EPOCH Tutorials - Using and Interpreting Late Medieval French Pardon Letters, the ‘JJ series’

EPOCH
9 hours ago
5 min read

Eleanor J. Bailey │ University of Sheffield

Interested in pursuing your own research using pardon letters from late medieval France? They can be found in the ‘JJ series’ of the Archives Nationales, Paris, which constitutes the records produced by the medieval chancery of the French crown. Almost all the fourteenth-and-fifteenth century records have recently been scanned and digitised through the HIMANIS project for public access.

The records of the medieval French chancery were kept in the ‘Trésor des Chartes’ (Treasury of Charters), a physical building in Paris near the Palais de Justice and Sainte-Chapelle that lends its name to this archive. The keeping of these archives originates in the early thirteenth century, and despite their history as royal records they survive as an original composition after being spared during the French Revolution (and later catalogued alphabetically). One-hundred-and-ninety-nine manuscripts from the medieval chancery, including the ‘JJ series’ of one-hundred-and-seventy-nine registers dating from 1302 to 1483, are now publicly available online, which amounts to 83,000 images of text.

In their surviving form the registers take the form of bound books, the leaves of which were assembled in rough but imperfect chronological order as different bundles of letters were copied into the chancery record by scribes. The petitions for pardon and decision-making around delivering a positive verdict would have passed through the head of the chancery called the chancellor (or chancelier) and his subordinates. The ensuing letter of pardon from the chancery was distributed to the local garrisons and administrative centres of royal rule for enactment and copied into the chancery’s records for the crowns’ safekeeping. Registers JJ70-JJ187 roughly cover the period of the Hundred Years War. From King Henry V’s landing in Touques to begin the conquest of Normandy on 1 August 1417 until 1437, when the English lost control of Paris and the garrison was compelled to leave on 17 April, the researcher should turn to registers JJ170-175.

The registers for the period 1417-1437 are a mixture of predominantly Middle French with some Latin, and pardon letters were always written in Middle French. This is because these documents were intended for distribution and thus had to be comprehendible by local officers in the vernacular tongue. The pardon to Guillaume du Croq on the 5 April 1431, for example, was addressed to the ‘bailli of Rouen’ and his various officers and lieutenants. Baillis in France served an approximate role to the escheator and sheriff in England, meaning they were appointed by the king to serve as local administrators in matters of finance, land ownership, and justice in the courts. Baillis were the main point of contact between central and local governance, which included responsibility for royal pardons, freeing those who had been imprisoned, returning their possessions to them, and receiving oaths of allegiance on the King’s behalf in return.

The letters were also highly formulaic. They always ‘imposed perpetual silence on our prosecutor’ (‘imposi sur ce silence perpetuel a notre procureur’) of the crime committed (effectively striking it from the public record), the subjects were almost always restored to ‘good name and renown’ (‘bonne fame et renomee’), and grace was almost always imparted ‘because mercy is preferred to the rigour of justice’ (‘misericorde preferer a rigueur de justice’).

Despite these conventions reading Middle French pardon letters can be tricky (see the ‘Dictionnaire du Moyen Français’, 2023, available online at: http://zeus.atilf.fr/dmf/, for help identifying archaic French words and phrases). Scribes often abbreviated common words to save space and time, and transcribing these sources requires researchers to recognise patterns of abbreviation. In the below example, the scribe has condensed the phrase ‘entendement que pour’ (‘understanding that for’) with three common short hands. ‘Entendement’ is written ‘entendemt’ with a horizontal tilde stroke to indicate the missing letters. The word ‘que’ is abbreviated as the letter ‘q’ with a rounded stroke above it, and ‘pour’ (‘for’) as ‘po’ with a flourish that looks like a letter yogh ‘ʒ’. Once known these abbreviations are easy to spot, but fluency in these documents takes time.

This image is a screenshot of a manuscript in Middle French scanned and digitised through the online HIMANIS Chancery project. The manuscript is a register of chancery of the French crown from the ‘Trésor des Chartes’ (Treasury of Charters) and depicts the pardon letter obtained by Guillaume du Croq in April 1431. The screenshot shows an example of three common scribal abbreviations in the Middle French phrase ‘entendement que pour’. — Screenshot of the pardon to Guillaume du Croq, Trésor des Chartes register JJ175, n. 88 (p. 62 online), digitised through the online HIMANIS Chancery project, http://himanis.huma-num.fr/app//index.php/ui/show/chancery/349/62.

However, to make the transcription process easier the digital HIMANIS platform utilises probabilistic indexing technology. Once a page in a document has been selected, double-clicking a word will produce a list of likely translations with a confidence percentage. This can be incredibly useful, though confidence rates vary dramatically across documents and the researcher must often exercise their own judgement to fill in the gaps themselves. In the below example, the probalistic indexing picks up on ‘debilite’ (debilitated) but completely misses ‘membres’ (limbs).

Screenshot of a manuscript in Middle French scanned and digitised through the online HIMANIS Chancery project. The manuscript is a register of chancery of the French crown from the Trésor des Chartes (Treasury of Charters) and depicts the pardon letter obtained by Guillaume du Croq in April 1431. The screenshot demonstrates the probabilistic indexing feature used in the digitisation process. Words can be double clicked and highlighted to generate a drop-down table of likely translations with a confidence percentage, though by comparing the two images it is made clear that the scanning software does not consistently identify written words.

The probabilistic indexing allows for some search capabilities, though the varying rates of confidence mean that results are never comprehensive. Despite these limitations, however, the digitisation of the Trésor des Chartes archives allows researchers to access a huge amount of material on the lived experiences and testimonies of those living in the French royal domain in the late medieval period. By pursuing them, historians will be able to uncover in piecemeal more about what normal day-to-day life and exclusion from this looked like in the late medieval period. For an example of how this can be done, see the EPOCH Feature by the author.