In Cecil’s column about Roman numeral on clocks, he mentioned the old “balance” argument about “IIII” was used because it balanced out “VIII”. He also mentioned that the subtraction shortcut wasn’t commonly used in Roman times, and maybe the clock makers simply copied the Romans.
I vote for the “balance” argument. If it was simply a matter of copying the pre-shortcut method of the Romans, then why write 9 as “IX” and not “VIIII”?
What I don’t get is the speculation about how the Romans wrote the number 4. Surely that would be as well documented as anything from 2,000 years ago could be? There must be dozens (at least) of inscriptions that include that numeral.
At the end of the column he suggests that IX may once have been VIIII. But he doesn’t speculate on why it changed but IIII did not.
The most obvious hypothesis is perhaps that VIIII changed because it was so wide, and IIII is narrower (especially if the characters are not fixed width; V will usually be much wider than I). And, in turn, VIIII would have to sit next to VIII, whereas the biggest neighbour to IIII is III.
So I’m saying it’s possibly an aesthetic thing, but not about balance, just about fitting the characters in without them getting too close.
Wikipedia notes that Cecil’s questioning of the clockmaker’s suggestion (that IIII allows four castings of the same figure to produce all of the letters necessary) is rendered moot by a creative arrangement of the letters. That is, not casting four copies of “XVIIIII” but rather of “VIIIIIX”. They can then be separated into groups (taking advantage of the rotational symmetry of the X to turn IX into XI) so that each number on the dial appears cast as a single piece.
Powers &8^]