september has the word “seven” tucked inside it, yet it is the ninth month of the year. october means “eighth”, and it is the tenth. november and december drag the same gap along: their names promise a number they no longer hold. it is no translation slip and no coincidence. it is the fossil trace of a calendar the romans broke with their own hands, half from superstition, half from political corruption.
according to the chronicles, the first roman calendar had only ten months. it began in march — the month of mars, when the season of war and ploughing reopened — and ended in december. the numerical names fitted perfectly back then: the seventh month was, indeed, the seventh. the problem was what was left out.
what happened to the more than sixty days of midwinter that fell between the end of december and the start of the new march? quite simply, they were not counted. for a farming and military society, winter was dead time: no one sowed and no one waged war, so the calendar ignored it altogether. it was a nameless limbo, an administrative void of two long months in the middle of the year.
tradition credits numa with the first patch, adding january and february to plug the hole. but then a second obsession took hold, this time religious. the romans were deeply superstitious about numbers and believed even numbers brought bad luck, so they insisted that most months should have 29 or 31 days, never an even count. the trouble was that the sums refused to add up: to adjust the reckoning to a reasonable year, one month had to be even by force. the casualty was february, the last month of the old year: it was cut down to 28 days and, on top of that, loaded with the rites of purification and of the dead.
time became a weapon. if the magistrate of the day was an ally of the pontiffs, they lengthened his year; if he was an enemy, they shortened it.
even with the patch, the roman year drifted hopelessly out of step with the seasons. to correct the drift, the pontiffs — the priests who guarded the calendar — had the prerogative of inserting an extra month from time to time, the mensis intercalaris (also called mercedonius), slipped inside february after the feast of the terminalia, splitting the month in two. in theory, it was a technical mechanism for astronomical adjustment, but also one open to abuse. ancient sources such as macrobius and censorinus charge that the pontiffs manipulated the intercalation for political ends: since magistrates held their offices for a year, lengthening or shortening it had direct consequences, and they were accused of slipping the month in to hand extra weeks to an allied consul, or of dropping it to push a rival out sooner. modern historiography confirms that the mechanism was vulnerable to such abuse, but the evidence of systematic manipulation is scarce and debated: much of the drift was also down to negligence and to war.
the result was chaos. by the middle of the first century bce, the accumulated manipulations and neglect had pushed the civil calendar nearly three months out of step with the sun: the equinox fell where it should not, harvests no longer matched their festivals and no one knew any longer what day they were living in. it took the absolute power of julius caesar, advised by sosigenes of alexandria, to impose, in 46 bce, a radical reform — the julian calendar, with its 365 days and its leap day — that restored order and, with slight later adjustments, still governs our lives. the misplaced names from september to december are the only thing that survived of that broken system: a fossil reminder of when, in rome, even time was for sale. and that same capacity of kings and elites to bend institutions to their will was about to provoke the crime that would set the whole monarchy ablaze.
