Yesterday I thought of a great new project that could be a nice little article, or, if I had a grad student with a background in classical archaeology, as a nice little thesis, or, if someone else wants to work with me, a co-authored paper. Heck, if you scam my idea, more power to you – I will cite you widely if it’s good, and mock you widely if not! You see, the Epigraphische Datenbank Clauss-Slaby is a searchable full-text database with over 350,000 Latin inscriptions (including over 20,000 images). You can enter a word (e.g, Germaniae) and it returns all the inscriptions that have that word. Nifty, huh? Just in mucking about with EDCS today I discovered two or three things that will be coming out in my book that are in need of revision, which makes me only a little bitter.

Now of course I’m not a classicist (I have three terms of Latin under my belt, but that’s hardly enough to make me an expert), but I do know a thing or three about Roman numerals. The study of Roman numerals is sorely neglected in modern epigraphy, which is a shame because there are some really interesting social questions to be asked relating to regional identity and literacy (the sort of stuff, e.g., that Greg Woolf does). We think that we know Roman numerals: just take I, V, X, L, C, D, and M, string them together in groups of no more than three, use subtractive notation for numbers like 9 and 44, and you’re done. But it isn’t so simple.

The Roman numerals are not a static and unified system; there are various expressions for the same number (e.g. XVIII vs. XIIX for 18, or XXXXX vs. L for 50). Back in the 1950s, Arthur and Joyce Gordon did some interesting statistical analysis, indicating some potential sources of this variability (chronological, regional, and textual), but he didn’t have the sort of massive resources that the EDCS provides. So, for instance, it is often said that IIIII for 5, XXXXX for 50, and CCCCC for 500 (i.e., not using the sub-base signs V, L, and D) are particularly found in African inscriptions. Well, a quick search for ‘CCCCC’ and ‘XXXXX’ suggest to me that this isn’t a full explanation. Are certain types of inscription more likely to contain these variants? Could we be dealing with a chronological difference? Could we be dealing with a variant typical of minimally literate writers, or writers of informal texts? Or could it be that the shorter forms are used when there’s less room on the medium, with longer variants used when space is not at a premium? I have no idea, but the only way to find out would be to build a list of inscriptions that use these variants, map them in time and space, and evaluate them in terms of the texts in which they occur.

Now, there are some methodological complexities: some of the interesting variation is between different forms for the same character, and there is no way to search for that. Some of the Roman numeral forms (the use of a horizontal bar or *vinculum* over a numeral to indicate multiplication by 1000) aren’t represented consistently, or at all, so one would just need to rely on other published material to find the relevant inscriptions. And quite a lot of the project would require taking the database results and then referring to the *Corpus Inscriptionum Latinarum*. But ultimately it would be taking what seems to be a rather dry subject (variability in Roman numerals) and potentially correlating it with variability in social identities (class, ethnic, professional). Well, I think it’s cool, anyway.

It seems cool to me but I’m not volunteering. :) Good for you for throwing the idea out there.

Although i am no expert, you may want to see if there is a correlation with the written numeral variations and how the native languages of the time, in any given region, understood mathematics (good luck with that). With what we know of as a common use of roman numerals, you have to begin your mental calculation moving from left to right, indicating the highest integer to the lowest integer. However, there are times when the integer values are not in order of highest to lowest, and you have to perform a different kind of mathematics where subtracting an integer-value group from the next highest integer-value is necessary. To some people, this is about the same as asking them to subtract negative numbers; the logic baffles them no matter how much they try. It is sad, but in knowing this, we can assume that some languages may not have been developed in cultures where mathematics was a strong-suit, and maybe, just maybe, that is the reason for ‘IV’ being represented as ‘IIII’. It may also simply be that the culture’s mindset was more ‘straight-forward’ in thinking and found it ridiculous to make things ‘unnecessarily’ complex. With cultures like that, i would anticipate their native mathematics to be non-existent in a written capacity, or just as simple as a ‘this many’ counting system. I am, however, not prepared to be bold, stating any of this as fact without further research.

I am doing a project for school and I wanted to know how to write really big numbers as roman numerals. Numbers bigger then 1 million.

the exact number i am trying to write in roman numerals is 602000000000000000000000

(6.02 X 10^23)

if you could help me out that would be great

There weren’t really any normal Roman numerals above 1000, although there were some unusual techniques used occasionally that could get you as high as 100,000,000. But numbers of the size you’re talking about just weren’t expressible in Roman numerals – they weren’t necessary!

It would be a VI with 23 lines above it followed by an II with 21 lines above it.

Or, if you don’t want that many lines above the numbers, you could put a little box containing the number of lines meant to go over the number.

}{