Historical mysteries · Read from the letters, 1669 to 1703
The letters that passed between the prisoner's jailer and the King's ministers from 1669 to 1703, the Bastille's own registers and the burial entry, read in full and tested against twenty proposed candidates.
On Thursday 18 September 1698, at three in the afternoon, a litter came through the gate of the Bastille, the royal fortress-prison in Paris. In it sat the prison's new governor, Bénigne de Saint-Mars, who had guarded state prisoners for the French crown since 1665, first in the Alps and then on an island off the coast of Provence, and beside him a prisoner whose face was covered by a mask. Étienne du Junca, the King's lieutenant at the Bastille, who kept the book in which every arrival was entered, wrote that Saint-Mars had brought "an old prisoner he had at Pignerol, whom he keeps always masked, whose name is not said". The man waited in a tower room until dark. At nine that night du Junca himself led him to the third chamber of another tower, furnished some days before on Saint-Mars's orders, where a sergeant who had come with the governor would bring his food.
He was there for five years. On Sunday 18 November 1703 he felt unwell coming out of mass. He died the next evening at about ten, du Junca wrote, "without having had a great illness", still "the unknown prisoner, always masked with a mask of black velvet". He was buried the following afternoon in the cemetery of Saint-Paul, the prison's parish. The burial register gives a name du Junca called "as unknown" as the man, Marchioly, and an age, "forty-five years or thereabouts", signed for by two officers of the Bastille who had never seen his face. Forty livres were paid for the burial, and no paper of the Bastille gives his name. The one letter about him from the minister in charge of the prison calls him "your prisoner from Provence".
The state never said who the man was or why he was held, and writers supplied answers. In 1751 the writer Voltaire described the mask as having a chin-piece with steel springs, and later accounts called it iron; in 1771 he let it be understood that the prisoner had been an elder brother of Louis XIV. Alexandre Dumas made him the King's twin in a novel of the 1840s. Since 1745 some fifty identifications have been offered, and the historian Jean-Christian Petitfils counts fifty-two: a royal brother, a diplomat from Mantua, the Duke of Monmouth (Charles II's illegitimate son), the playwright Molière, the musketeer d'Artagnan, the disgraced finance minister Nicolas Fouquet.
The evidence comes from the years before the Bastille, and it is incomplete at one point. Saint-Mars had held the man since 1669: at Pignerol, a French fortress in the Alps (now Pinerolo, in Italy) where he guarded state prisoners until 1681; at Exilles, a smaller fort in the same mountains, until 1687; and on the island of Sainte-Marguerite, off Cannes, until 1698. All that time he wrote about his prisoners to Louvois, the King's war minister in Paris, who had charge of the state prisons; the minister wrote back, and hundreds of the letters were kept. They were first printed in 1801 by Pierre Roux-Fazillac, a former member of the revolutionary Convention, and every serious argument since has been made from them. The letters say where each prisoner was moved and when, but never why this one was held, and they fail at the one point where a name is needed: in January 1687 one of two unnamed prisoners died at Exilles, the other went on to the island and then to the Bastille, and the letter that said which of the two had died is lost.
Three answers have been given. Most historians who have worked from the letters since 1890 hold that the masked man was a prisoner whom Louvois called "only a valet", a manservant, arrested in July 1669 on the Channel coast and delivered to Pignerol that August under orders that no one should see him or hear him speak. Jules Lair, the biographer of Fouquet, argued this in 1890. Georges Mongrédien, a historian of the seventeenth century, narrowed the choice in 1952 to two men, this valet or the diplomat from Mantua, and would not decide between them. Petitfils (1970 and 2003), the writer John Noone (1988), the American historian Paul Sonnino (2016) and the biographer Josephine Wilkinson (2021) decide for the valet. The nineteenth century's answer, given by Roux-Fazillac and defended last in 1894 by Frantz Funck-Brentano, who catalogued the Bastille's papers, was that diplomat, Ercole Antonio Mattioli, whom the French had seized in 1679 for double-dealing over the sale of a fortress; the name on the burial register is three letters from the spelling Saint-Mars used for his. The popular answer, from Voltaire to the film-maker Marcel Pagnol, is a brother or twin of the King, for whom no document exists. Others have proposed a disgraced nobleman who bore the valet's name (Harry Thompson, 1987), d'Artagnan (Roger Macdonald, 2005) and Fouquet himself, his recorded death in 1680 being a fiction (Jean-Marc Dabos, 2025).
Those who accept the valet are divided by two further questions, which the letters do not answer. The first is who he had been. Petitfils makes him a servant of Henrietta of England, Louis XIV's sister-in-law, known at court as Madame, who in 1669 and 1670 carried the King's secret negotiation with her brother Charles II of England. Sonnino makes him a valet of the man who managed the fortune of Cardinal Mazarin, the King's chief minister until his death in 1661. The second is why he was kept for thirty-four years, and masked. Either he knew something that stayed dangerous, or, as Petitfils and Bernard Caire, who examined the original letters in the 1980s, have both argued, the secret was of small importance and the appearance of mystery was Saint-Mars's own doing.
This report takes the letters as the evidence and reads all of them. Every letter and register entry of the Pignerol, Exilles, Sainte-Marguerite and Bastille correspondence printed between 1801 and 1953 was read, one hundred and eighty entries in all, and where the editions disagree on a word both readings are recorded. From them seven dated facts about the man's custody were taken, each resting on a named letter or register entry: the arrest and delivery of 1669, the permission of 1675 to serve Fouquet, the confinement with Fouquet's other servant in 1680, the move to Exilles in 1681, the death of one of the two in 1687, the entry into the Bastille in 1698 and the death of 1703. Twenty candidates were tested against the seven. A candidate documented alive and free, in another prison, or dead at any of those dates fails.
Only the valet of 1669 passes every documented point, and one point is not fully documented: no letter says which of the two Exilles prisoners died in January 1687. The other was La Rivière, who had come to Pignerol with Fouquet as a free servant in 1665. From two later letters and the subject matter of the whole file this report argues that the survivor was the valet of 1669, and puts that at 85 per cent. Two other steps in the chain of custody carry their own weights, 99 and 93 per cent. Multiplied, they give 78 per cent that the man buried as Marchioly was the valet arrested in 1669, 14 per cent that he was La Rivière, 3 per cent Mattioli, and 5 per cent someone the letters never name. The weights are judgements about the documents rather than measurements, and the report shows how the headline moves when they change. On the two questions the letters do not answer, the report reaches no name for the valet's master, and finds that no minister ever ordered a mask: the mask was the jailer's practice from 1687, approved in general terms by the King in 1698, and the length of the detention follows the practice of the King's prisons, which had no procedure for releasing a man held for what he knew and in this file used none.
The panel below states the three answers with their probabilities. The report then sets out the method and the seven fixed points; follows the letters in order through the arrest of 1669, the eleven years at Pignerol, the years at Exilles, the missing link of January 1687, the island and the Bastille; tests the twenty candidates in a table; takes up who the valet was and why he was held; and closes with the strongest cases against the verdict, what would overturn it, what in it is new and what comes from earlier historians, what could not be read, and the sources.
Verdict
Figure 1. The four outcomes, weighted as in the text. Made for this report from the probability arithmetic below (three link weights, 0.99 × 0.93 × 0.85).
Probable The evidence clearly favours one answer over all the others, but does not prove it. The percentage moves when the same evidence is weighed differently. Which of the two Exilles prisoners died in January 1687 is decided by inference rather than by any document, so the headline is the product of three judged weights and is printed with its sensitivity below.
Not identified. No document names his master, his birthplace or his employment. The two current hypotheses, Petitfils's servant of Madame carrying letters in the negotiation that produced the secret treaty of Dover in 1670, and Sonnino's valet of Antoine-Hercule Picon, the manager of Mazarin's fortune, are both inferences from the same silence. Petitfils's one concrete name, a Gaspard Eustache baptised at Calais on 27 December 1643, is no more than a matching surname in a parish register; Sonnino writes that he found no Auger or Dauger among the Picons' servants. Against each other the two hypotheses are roughly even, and together they are less likely than an employment the record does not point to at all.
The letters say the secret was something he had been employed at or had done before 1669 ("ce à quoi il a été employé auparavant que d'être à Pignerol", 1678; "ce qu'a fait vostre ancien prisonnier", 1697), and that the object of the orders was that he should never speak of it, above all to Lauzun, a courtier held beside Fouquet at Pignerol from 1671 to 1681 whose release was expected. No minister ever ordered a mask. The mask is first attested in a rumour of September 1687, and its use at the Bastille was Saint-Mars's practice, approved in general terms by the King in August 1698. The long detention is the ordinary working of Louis XIV's state prisons, which released nobody whose secret was still thought live, continued under a minister (Barbezieux, who succeeded his father Louvois in 1691) who inherited the precautions, and a jailer whose posts and pay depended on continuing to hold him.
Agrees with the position most historians who have worked from the letters since 1890 have held, which is that the masked man was this valet. The question is still argued. Jules Lair (1890) first identified the masked man with the valet of 1669; Georges Mongrédien (1952) reduced the choice to Mattioli or the valet, showed that La Rivière died at Exilles in 1687, and declined to choose; Petitfils (1970, 2003), Noone (1988), Sonnino (2016) and Wilkinson (2021) decide for the valet, while Funck-Brentano (1894) held for Mattioli, Thompson (1987) for Eustache Dauger de Cavoye, the disgraced nobleman who bore the valet's name, Macdonald (2005) for d'Artagnan, and Sarah Madry, in a 2016 article in the Cahiers du dix-septième, a journal of seventeenth-century studies, argues that the word "valet" itself cannot be relied on. This report reaches the valet at a lower confidence than his proponents state, because one link in the chain, the death of one of two unnamed prisoners at Exilles in January 1687, is decided by inference and not by any document.
What moves the headline. The 78% is the product of three links: that the Bastille prisoner of 1698 was the man Saint-Mars brought from Sainte-Marguerite (put at 99%), that the island prisoner of 1698 was the prisoner brought from Exilles in 1687 (put at 93%, against the thesis of Stanislas Brugnon, an archivist who reconstructed the island's prison register, that the island prisoner died in 1693), and that the 1687 survivor was the 1669 valet rather than La Rivière (put at 85%, on the strength of Barbezieux's "twenty years" of 1691 and "what he has done" of 1697). Read the 1691 letter as a loose round number (60%) and the valet falls to about 55% with La Rivière at about 37%. Give Brugnon even odds and the valet falls to about 42%. The 78% and every figure derived from it are conditional on those three weights, which are judgements about the documents rather than measurements; the arithmetic is published with the working files for this case. The candidate table and the sensitivity are below.
The beginning and the end of this prisoner's custody are documented, and the years between them are disputed. The end is the Bastille: du Junca recorded on 18 September 1698 that the new governor Saint-Mars had arrived with "an old prisoner he had at Pignerol, whom he keeps always masked, whose name is not said", and on 19 November 1703 that "the unknown prisoner, always masked with a mask of black velvet" had died. The parish of Saint-Paul buried him the next day under the name Marchioly, aged "forty-five years or thereabouts". The beginning is a letter of 19 July 1669 in which Louvois tells Saint-Mars to prepare a dungeon at Pignerol for "the man named Eustache d'Auger", adding that "as he is only a valet" he needs no great furniture. The question is whether the same man stands at both ends, and, if so, who he was and why the state kept him.
The method is the one Jules Lair used in 1890 and Georges Mongrédien in 1952, applied to the whole printed record rather than to extracts. Every letter of the Pignerol, Exilles, Sainte-Marguerite and Bastille correspondence that the nineteenth- and twentieth-century editions print was read. The editions, each the work of a historian who printed part of the file, are Roux-Fazillac (1801), Joseph Delort (1825 and 1829), François Ravaisson's Archives de la Bastille (vols. III and VII to XI, 1866 to 1880), Théodore Iung (1873), Jules Loiseleur (1882), Marius Topin (1883), Lair (1890), Funck-Brentano (1894 and 1901), Émile Laloy (1913), and Mongrédien's article of 1953; the report cites them below by surname, volume and page. One hundred and eighty entries were transcribed with the archive reference each edition gives, and the variants between editions were tabulated (the transcriptions and the tables are published with this report as working files). From that record seven fixed points were taken, each resting on a named document:
A candidate who is documented alive and free, or in another prison, or dead, at the date of any fixed point fails that point. The age on the burial register is treated as a weak test, because the men who supplied it did not know the prisoner and the same entry carries a false name. Points (a) to (d) and (f) to (g) are documented as events. Point (e) is documented as an event but not as an identity: no surviving letter says which of the two Exilles prisoners died. Point (e) is therefore argued separately below.
Two cautions on the record. First, no original manuscript was consulted. The letters survive in three places: the Archives nationales series K 120 (Saint-Mars's own file of the letters he received, to 1681), the war-office registers now at the Service historique de la Défense, the French army's historical archive at Vincennes (series A1, which the editions cite as "P. x, v. y, Dépôt de la guerre", page and volume), and the Bastille's own registers of entries and exits, now in the Arsenal library in Paris (mss. 5133 and 5134). All were read through the editions, with each edition's page and reference recorded. Six documents were also seen in the photographs of the originals published on masque-de-fer.com by Jean-Marc Dabos, who runs the site. Second, the editions disagree on words that matter. Where a reading changes the sense, both readings are given, and the report says which is the more literal transcription.