A numeral system is a consistent scheme for writing every number with a finite set of symbols β and humanity has invented well over a hundred of them. Each one answers the same five design questions: what base, which symbols, does position carry meaning, is there a zero, and can ordinary people calculate in it. The system on your keyboard β positional base-10 with a true zero β won because it writes any number in few symbols and makes arithmetic mechanical.
Marks required, on a logarithmic scale β from nearly two thousand tally strokes down to four digits. This single chart is the history of numeration: each system buys shorter numbers, until the positional systems make length grow only with the logarithm of the number.
The zero story, briefly: sign-value systems like Roman numerals never needed a zero β nothing written means nothing counted (medieval scribes who did need it borrowed nulla, or just N). Positional systems can't live without one: in 205, the zero holds the tens place open, and without it 205 and 25 collapse into the same marks. Babylonian scribes managed for centuries with a risky blank space before inventing a placeholder wedge; the Maya had a shell-glyph zero; but treating zero as a number you can calculate with was codified in India β Brahmagupta wrote down its arithmetic rules in 628 CE. Our Roman numeral converter tells the nulla story, and the binary converter shows zero holding places in base 2 all day long.
What a numeral system is β base, symbols, place value, zero, and arithmetic rules β why positional notation with zero defeated sign-value systems, and the memorize-vs-length bargain every civilization struck.
From notched bones older than agriculture to the digits in your calculator spans roughly forty thousand years β with the decisive inventions crowded into the last five. The pattern repeats everywhere: tallies, then grouped symbols, then sign-value systems, then positional systems, then positional with a true zero β each step trading a little memorization for shorter, more calculable numbers. Ancient dates below are conventional scholarly estimates from surviving artifacts.
Years before the present on a logarithmic scale. Notice the compression: counting itself is Ice Age technology, but everything that makes modern arithmetic possible happened in the last few thousand years β and the digits you use daily won Europe less than eight centuries ago.
The strangest chapter is the last one: Hindu-Arabic numerals reached Europe through al-Khwarizmi's 9th-century treatise (his name became the word algorithm) and Fibonacci's 1202 Liber Abaci β and then fought Roman numerals for three hundred years. Florence banned the new digits in bookkeeping in 1299, partly on the argument that a 0 was too easy to doctor into a 6 or a 9. Merchants won the argument the way merchants do: the new numbers were simply better for business, and by the 1500s the abacus-versus-algorithm contest was over in Europe. (The abacus, of course, never conceded elsewhere β see our Interactive Abacus for the 1946 rematch it won.)
Timeline of numeral systems β Lebombo and Ishango tally bones, Sumerian cuneiform, Egyptian hieroglyphs, Babylonian base-60, Chinese oracle-bone and rod numerals, Greek, Roman, Maya, and the Hindu-Arabic digits with zero.
Numeral systems fall into four families β unary tallies, sign-value systems like Egyptian and Roman, alphabetic systems like Greek, and the positional family that includes Babylonian, Maya, Chinese rods and our own digits. Only the last family survived contact with algebra. The tradeoff is precise: Greek alphabetic numerals demand 27 memorized symbols but write 1,987 in four letters; a tally needs one symbol and 1,987 strokes.
| System | Era | Base | Positional | Zero | 1987 looks like |
|---|---|---|---|---|---|
| Tally marks | > 20,000 BCE | 1 (groups of 5) | no | no | |||| |||| β¦ (1,987 strokes) |
| Egyptian | ~3000 BCE | 10 | no β sign-value | no | πΌ π’Γ9 πΓ8 π€Γ7 (25 glyphs) |
| Babylonian | ~1800 BCE | 60 | yes | placeholder, late | 33 ; 7 (two positions, 13 wedges) |
| Greek alphabetic | ~600 BCE | 10 | no β alphabetic | no | Ν΅Ξ±Ο ΟΞΆ (4 letters) |
| Roman | ~500 BCE | 10 (with 5s) | no β sign-value | no (nulla, later) | MCMLXXXVII (10 letters) |
| Chinese rods | by ~400 BCE | 10 | yes | empty square, later γ | δΈδΉε «δΈ in rod columns (4 positions) |
| Maya | ~300 BCE | 20 (18 in calendars) | yes | yes β shell glyph | 4 Β· 19 Β· 7 (three positions, 14 marks) |
| Hindu-Arabic | zero by 628 CE | 10 | yes | yes β full number | 1987 (4 digits) |
The alphabet each system demands. Babylonian wrote everything with two wedge shapes; Greek merchants memorized twenty-seven letters to earn four-symbol numbers. Pair this with the marks chart on The Big Idea tab and you can price any system's bargain exactly.
Why arithmetic settled the contest: sign-value systems read beautifully but calculate terribly β Roman multiplication was a specialist's craft performed on a counting board, with numerals only recording the answer (our Roman converter demonstrates the authentic merge-and-trade addition). Positional systems make the algorithms you learned in school possible: line up the places, work column by column, carry. The moment commerce needed division and algebra needed manipulation, place value stopped being one option among many and became the only game in town.
Comparison of numeral systems β base, positionality, zero, era and sample numbers for tally, Egyptian, Babylonian, Greek, Roman, Chinese rod, Maya and Hindu-Arabic numerals, with the symbols-versus-length tradeoff charted.
Numeral systems up close β Roman numerals and their modern survivals, Babylonian sexagesimal, Maya vigesimal with the shell zero, Greek alphabetic numerals and isopsephy, and the Thai, Lao and Khmer digits still in daily use.
A whole number from 1 to 999,999 β watch it change costumes on the right, then follow any system deeper with its full converter.
| Converter | What it does |
|---|---|
| Binary | Decimal β base 2 with exact fractions, two's complement, IEEE 754 float anatomy, text-to-binary |
| Octal | Base 8 with chmod permission decoding and the binary bridge |
| Duodecimal | Base 12 with authentic Pitman digits β the case for counting in dozens |
| Hexadecimal | Base 16 with bitwise arithmetic and a color lab |
| Negabinary | Base β2, where negative numbers need no minus sign |
| Roman numerals | Conversion, authentic merge arithmetic, a live Roman clock, and the missing-zero story |
| Egyptian numerals | Hieroglyphic powers of ten |
| Greek numerals | The 27-letter alphabetic system |
| Mayan numerals | Dots, bars, and the shell zero in base 20 |
| Phoenician numerals | The merchant script behind the alphabet itself |
| Thai numerals | The living digits of Thailand |
| Babylonian numerals | Base-60 in two wedge shapes β the first positional system |
| Chinese rod numerals | The counting-rod system that made China decimal and positional early |
| Tally marks | Where it all began β one mark, one thing |
| Hebrew numerals | The alphabetic system behind gematria |
| Armenian numerals | Another alphabet moonlighting as digits |
| Sexagesimal | Base 60 for degrees, minutes and seconds β Babylon's living legacy |
| Balanced ternary | Digits β1, 0, +1 β the "most beautiful" base, per Knuth |
| Quinary & senary | Bases 5 and 6 β one hand, and one hand plus a fist |
| Interactive abacus | Soroban and suanpan with working beads β arithmetic as furniture |
| Ancient units | Cubits, stadia, talents and amphorae in authentic ratios |
| Pascal's triangle | Binomial coefficients and the patterns hiding in them |
A closing thought worth stealing: numerals are the oldest software. The hardware β ten fingers, clay, papyrus, silicon β keeps changing, but each system is a set of encoding rules running on whatever substrate history provides, and upgrades propagate the way software upgrades always have: slowly, against resistance, and then all at once. You are reading this in the third millennium of a format war that Hindu-Arabic positional notation won so completely that we call its digits, simply, numbers.
Try any number in binary, octal, duodecimal, hexadecimal and Roman numerals, then explore the full converter collection β from Egyptian hieroglyphs and Mayan dots to the interactive abacus β with classroom activities for teaching numeral systems.