Historical figures with march birthdays: guide

Before there were years with numbers everyone shared, there were reigns, feasts and ages at death. Most birth dates for people who lived before civil registration were built from those three materials, and each leaves its own kind of error in the finished date. Knowing which material a date came from tells you which check to run.

What to take away

• A pre-modern date was often written as a position in a reign or a distance from a feast, and converting it needs the reign's start date and the local feast calendar.
• Many historical birth years were never recorded. They were computed backwards from an age at death, and the computation has a built-in width.
• An entry for a historical figure should say which material its date came from, because the right check depends on it.

Dating by reign

A regnal year counts from the day a reign began, not from 1 January. "The fourth year of the reign" therefore spans parts of two calendar years, and converting it needs the exact accession date, which is itself sometimes disputed. A document dated by regnal year alone gives a window of up to a year, narrowed only if a month and day are also given.

Two errors follow. A regnal year converted to the calendar year in which it began, without noting that it ran into the next, silently picks one side of the window. And a regnal year from one ruler applied to another's reign, easy to do where rulers share a name and a number, moves an event by decades while producing a perfectly plausible date.

Dating by feast

Medieval and early modern documents often give a date as a feast, or its eve or morrow, from the calendar of saints in local use. A fixed feast is a definite day in its own calendar, so the conversion is exact if the calendar is known. But feasts varied by diocese and by religious order, some were moved when calendars were reformed, and a few are movable feasts tied to Easter that fall on a different date every year. A date given as the Tuesday after a movable feast needs the year to be known before it can be resolved, which is circular when the year is what you were trying to find.

Material	What it gives directly	What it needs to become a date	Width of the result
Regnal year alone	A span across two calendar years	The accession date	Up to a year
Regnal year with month and day	A single day	The accession date and the calendar in force	A day, if the reign is right
Fixed feast	A day in the local calendar	Which local calendar	A day
Movable feast	A day relative to Easter	The year, and the Easter rule then in use	A day, once the year is fixed
Age at death	A birth year by subtraction	The death date and the meaning of "aged"	One to two years

The backwards birth year

For a great many historical figures no record of the birth survives, and the birth year in every reference work was computed: a tomb, a chronicle or a legal inquiry gives the age at death, and someone subtracted. That computation has a width. "Aged forty" means at least forty and under forty-one in modern usage, in older usage may mean in the fortieth year, and may be a round number offered because nobody knew. Formulae such as "forty years and more" were standard in some legal records and mean exactly what they say: a lower bound.

So a birth year derived from an age at death is a range, typically two years wide, and a reference work that prints a single year has picked one end of it. Two references disagreeing by a year on such a figure are not in conflict. They are reporting the same range from different ends. The famous-February page explains why year errors behave worse than day errors in modern records, and the same holds here, with the difference that the range was there from the start.

What a historical entry should carry

• The material the date came from: a register, regnal dating, feast dating, or an age at death.
• The conversion inputs used: accession date, feast calendar, Easter rule, or the death date and the reading of "aged".
• The width of the result, stated as a range where it is one.
• The edition or editor whose conversion is being repeated, since conversions differ.

The January index grades every entry by what it rests on, and a computed year rests on an inference and should say so. The March pillar explains why even the calendar those feasts sat in was an approximation, and the April page shows that the gap between an event and its record did not close when registration became civil.

Common questions

Is a birth year computed from an age at death worth publishing?

As a range with its derivation, yes. As a single year with nothing attached, it is a guess presented as a record.

Why do two reliable references give different years for the same medieval figure?

Usually because they resolved a regnal year, a feast or an age differently, and both were reasonable. Look for the working, not the answer.

Does any of this affect living people?

No. It affects how a monthly page should treat its older entries, and it is a reason to hold them to a different standard of precision than modern ones. The actors' page for February shows the opposite case, where the precision exists and should be withheld.