Several problems arise in interpreting dates in Latin documents, especially before the adoption of the Gregorian calendar reforms, which occurred at different times in different countries. Sometimes the best we can say about a date found in a medieval document is that it contains an error!
The year itself is sometimes in question, because sometimes the date of the new year was reckoned according to local or personal preference from March 25 (March 24, 1492 was followed by March 25, 1493), at other times from January 1 (December 31, 1492 was followed by January 1, 1493), and occasionally literally "computed from the Nativity of the Lord" (December 24, 1492 was followed by December 25, 1493). A collection of documents may turn out to contain records computed under two or more systems, originating from different courts or institutions. In Latin church records from the time of Napoleon, it is even possible to find dates from the French Republican Calendar!
Dates are frequently specified by reference to a date in the liturgical calendar that has just occurred, or that will soon occur, such as "the Tuesday after Easter 1407" or "the Wednesday before Christmas 1419". In some cases, the date referred to is a fixed date, such as Christmas or the feast of the nativity of St. John the Baptist, in others it is a movable feast such as Easter, Ascension, or one of the Sundays in Lent. To establish these dates, we need an accurate calendar for the year, starting on the correct date and correctly identified as a common year or leap year, a method for determining the date of Easter, a table of the other movable feasts, and a table of the principal fixed feast days. Sometimes the saints mentioned in dates are not widely known, or else not described in enough detail to be sure which one of that name was meant.
Years are usually expressed in the ablative case, preceded by Anno Domini. The year is often written out in words, frequently abbreviated, or else in Roman numerals or a combination of words, abbreviations, and Roman numerals. It is not uncommon to find Roman numerals that do not follow modern practice, such as mo iiijc xlviiii for 1449 or mo iiiic lxxscdo for 1472. The words used in Latin dates are generally the "-th" forms, known as "ordinal numbers". In English, they are first, second, third, fourth, etc. Note that "ordinal numbers" in Latin are generally used as adjectives. They follow the usual -us, -a, -um declension pattern for adjectives. The words typically encountered in expressing years in Latin are the following (all in the ablative case, masculine singular to agree with Anno Domini):
The names of the month are usually in the genitive case, in a phrase such as "mensis Augusti", in the month of August. They are frequently abbreviated, but can almost always be distinguished even when badly faded or damaged. In Roman times, the names of the months were adjectives, and so they appeared with different endings in some situations. While the medieval documents I have seen treat the names of the months as nouns (and thus use them in the genitive case), it is possible that you will find an exception, perhaps if a scribe had learned his Latin from a classical scholar.
The days of the month are again "ordinal numbers", but they are usually found with feminine ablative endings, agreeing with Die, the ablative singular of Dies, which is supposed to be masculine, but usually used as a feminine noun when expressing dates. Thus, we usually see days of the month expressed like this: Die trigesima mensis Augusti — with the day as a feminine ablative adjective, and the month in the genitive case. However, other constructions are possible. For example, in the situation where a document is cited that was created on a particular date, we might find "diei", the genitive case of dies, followed by the day of the month in the genitive case: diei trigesimae (or Medieval trigesime) mensis Augusti. The days of the month are therefore the same as the corresponding numbers used to express years, but with feminine endings. However, in some areas and during some periods, you may find the days of the month with the masculine endings instead, if dies has been treated as a masculine noun.
In Roman times, days of the month were expressed by a system that today seems unnecessarily obscure and complicated, referring to the number of days until the next "special day" in the calendar. This system continued to be used in at least some documents as late as the middle of the 14th Century. The words to look for are kalendas (usually abbreviated, something like kl or kal with a line through the word), nonas, or idus, in place of the word mensis. However, the letter k is encountered so seldom in Latin, you might not recognize it. When you can't find the word mensis in a date, and there is some other word or abbreviation in its place, suspect kalends, nones or ides! (Kalends, nones, and ides are the English words for kalendae or medieval kalende, nonae or medieval none, and idus, respectively. These words are always plural, and in theory should be in the accusative case, as kalendas, nonas, and idus. One explanation for these "special days" is that the kalends was originally the day of the new moon, the ides the full moon, and the nones the half moon preceding it, these days having religious or ritual significance.)
Another wrinkle is that the day before the kalends, nones, or ides is indicated by the word pridie. Pridie kalendas februarii is the day before February 1, or in other words, January 31.
Once you realize that you are looking at such a date, the next problem is how to convert it into a "normal" date. A typical example is likely to be written something like "die viii kal Augusti anno dni mo ccco xxxiiio". The kalends of August is August 1. Count backwards 8 days from August 1, counting August 1 as the first day (inclusive counting). The result is July 25, 1333. The same system applies to dates reckoned from the nones or ides; you count backwards with "inclusive counting".
The kalends always falls on the first of the month. For the nones and ides, the situation is slightly more complicated. For the four months that were originally "long" months in the pre-Julian calendar (March, May, July, and October), the ides falls on the 15th, and the nones 9 days earlier (by inclusive counting — nones comes from the word for 9), on the 7th. For all other months (January, February, April, June, August, September, November, and December), the ides falls on the 13th and the nones again 9 days earlier (by inclusive counting), on the 5th. Another oddity occurs in leap years, where the actual leap day was on one of the two days which together would be reckoned as February 24 (see below). The first of these two identically numbered days (but considered the second, because the system always counts backwards from the kalends) is written die bis vi kal marcii, literally the second of the sixth day before the kalends of March. But the easiest way to convert the dates is to use a calendar conversion web site such as this Calendar Utility (select the second line, and use the dropdown menus to enter the day, kalends etc., month, and year, then click on Start). Wikipedia articles on Roman Calendar and Julian Calendar provide useful background.
Given the existence of multiple dates for the new year, the question naturally arises, what to do in leap years? As best we can determine, for the purposes of determining which year was to be a leap year, the year was taken to begin on January 1. Further, the "intercalary day" in such years was placed after February 23, resulting in two days numbered February 24 (but without interrupting the normal sequence of the days of the week). At some later date, the intercalary day was moved, and designated February 29. Exactly when this happened is not clear (to me, at least), but it may have been in the late middle ages. Meanwhile, the century years continued to be counted as leap years until the Gregorian calendar reforms in 1582, which placed the additional restriction that only the century years divisible by 400 were leap years. Thus, 1700 was the first century year that was not a leap year. In order to retain some sense of sanity in medieval dates, it seems best to imagine that the days of February in leap years were numbered 1 through 29, unless a date is encountered (in the range February 24-28 in a leap year) that requires further explanation. When the date is stated to be "the last day of February" in a leap year, it seems most logical to call it Feburary 29! Note that the Feast of St. Matthias (not St. Matthew) is affected by this problem. It was traditionally celebrated on February 24, but in leap year it moved, either to the second day numbered 24, or, once the days in the leap year month of February were numbered 1 to 29, to February 25. But if we find a document dated on the Feast of St. Matthias, 1407 (style of the Annunciation), we would probably report it for purposes of history as February 25, 1407 (Old Style), because that month fell in the leap year 1408 computed from January 1 — St. Matthias would be celebrated on the day number 25, because that month was part of the leap year beginning January 1, 1407 (written in the style of the Annunciation).
Finding calendars on the internet: The calendars used by Wikipedia appear to be correct. The calendars prior to the Gregorian reforms (1582) correctly show that the century years up to 1600 were leap years. Exactly what happened during long period when the Gregorian reforms were adopted by one country after another is another topic entirely; you will have to investigate carefully to be sure what happened in any particular place! There are other internet calendar sites that "forgot" that all the century years prior to 1700 were officially leap years, and therefore all of their calendars up to and including 1500 will be wrong! "Calendar Donkey" is one of the sites that made this mistake! See below (Movable Feasts) for a link to the correct dates of Easter, which match the calendars on the Wikipedia.
The days of the week are normally found in the genitive case, even when the rest of the date is in the ablative case (the list traditionally begins with Monday).
Some Catholic church records, notably in southern Germany, and possibly elsewhere, include dates in the French Republican Calendar during the period when Napoléon was active in Europe. In all the examples we have seen, the date was also given in the "normal" Gregorian calendar, but it would not be surprising to find documents where the only date is expressed in the French Republican Calendar, and the date is written in Latin. For an explanation of the calendar, see this site (opens in a new window, to prevent problems getting back to our site!), which includes a simple and reliable date converter (opens in a new window, for the same reason!).
The Latin expression for the "French Republic" is "respublia gallicorum", the Republic of the French (plural, meaning the French people). Respublica is a compound word, originally two separate words, "res publica". As a compound word, each part of it declines separately. The word res ("thing") follows the fifth declension of feminine nouns, while publica is a feminine adjective following the usual -us, -a, -um pattern.
In Latin, these are normally encountered in the genitive case, because they are cited in a phrase such as "on the day of", "on the feast of", "on the Tuesday after the feast of", etc. Therefore, we list them here in the genitive case. Alternative spellings (medieval) are in parentheses. Any of the words ending in "-tionis" can also be spelled "-cionis". Today Sanctus is the usual title, but in medieval records we often see Beatus. These are the most common ones that we have found in dates, but different saints may be more popular in different regions.
The "Octave" of a feast day is exactly one week after the date of that feast, whether or not it was a movable feast. Various liturgies might be prescribed for the "Octave" of a particular feast. In the Roman tradition, dates were expressed by "inclusive counting". If something was supposed to occur one week after the present date, that duration was expressed in Latin as "in eight days", just as is done in modern French, for example — the present day is counted as 1, rather than as 0. For two weeks, both Latin and French would say "in fifteen days". When confronted with Latin statements about durations and dates, remember "inclusive counting" !
The "Vigil" of a feast is the day before that feast, in English the "Eve" of the feast day. Thus, All Hallow's Eve (Halloween, October 31) is the eve or vigil of the feast of All Saints (November 1).
"Feria" (genitive Feriae or Ferie) is a word with an obscure history. It should always be followed by a number: secunda, tertia, quatra, quinta, or sexta. Feria secunda = Monday, feria tertia (tercia) = Tuesday, feria quatra = Wednesday, feria quinta = Thursday, feria sexta = Friday. Evidently the reason for this convention was to avoid using "pagan" names of the days of the week. You might find a date such as "die feria tercia post octabas Purifficationis Beati Marie Virginis", meaning, the Tuesday after the octave of the feast of the Purification or Candlemas. Candlemas falls on February 2, the "octave" is one week later, February 9, and you would then find the calendar for the year in question and locate the first Tuesday after February 9.
The date of Easter, in turn determined (roughly speaking) by the date of the first full moon after the (northern hemisphere) spring equinox, determines the date of many feast days in the liturgical calendar. In fact, the original date of the equinox was March 20 when the system was initiated in 326 AD, and the full moon dates are "ecclesiastical full moon" dates, not actual astronomical full moon dates. But the effect is that Easter can occur as early as March 22 or as late as April 25. See the convenient table compiled by Father Kevin Michael Laughery. Over the centuries, there have been multiple opinions about various items needed for the computation of the date of Easter. More than once, we have found documents or church records that imply a date for Easter that does not match any of the known variants. However, before the Reformation, and before any of the calendar reforms, in the part of the world subject to the Church of Rome, the computation should have been the same over a period of about 12 centuries.
It seems logical to use a calendar with the first day of the year on January 1 (or December 25), so that there will always be exactly one Easter for each calendar year. If the first day of the year is taken as March 25, it would be possible to have two Easters in the same year, because Easter can occur as early as March 22. But this logic may not have impressed the notaries of another age, who were accustomed to begin the year at the Annunciation, March 25. What did they do when they needed to date a document by reference to Easter in a year when one fell after March 25 and the next one fell before the next March 24? If anyone has found a document that bears on this question, we would like to see it! The sequence of the liturgical calendar, which begins at Advent, would not be disturbed; it is only when one dates a document with something like "die Lune post festum Pasche" in a year figured from the Annunciation that an ambiguity could arise, and only then when Easter was unusually early, such as 1410 and 1421 (March 23, with the previous Easter falling on April 7 of the same Annunciation year).
The "movable feasts" most often encountered in dates in Latin documents are those computed from the date of Easter. The "movable feasts" are generally those that fall on a Sunday, so we normally find either "in festi" plus the name of the feast (genitive case), or else "die Jovis (or some other day of the week) post (or ante) festum" plus the name of the feast, or else "die Jovis post Dominicam" plus the genitive, or "Dominicam de" plus ablative of the name of the feast. For those feast days whose names consist of the first word or words from the introit, those words are not declined, resulting in names such as Dominicam de Oculi Mei, Dominicam de Judica Me, etc. Remember that abbreviations are always possible!
The Sundays of Lent are usually identified by the first words of the traditional introit for that feast day. However, there are alternate names for some of these Sundays. We include here also the three Sundays before Lent, the "Septuagesima" period. While the names of these three Sundays mean 70th, 60th, and 50th, the first two are misnomers, since even by "inclusive counting", they are actually only 64 and 57 days prior to Easter.
The Sundays (and two Thursdays) of the Easter season also have peculiar names. The feasts in this sequence that fall on Thursdays (Ascension, Corpus Christi) are often moved to the following Sunday in modern practice.
A hint for those who want to compute the movable feast days using Microsoft Excel: Unfortunately, the designers of Excel made the stupendous error of imagining that dates prior to 1900 were of no importance in this age of computers! In order to "fix" Excel so that it can interpret and manipulate dates prior to 1900, you will need additional programming or an "add-in". The one that works for me is XDate by J-Walk and Associates, Inc. It was easy to install, and it made the problem of computing movable feast days during medieval times easy. It got the leap years right, too! It is possible that later versions of Excel no longer have this problem, so you may want to experiment before installing an add-on.
Found on Google Books, a very detailed glossary or dictionary of medieval calendar terms: R. T. Hampton, Medii Aevi Kalendarium, vol. 2 (London: Henry Kent Causton & co., 1841), "Book IV: Glossary".
As always, please let us know if you find documents that show other ways of expressing dates, additional feast days, etc.!