Roman Numerals: The Rules, the Exceptions and Why IIII Exists
Roman numerals look simple until you try to validate one. Reading a numeral is a matter of adding symbols and subtracting the occasional smaller one, but deciding whether a string is a well formed numeral involves repetition limits, a fixed list of permitted subtractive pairs and an ordering rule. This converter works in both directions and validates strictly, telling you the correct form rather than simply refusing. Everything runs in your browser.
Table of Contents
The Seven Symbols
I, V, X, L, C, D and M. One, five, ten, fifty, one hundred, five hundred and one thousand. Seven symbols, and everything else is a combination of them.
They alternate between ones and fives. One, five, ten, fifty, and so on, which reflects a counting system built around fingers on one hand and then both.
The system is additive at heart. Symbols are written largest first and added together, with subtraction as a later refinement rather than a founding principle.
There is no positional notation. A symbol means the same thing wherever it appears, which is why arithmetic in Roman numerals is so awkward.
How to Use This Converter
Choose a direction and type. Both convert as you type, and the swap button moves the current result into the opposite field.
The arithmetic panel shows the working. Each symbol or subtractive pair appears as its own block with its value, so a numeral like MCMXCIV becomes four blocks rather than seven letters.
Invalid input still gets an answer. Enter IIII and the tool reads it as four and tells you four is written IV, with a button to correct it.
The examples cover the awkward cases. Four, nine, forty and 1994 are where most confusion actually lives.
The Rules for Writing Them
Largest to smallest, left to right. The default ordering, and any departure from it must be one of the permitted subtractive pairs.
I, X, C and M repeat at most three times. Four of the same symbol in a row is what the subtractive forms exist to avoid.
V, L and D never repeat. Two fives is ten, which already has its own symbol, so VV is never correct.
These rules together give exactly one form per number. Which is why the round trip works as a validator: there is only one right answer for any value.
Subtractive Pairs and Their Limits
Only six pairs are permitted. IV, IX, XL, XC, CD and CM. Everything else that looks like a subtraction is wrong.
Only one step down is allowed. I can precede V or X but never L or C, which is why IC is not ninety nine. That number is XCIX.
Only I, X and C subtract. V, L and D never appear before a larger symbol, so VL is not forty five.
MCMXCIV has three in a row. M, then CM for nine hundred, then XC for ninety, then IV for four. It is the standard example precisely because it stacks all three.
Why Clocks Show IIII
Clock faces routinely use IIII rather than IV. It is a genuine and very old convention, and it is not correct Roman numeral notation.
The usual explanation is visual balance. IIII opposite VIII balances the dial better than IV does, and the four heaviest numerals then sit symmetrically.
Other explanations exist. Casting efficiency for the moulds, and a reluctance to use the first two letters of a god's name, are both cited, and none is settled.
This tool rejects it. Deliberately, since the point of validation is to identify what is standard. It still tells you the value and the correct form.
Why the Range Stops at 3999
M is the largest symbol. With at most three repetitions, the largest expressible value is 3999, written MMMCMXCIX.
The vinculum extended it. An overline above a numeral multiplies it by a thousand, so a barred V is five thousand.
There is no agreed plain text form. Overlines are typography rather than characters, and the conventions people use to fake them in text do not agree with each other.
Which is why converters stop here. Producing something above 3999 would mean inventing a notation and presenting it as standard.
The Missing Zero
There is no Roman numeral for zero. The system counts things, and counting nothing did not require a symbol.
Medieval scholars wrote N for nulla. The Latin for nothing, used in astronomical tables, but it was never part of the numeral system proper.
The absence is why arithmetic is so hard. Without zero there is no place value, and without place value there is no column arithmetic, which is a large part of why the system was eventually replaced.
Negative numbers are equally absent. For the same reason, and no convention for them ever developed.
Where They Are Still Used
Clock faces, monarchs and popes. The traditional survivals, where the numeral carries a sense of formality that Arabic digits do not.
Film copyright dates and book front matter. Copyright lines in Roman numerals were once near universal, and page numbering for prefaces still uses lower case numerals.
Sporting events and sequels. Numbered championships and film sequels use them for the same reason: a Roman numeral reads as more consequential.
Outline numbering. Legal documents and formal outlines still nest Roman numerals inside Arabic ones as a level marker.
Common Mistakes to Avoid
Writing IIII for four. Correct on a clock face by convention and incorrect everywhere else.
Writing IC for ninety nine. Only one step down is permitted, so it is XCIX.
Repeating V, L or D. Two of any of them is a value that already has its own symbol.
Assuming any converter validates. Many read malformed numerals silently and return a number, which hides the error rather than reporting it.
Frequently Asked Questions
On strictness: this converter rejects numerals that can be read but are not well formed, and tells you the standard form instead. If you are transcribing a clock face or a historical inscription, expect it to disagree with what is written there.