Roman Numeral Converter

Seven letters, no zero, and no place value. A Roman numeral is a sum you read left to right, and the only complication is that six pairs run backwards.

Building one

2026 = 1000 + 1000 + 10 + 10 + 5 + 1
        M      M     X    X    V   I

     = MMXXVI

Work down from the biggest letter that fits and keep
going. That is the whole method.

Take the largest letter that fits, write it down, subtract, repeat. Since M is the biggest letter and it is worth a thousand, four M in a row is the practical ceiling, which puts the top of the ordinary system at 3,999.

The six pairs that go backwards

IV = 4     one before five
IX = 9     one before ten
XL = 40    ten before fifty
XC = 90    ten before a hundred
CD = 400   a hundred before five hundred
CM = 900   a hundred before a thousand

Six pairs, and no others. IL is not 49 and IC is
not 99, whatever they look like.

A smaller letter placed before a larger one gets taken away instead of added. Only six combinations are allowed, and the rule behind them is that you can only subtract a power of ten, and only from the next two values up. So I can precede V and X, but never L or C.

This is where the tattoo goes wrong. People write IIX for eight or IC for ninety-nine, and both are read here instead of refused, with the usual spelling offered alongside.

Bigger than 3,999

Put a bar over a letter and it is worth a thousand times more. So a barred V is five thousand and a barred M is a million, which takes the system as far as 3,999,999.

The bar is drawn here with a combining overline, U+0305, which means it survives copy and paste into anything that handles Unicode properly and falls apart in anything that does not. If it matters, use an image.

Before the bar caught on, Romans used the apostrophus instead: a series of brackets around a symbol. Those survive in Unicode as ↁ, ↂ, ↇ and ↈ, and they are read here too.

What the Romans actually wrote

I II III IIII V VI VII VIII IX X XI XII

Almost every clock face uses IIII, not IV. It balances
the VIII opposite it, and a clockmaker casting numerals
could cut them from four moulds instead of five.

The tidy system taught in schools was settled in the Renaissance, not in Rome. Inscriptions are full of IIII, XXXX and VIIII. A scribe writing IIXX meant eighteen, two taken from twenty, which no modern reader would get.

Reading here is deliberately generous. Anything that can be read is read, and the usual modern spelling is shown next to it and not in place of it. One place that generosity stops: IIXX comes out as twenty, because the pairwise rule everyone now uses reads it as one, then nine, then ten.

No zero, and no fractions either

There is no symbol for nothing because the system never needed one. Place value is what makes a zero necessary, and Roman numerals have no places to hold. The word nulla got used when a zero was genuinely wanted, mostly by medieval astronomers.

Fractions existed and were twelfths, marked with dots and an S for a half. They are not in the modern system and are not here.

Where you still meet them

Copyright dates on films, monarchs and popes, Super Bowls, book prologues, clock faces, and the year carved on a building. Also chemistry, where the Roman numeral in iron(III) is an oxidation state, not a count.

One practical warning for anyone converting a date for something permanent: check it twice and get somebody else to check it. A misplaced I is a different year and, on a tattoo or a headstone, a different problem.