Lady Day (and the vexing matter of assigning dates to documents)

I just realized that today is Lady Day – that is, Annunciation. Once upon a time, this was New Year’s Day. It probably derives from the date being originally set on the Spring equinox, which makes a pretty sensible first day of the year if your concept of the world derives from observations of the sun. (Not that the autumn equinox nor the solstices would be any worse, but Spring = growth = life = youth and we all grow old and die, so I would argue Spring is the more likely candidate for ‘beginning of the year’ than the other seasons.)

At the moment I’m thinking about dates. Or to be exact, I am working on dating documents. One might think that dating was one of the easier tasks in creating an edition – the difficulties of determining dates of some documents notwithstanding. As I’m working on an edition of, mainly, letters, happily they almost always include a dateline.

If only it was this simple. I have found to my cost that there were three problems which I needed to solve.

1) New Style vs. Old Style

a. The Gregorian Calendar

In the early 1600s, two (main) calendars were used in Europe: the Julian calendar, dating back to Antiquity, and the Gregorian calendar, newly installed by Pope Gregory XIII in 1582. The Gregorian calendar was designed to deal with the problem of solar years being slightly longer than 364 days, which the Julian calendar failed to adequately compensate for, for which reason by 1600 festivities originally set by solar dates (namely the equinoxes and solstices) had drifted 10 days in the solar calendar. In addition to improving the correlation between the numerical and solar calendars, the Gregorian calendar shifted the entire year ten days forward in order to fix the present discrepancy, so that ten days were “skipped” or “lost”. (Or twelve days in the case of UK, which adopted the Gregorian calendar in 1752, by which time the discrepancy between calendars had increased to twelve days. This, by the by, is why the fiscal year in the UK starts on 6 April: the commercial sphere found it easier to stick to a 365-day period for calculating annual rents, taxes, etc, than to skip the 12 days and recalculate everything.)

Anyway, there were two obstacles for all of Europe to happily adopt the new calendar. First, the Gregorian calendar was established by the Pope – and this being the period of the Reformation, many Protestants states did not rush to embrace things of Catholic origin. Second, old habits are hard to kick, and it took time for innovations to spread and take root. Long story short, Catholic Europe adopted the new calendar immediately or very soon, and the Protestant north did not. England did not convert to the Gregorian calendar until 1752, so that for nearly two centuries the British Isles and most of continental Europe followed different calendars. (Interestingly, Tuscany hung on to the Julian calendar until 1750, and the United Provinces of the Netherlands adopted the new calendar in 1583, so this matter did not blindly follow the religious divide.)

The Gregorian and Julian calendars very quickly became to be referred to as “New Style” (NS) and “Old Style” (OS) (or some other variants thereof).

b. New Year’s Day

The other (to a modern eye, annoying) problem with pre-Gregorian dates is that the beginning of the year was calculated from 25 March (see top). [ETA: But see the bottom of this blog entry.] In other words, a full year ran from 25 March to 24 March. Thus, for dates between January 1 and March 24, the year is one less than in our modern reckoning. However (and here it gets interesting), starting with Venice in 1522, European states slowly started to adopt calculating the beginning of the year from January 1. Much of the Catholic block had already shifted to this reckoning years before the (creation and) adoption of the Gregorian calendar.

To sum up a bit, here is a comparison of dates, depending on where the writer (dater?) was located:

      England (OS):   15 January 1605    26 March 1606    24 December 1606
      Spain (NS):       25 January 1606    5 April 1606           3 January 1607

However, the above examples are drawn from a theoretical ‘official’ calendar. In practise, many in England, and many Englishmen on the continent, used New Style dates. The catch is that while some datelines do note that the date is “stilo antiqua” or “nova”, most dates are not accompanied by a mention of which calendar they are following. This is not a problem in many contexts where the style can be guessed, and especially between 26 March and 21 December, when at least the year will be the same. But as is evident from the above examples, a date in the problematic quarter could potentially be in either year.

The traditional way to counter this problem in historical studies is to use double-dating: each year is marked so that the year according to modern reckoning is clear. For instance, 1605/6 indicates 1606 in our reckoning. This double-dating is of course only used for dates falling in the problematic quarter. (Double-dating was also used by an increasing number of English writers before 1752.)

Having shown that establishing the correct date of a document is far from a trivial or simple matter, let us move on to the second problem.

2) Assigning dates to problematic or undated documents

This was partly covered in 1b. above, but keeping in mind the problem of establishing the dates of documents according to modern reckoning as outlined in part 1, this can be a daunting task. In practice, it boils down to comparison with other documents of (ostensibly) the same period, namely other letters, and the lists and abstracts mentioned under part 2. The clues include mentions of dates and contents of previous letters as repeated in other, clearly dated, letters, and in lists and abstracts. They also include comparing information in the problematic letters with known historical dates (such as the Gunpowder Plot). The editor may also have to resort to palaeographical analysis to determine the dating of letters – this is something I have done both for entire documents (where the handwriting was “too early” for the year attributed by the repository), and for specific instances (where others had previously read “1603” I established that the number was in fact “1605”).

Sometimes a document cannot be dated to a specific date. In these cases the editor has no choice but to also use span dates, and ante quem and post quem dates.

3) Deciding on a policy of dating

What is the date of a document, and what date(s) should the editor assign to documents?

For letters, these questions have a seemingly obvious answer: the date of writing. But what about letters started on day 1 and finished on day 3? What about letters with postscripts dated 5 days after the beginning of the letter? And do we need to consider the date of sending of the letter, instead of the date of writing? (Perhaps more relevant in more modern times, where postal markings reveal dates of conveyance.)

– The easy option is to use the last date in a letter, unless the editor feels that there is some justification in assigning a span date (“1-5” or “1×5 January 1605”).

But what about copies of letters (or other documents)? Does one assign them the date of the writing of the original document (see above), or is the date of composition of the copy the one to use?

– The text-based solution is the easiest, and perhaps makes the most sense. That is, to use the date of the original document, and mark the date of copying (or estimation thereof) in the notes. Using the date of the original document allows for the copy to function as a surrogate when the letters are viewed in chronological order, so that there are no leaps in the textual continuity.

Moving on, what about notes, memoranda and the like?

– Here a date of composition seems the obvious one to use.

But what about abstracts of letters (sent or received), or lists of correspondence (sent and/or received)? These often lack a date of composition, but do contain dates of letters they refer to.

– The only sensible approach would be to calculate a likely date of composition, based in part on the contents of the document (ie. on or, which is more usual, some time after the date of the last letter in the list or abstract).

Finally, a technical point: when creating a digital edition, the editor may be forced to assign specific dates when none are evident, in order to satisfy the requirements of the computing process. These should always be noted both in the notes for the document, and within the encoding.


This blog ran much longer than I thought it would. I suppose it is evidence of how the matter is far from trivial, and requires proper editorial consideration and attention. Yesterday I spent a similar while thinking about the ordering of documents in editions, and what kind of numbering or IDs to assign to them. Maybe tomorrow I’ll tackle misplaced sheets and letters – the disparity between archival ordering and chronological ordering can pose tricky problems..


ETA: regarding dating the New Year in Early Modern England

It had not crossed my mind to actually look at the evidence to see when people in Early Modern England referred to New Year. So I had a look in the Corpora of Early English Correspondence, searching for “new year”. And what did I find? In the CEEC, 27 hits for “new year”, almost all of them pointing to 1 January (in some cases the date could not be determined). In the CEEC Supplement, 1 hit for 1 January, and 1 for 25 March. (There may be more instances in the CEEC and CEECSU: this was a cursory search.)

The next step would be to try to determine from the CEEC or other corpora – or, since the CEEC corpora are compiled from editions (which often modernize dates), directly from manuscript sources, whether people changed the year in their datelines after 1 January or after 25 March. I mean people in England, and in particular people like those in CEEC who referred to 1 January as New Year’s Day. ..but maybe I need not do this now. I do need to remember to eschew generalisations which I can’t justify, but I can definitely make claims about the material in my edition, where this does not seem to occur: people seem to be consistent in using NS/OS according to their surroundings or, when deviating from their given practice, make this change explicit.

Leave a comment

Your email address will not be published. Required fields are marked *