Notches on a Bone, All the Way to Infinity The World of Numbers

Long before any number was ever written as a symbol, the Lebombo Bone (from the Lebombo Mountains, roughly 35,000 years old) carries 29 carved notches, likely a lunar tally, and the Ishango Bone (from the Democratic Republic of Congo, roughly 20,000 years old) carries three separate columns of notches, one of which groups exactly the numbers 11, 13, 17, and 19 together -- every prime number between 10 and 20. Whether or not its carver understood 'prime' as a concept, the grouping itself is a genuine, deliberate act of counting and classifying, the earliest ancestor of the Natural Numbers (N = {1, 2, 3, 4, ...}) this chapter starts from. Centuries later and much closer to home, the Indus Valley Civilisation standardised weights for trade at sites like Lothal, and Vedic and Buddhist texts named numbers at a genuinely staggering scale: the Rigveda already uses powers of 10, and later texts like the Lalitavistara name a number as large as 10 to the power 53 (tallakshana). Naming a number that large, correctly and consistently, already strains a system built purely on tally marks and words -- a strain this chapter's whole story exists to resolve, one number system at a time.

Zero began as a philosophical idea long before it became a working number: the concept of Shunyata (emptiness) appears in the Upanishads and Buddhist thought centuries before anyone used a symbol for it in arithmetic. The Bakhshali Manuscript uses a dot (bindu) to mark an empty place value, and it is Brahmagupta, in his Brahmasphutasiddhanta (628 CE), who is credited with the first attempt to give zero a genuine set of arithmetic rules rather than treating it as mere absence: a + 0 = a, a - 0 = a, and a x 0 = 0. Turning 'nothing' into a number with its own rules, rather than a gap to be skipped over, is precisely what makes zero usable in calculation at all, and it is the hinge this chapter's next number system swings on.

Brahmagupta introduced negative numbers through the language of commerce: a Dhana (fortune) is positive, a Rina (debt) is negative, and this framing makes every rule of integer arithmetic read like ordinary financial common sense rather than an abstract symbol-pushing rule. A fortune plus a fortune is a fortune (5+4=9); a debt plus a debt is a debt (owing 5 and borrowing 4 more leaves a debt of 9); a debt times a fortune is a debt (taking on 4 debts of 3 each leaves a total debt of 12, so (-3) x 4 = -12). The one rule that resists common sense at first glance is a debt times a debt: (-3) x (-4) = +12. Brahmagupta's own commerce framing resolves even this: if someone takes away four of your debts, each worth 3, you are immediately 12 richer, so removing (subtracting, hence the double negative) four debts of 3 genuinely does make you 12 units better off. Integers, written Z (from the German Zahlen), are exactly the set of whole numbers extended to include every fortune and every debt, alongside zero itself.

A field divided among three children, or a recipe needing half a cup of ghee, both need a number system finer than whole fortunes and debts: the Rational Numbers, Q, defined as any number expressible as p/q for integers p and q with q never zero. A single rational number has infinitely many equivalent disguises (1013/2026 is exactly 1/2, since 1013 x 2 = 2026), and by convention it is written in lowest terms, with p and q sharing no common factor. Rationals obey closure: adding, subtracting, or multiplying any two rationals always gives another rational, and dividing does too, except by zero. Placed on a number line, a rational's exact position comes from dividing the unit interval into q equal parts and counting p of them; its absolute value, written |x|, is simply its distance from zero, always zero or positive, and the distance between any two rationals a and b is |a - b|. A genuinely surprising property follows from all this: between any two rational numbers, however close together, another rational number always exists -- simply average them, (a+b)/2, and the result always lies strictly between the two. Repeating this averaging trick endlessly proves something remarkable: there is no 'next' rational number after any given one; rationals are infinitely dense, packed with no gaps of their own making, anywhere a person might look.

Around 800 BCE, Baudhayana's Shulbasutra, a manual for constructing fire altars with geometric precision, ran into a length that no fraction could capture exactly: the diagonal of a unit square. Using what today is called the Baudhayana-Pythagoras theorem, a unit square's diagonal d satisfies 1 squared + 1 squared = d squared, so d squared = 2, and d is the square root of 2. Around 400 BCE, a Pythagorean named Hippasus is credited with the first proof that this length genuinely cannot be written as any fraction at all, using proof by contradiction: assume the square root of 2 equals p/q in lowest terms; squaring gives 2 = p squared / q squared, so p squared = 2 q squared, meaning p squared is even, which forces p itself to be even (write p = 2k); substituting gives 2 q squared = 4 k squared, so q squared = 2 k squared, meaning q must be even too. But if both p and q are even, they share a factor of 2, directly contradicting the assumption that p/q was already in lowest terms. The contradiction proves the assumption was false: the square root of 2 is irrational. The same length can be constructed exactly with only a compass: build a right angle with two sides of length 1, and the hypotenuse is exactly the square root of 2, a length that is perfectly real and perfectly constructible, just never expressible as a clean ratio of whole numbers. Centuries later, Aryabhata (499 CE) gave pi as the approximation 3927/1250 = 3.1416, explicitly calling it asanna, meaning approximate, showing he understood it was not exact; a rigorous proof that pi itself is irrational did not arrive until Johann Lambert in 1761. In the 14th century, Madhava of Sangamagrama, from the Kerala school of mathematics, found an exact infinite sum equal to pi/4: 1 - 1/3 + 1/5 - 1/7 + 1/9 - ..., continuing forever, centuries before equivalent results appeared in Europe.

Every rational number's decimal expansion falls into exactly one of two kinds: terminating (3/8 = 0.375, stopping cleanly) or repeating (5/11 = 0.454545..., looping forever). Which one happens is decided entirely by the denominator's prime factorisation, once written in lowest terms: if it contains only the primes 2 and/or 5, the decimal terminates (3/20 = 3/(2 squared x 5) = 15/100 = 0.15); any other prime factor forces a repeating decimal, and the reason is a genuine pigeonhole argument -- long division only ever produces a remainder from a finite list (for a denominator of 7, only remainders 1 through 6 are possible), so a remainder must eventually repeat, and once it does, the digits from that point on repeat too. Converting a repeating decimal back into p/q form uses a clean algebraic trick: let x equal the decimal, multiply by whatever power of 10 shifts one full repeating block past the decimal point, then subtract the original x to cancel the infinite tail entirely, leaving a simple equation to solve. Applied to x = 0.6 repeating: 10x = 6.666..., so 10x - x = 6, giving 9x = 6, x = 2/3. One repeating decimal carries an extra surprise: 1/7 = 0.142857142857..., and multiplying 142857 by 1, 2, 3, 4, 5, or 6 just rotates its six digits into a new starting position (142857 x 3 = 428571, for instance) -- a genuine 'cyclic number'. This property is not automatic for every prime denominator: 1/13's repeating block (076923) does NOT extend the same way across all of 1/13 through 12/13; instead, two separate six-digit cyclic families appear, not one, since 13 does not share 7's special property of having 10 as what number theory calls a 'primitive root'. At the far end of decimal behaviour sits a genuinely different kind of number: 1.010010001000010..., a decimal that never terminates and never settles into any repeating block at all, however far it is carried out. Such a number cannot be rational, since every rational's decimal must either terminate or repeat -- it is irrational by construction, built to defeat both possibilities on purpose.

Assembling everything: Natural numbers sit inside Integers, which sit inside Rationals, and Rationals together with Irrationals fill up the Real Numbers entirely, with no further gaps left along the line. Every step of this journey followed the same pattern: whole numbers couldn't hold a debt, so integers were invented; integers couldn't hold half a cup of ghee, so rationals were invented; and rationals couldn't even measure a unit square's own diagonal, so irrationals were invented alongside them, completing the reals. One crack remains, deliberately left open: what is the square root of negative 1? No real number squares to a negative result, since a positive number squared is positive and a negative number squared is also positive -- so root of -1 simply does not exist anywhere on this real number line, however far it is searched. This is not a mistake or a dead end; it is a genuine, honest gap, and it is exactly the gap that imaginary numbers are built to fill, a story deliberately saved for a later year rather than rushed here.