according to wikipedia, the largest number Roman Numeral system can represent is represented like following: (answer below has much bigger number represented)


why can't it be represented like following?

  • 2
    This looks like a good question for Latin Language – jk - Reinstate Monica Aug 9 '19 at 11:29
  • 1
    @curiousdannii IMHO technically, Roman Numerals is a language of its own. – Sajuuk Aug 9 '19 at 12:59
  • 3
    @Sajuuk Maybe in a Computer Science definition of "language", but definitely not a language in the scope of linguistics. – jk - Reinstate Monica Aug 9 '19 at 13:14
  • 1
    @Sajuuk Nope, it's not a language in the sense of languages studied by linguists. It's only a code. – curiousdannii Aug 9 '19 at 13:33
  • Echoing jknappen, I think they would welcome this sort of question at Latin Language.SE – Mark Beadles Aug 9 '19 at 14:50

As a part-time computer scientist, I've tried to come up with different ways of analyzing the Roman numeral system so that the rules can be formulated in as general a way as possible, with no special cases per glyph. The one that satisifes me the most, for the time being, is one in which we analyze the seven glyphs as four sets that each have a multiplier — similar to how each column has a place-value multiplier in the Arabic numeral system.

These sets are I V X ; X L C ; C D M ; M (the Romans did not leave us perfectly symmetrical sets). You can only subtract the first glyph in each set from the second or third in the same set. This precludes IM.

To see this analysis in action, read this simple handout. It includes exercises to practice with and an online tool to verify your answers. For the sake of perpetual Stack Exchange availability, I'll copy a screenshot of the first page here:

Roman numeral handout first page

  • wonderful handout, I didn't find rule for how many same symbol can one place consecutively though. is it 3? – Sajuuk Aug 10 '19 at 8:28
  • Additionally, could you provide source/reference for this handout? is this grammar authoritative in any regard? – Sajuuk Aug 10 '19 at 8:49
  • @Sajuuk The source is the Wikipedia article on Roman numerals. It lists a few variants but I just described the most widely accepted one. Interestingly, I reread the article for the first time in about 8 years and found that it uses a similar analysis about place value. I just added a note about the source to the handout. – Luke Sawczak Aug 10 '19 at 14:41
  • As for how many of the same symbol: you can only do that for the x1 symbols. The limit is usually 3, but a common variant of the system allows 4. (I mentioned that in one of the exercises because it greatly increases the upper bound of numbers you can write! The checker tool linked in the handout also lets you write 4 of the same symbol.) – Luke Sawczak Aug 10 '19 at 14:42
  • Then there's this inscription, over a fireplace at Trerice. Yes, that is a numeral '3' on the end. – Colin Fine Aug 10 '19 at 23:05

It's because of the grammar of the Roman numerals (It's not the same as natural language grammar, but it's still grammar).

I does not combine with all characters. I combines on the left of V or X to mean -1. (Just like X combines on the left of L or C to mean -10, but combinations like XM does not work).

So the problem with MMMIM is that IM is invalid and cannot mean 999. To subtract anything from M, you can use C.


Maybe because you cannot skip so many slots, mind that 49 is XLIX, not IL, and 99 is XCIX, not IC. Anyhow, you can write much bigger numbers than that MMMCMXCIX (3999). By putting a horizontal line, called vinculum or overline above a Roman numeral, you multiply it by 1,000, so M with an overline is 1,000,000 that is one million. Using the overline, the biggest number that can be written is 3,999,999 (see the image).

Note, that the overline is a medieval feature. enter image description here

  • wikipedia used the word "another system is 'Vinculum'", is this some sort of extension to original Roman Numeral system? and by medieval feature, do you mean this is only used in europe since middle age? – Sajuuk Aug 9 '19 at 12:18
  • @Sajuuk - Yes, you are right, since the Middle Ages. – Yellow Sky Aug 9 '19 at 12:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.