigger than the one inside it.

This nesting is exactly the language of sets and subsets — each family is a set, and the smaller ones are subsets of the larger. If that framing is new to you, see /guides/sets-relations-and-functions. And if the whole topic still feels intimidating rather than interesting, that's worth addressing directly: /guides/why-math-isnt-your-enemy.

For builders

If you write code, you've already been using two of these families, probably without naming them.

• An int type is your machine's version of the integers ℤ — whole numbers, positive and negative. (With a catch: a real int has a maximum and minimum, while ℤ goes on forever. Overflow is what happens when a value walks off that edge.)
• A float (or double) is your machine's attempt at the reals ℝ — numbers with a fractional part.

Two things follow directly from the families above.

Integer division throws away the fraction. In many languages, 7 / 2 on two integers gives 3, not 3.5. That's not a bug — the result is being kept inside ℤ, where 3.5 doesn't exist, so it truncates. To get 3.5 you have to ask for real (float) division.

Floats are finite approximations, not the true reals. This one bites people, so it gets its own callout.

⚠️ Gotcha: a float is not an exact real number. ℝ is infinite in precision — some reals (like √2) need infinitely many digits. Your computer has finite memory, so it cannot store every real exactly; it stores the nearest value it can represent. That's why, in almost every language:

0.1 + 0.2  =  0.30000000000000004

The numbers 0.1 and 0.2 can't be held exactly in the machine's binary format, so their stored versions are a hair off, and the error shows up in the sum. Nothing is broken — it's the unavoidable cost of squeezing the infinite real line into a fixed number of bits. The practical rule: don't test floats for exact equality; check whether they're close enough.

Recap

• Each family of numbers was invented to make an operation always have an answer.
• ℕ (naturals): counting numbers. Subtraction can fail (3 − 5).
• ℤ (integers): add negatives — subtraction always works. Division can fail (1 ÷ 2).
• ℚ (rationals): fractions a/b — division works (except by zero). Decimals terminate or repeat. But some real lengths (√2, π) are provably not fractions.
• Irrationals: can't be written as a ratio; decimals never terminate or repeat.
• ℝ (reals): rationals + irrationals = the full, gapless number line. (Next extension, ℂ, handles √−1.)
• They nest: ℕ ⊂ ℤ ⊂ ℚ ⊂ ℝ — every member of an inner family belongs to the outer ones too.
• For code: int ≈ integers, float ≈ a finite approximation of the reals, which is why 0.1 + 0.2 isn't exactly 0.3.

Quick check before you move on:

[
  {
    "q": "Subtraction like 3 − 5 has no answer among the natural numbers. Which family was invented to make subtraction always work?",
    "choices": ["The rationals (ℚ)", "The integers (ℤ)", "The irrationals", "The reals (ℝ)"],
    "answer": 1,
    "explain": "Adding the negative numbers gives the integers, where any subtraction of integers produces an integer. Rationals fix division, not subtraction."
  },
  {
    "q": "What makes a number irrational?",
    "choices": ["It is negative", "It is larger than every fraction", "It cannot be written as a ratio of two integers, and its decimal never terminates or repeats", "It has a decimal point in it"],
    "answer": 2,
    "explain": "An irrational like √2 or π provably can't be expressed as a/b for integers a and b; written out, its decimal goes on forever with no repeating pattern."
  },
  {
    "q": "Given the nesting ℕ ⊂ ℤ ⊂ ℚ ⊂ ℝ, which statement is true?",
    "choices": ["Every integer is also a rational number", "Every rational is also an integer", "Every real number is also a natural number", "Every rational is also an irrational"],
    "answer": 0,
    "explain": "Each family contains the one inside it: an integer like 5 is the rational 5/1, so every integer is rational. The reverse fails — 1/2 is rational but not an integer."
  }
]

← Guide overview · Phase 2: Bases: Binary, Decimal, Hex →