A note on Roman numerals
IIX The irregular eight
Long before the school rules were settled, eight could be written as two taken from ten. A short note on that numeral, the people who carved it, and the other forms that never quite went away.
- IIXThe numeral
- X − IITen, less two
- VIIIEight
Two from ten
Most of us learned one tidy rule. Eight is VIII: five, then three more. Nine is IX: one taken from ten. Only a few pairs may subtract, and only one small sign at a time.
The Romans never agreed to that rule. Usage varied a great deal in antiquity, and even more in the Middle Ages. A writer was free to set two strokes in front of a larger sign and take both away. Read that way, IIX is ten less two, and there are many old examples of it. It is not a slip of the chisel. It is simply another habit.
Counting down
For its eights and nines, Latin counted down rather than up. Eighteen was duodeviginti, “two from twenty”, and nineteen was undeviginti, “one from twenty”. The written numerals sometimes counted down as well.
Officers of the Eighteenth Legion wrote their number as XIIX: ten, and then two from ten. The form appears on the cenotaph of the centurion Marcus Caelius, who fell with Varus in AD 9.
LEG · XIIX
Some of the public calendars known as the Fasti did the same, writing XIIX for the eighteenth day before the Kalends. The one calendar to survive from before Julius Caesar gives February its twenty-eight days as XXIIX.
The freedom had a price. Men of the Twenty-second Legion wrote their number as IIXX, after the spoken “two and twentieth”, and at least one ancient stonecutter seems to have read it as eighteen.
The watchmaker’s four
Look at a clock with Roman numerals and you will very often find IIII where school says IV, while the same dial still shows IX for nine. The habit goes back to very early clocks, such as the one in Wells Cathedral from the late fourteenth century.
It is far from universal. The clock in the Elizabeth Tower at Westminster, home of Big Ben, uses a plain IV. Many reasons are offered for IIII, from a better balance with VIII across the dial to simple tradition, and none of them settles the question.
The mix is older than any clock. The surviving entrance numbers of the Colosseum run from XXIII to LIIII, and gate forty-four is marked XLIIII: subtraction for forty, but none for four.
Other ways to write a number
A few of the forms that turn up in inscriptions, on calendars, in old books and on clock faces, next to the ones taught today. The strokes in gold are the ones taken away.
| Number | Today | Also written |
|---|---|---|
| 4 | IV | IIIIClock faces; the gates of the Colosseum |
| 8 | VIII | IIXMany old examples |
| 9 | IX | VIIIIThe additive habit, used less often |
| 17 | XVII | IIIXXRare, but recorded |
| 18 | XVIII | XIIX, IIXXXIIX for the Eighteenth Legion and the Fasti; IIXX in some seventeenth-century dates |
| 22 | XXII | IIXXThe Twenty-second Legion, after the spoken “two and twentieth” |
| 28 | XXVIII | XXIIXA Roman calendar, for the days of February |
| 40 | XL | XXXXOften, in old texts |
| 44 | XLIV | XLIIIIGate 44 of the Colosseum |
| 90 | XC | LXXXXThe additive habit, used less often |
About the name
This page takes its name from the numeral. IIX is eight, counted back from ten. A Roman reader would probably have taken it in at a glance, and any modern rule would mark it wrong.
That gap, between what people actually wrote and what the rules later said they should write, runs through the whole history of Roman numerals. The rules came afterwards. The numbers were already there.