Roman numeral converter
Fill in either box. Numbers become numerals in standard subtractive form, so 4 is IV and not IIII, and numerals are read back to numbers with anything malformed rejected rather than guessed at. If both boxes have something in them, the numeral wins.
How to convert Roman numerals
Two rules do nearly all the work. A smaller symbol placed before a larger one subtracts, so IV is four and IX is nine, but only I, X and C are allowed to subtract and only from the next two symbols up. I can precede V and X, X can precede L and C, C can precede D and M, and nothing else is legal. That single restriction is why 99 is XCIX and never IC, and why 49 is XLIX rather than IL. The second rule is that no symbol repeats more than three times, which is where IV comes from instead of IIII. V, L and D never repeat at all, since two of any of them is just the next symbol up.
Validation here is done by round trip rather than by a table of special cases. A numeral is read to a number, that number is written back out, and the two are compared. If they do not match, the numeral was not in standard form. IIII reads as four, four writes back as IV, so IIII is refused. The same test catches VV, XXXX, MMMM, IL and every other malformed string without anyone having to enumerate them. Case and surrounding spaces are ignored, so mmxxvi is read the same as MMXXVI.
The IIII on clock faces is a real exception and not a mistake. Clock dials have used it for centuries, alongside a normal IX on the other side, and the usual explanations are visual balance against the VIII opposite and the convenience of casting the numerals in three moulds. It is a typographic convention rather than notation, which is why this converter will not produce it and will not accept it.
The ceiling is 3999 because 4000 needs four M and standard form forbids it. Romans wrote larger numbers with a vinculum, a bar over the numeral meaning multiply by a thousand, so a barred V is five thousand. That notation has no agreed digital representation, it is not what anyone converting a date wants, and it is left out. There is also no zero and no negative number in the system, and no fraction: Roman fractions were a separate duodecimal set with S for a half and dots for twelfths, and they are not part of what this page does.
What people use it for
- Reading the copyright year at the end of a film
- Working out the date carved on a building or a monument
- Writing a year for a tattoo, a ring or an invitation
- Checking an event number such as a Super Bowl or an Olympiad
- Reading a preface or chapter numbered in numerals
- Confirming that a numeral someone has written down is actually valid
Questions
Standard form does not repeat a symbol more than three times. The IIII on clock faces is a typographic convention rather than correct notation.
3999, which is MMMCMXCIX. Anything above it would need four M in a row, and standard form does not allow that.
With a vinculum, a bar over the numeral meaning multiply by a thousand. There is no settled way to type it, so it is not offered here.
No. The system has no zero, no negative numbers and no place value. It was built for recording quantities, not for arithmetic.
Yes, historically, and they are a separate duodecimal system: S for a half and dots for twelfths. This converter handles whole numbers only.
Only I, X and C may subtract, and only from the next two symbols up. I can precede V and X, never C. The same rule makes 49 XLIX rather than IL.
I, X, C and M, up to three times. V, L and D never repeat, because two of any of them is simply the next symbol up.
It is not standard form. The check reads it to a number and writes that number back out; if the result differs from what you typed, it was malformed.
Yes, and surrounding spaces. mmxxvi and MMXXVI both give 2026.
The numeral box wins. Clear it if you want to convert a number.
2026. M is 1000, so MM is 2000, XX is 20 and VI is 6.
1994. MCM is 1900, XC is 90, IV is 4. It is the standard example because it uses three subtractive pairs at once.
MDCLXVI, which is 1666. All seven symbols appear exactly once, in descending order.
They did not exist as separate letters in Latin. I served for both I and J, V for both U and V, which is why the numeral set has seven letters and not nine.
So the front matter can be typeset and paginated after the body, without renumbering every page when the introduction grows or shrinks.
No. Alphabetical order puts IX (9) ahead of V (5), and XL (40) ahead of XX (20). Convert to numbers, sort those, then convert back.
No. Unicode carries Ⅰ Ⅴ Ⅹ and the rest as single characters for round-tripping old documents, and they are best avoided: they defeat a search for the ordinary letters and confuse screen readers. Type the capital letters, which is what this page gives you.