Death in Amsterdam

The last days of great men or women hold a morbid fascination for us, particularly when they are attended by mystery or just obscured by ignorance. Try as we may, forensic-standard evidence is seldom available to clarify these issues when the events took place more than two centuries ago (as we shall see, even the documents that survive can mislead and do not always bear the weight we want to place on them), and that I fear is the case with the death of Jean-Baptiste Perronneau.

We know that it happened in Amsterdam in November 1783. That much was established by the indefatigable Maurice Tourneux (1849–1917), who, despite being by profession an archivist and bibliographer rather than an art historian, wrote the first serious account of Perronneau in a series of articles for the Gazette des Beaux-Arts in 1896 which were issued also as an offprint in 1903. His work was absorbed by Léandre Vaillat (1878-1952, best known as a dance critic) and Paul Ratouis de Limay (1881-1963, a librarian with a particular interest in pastel, and Desfriches’s great-great-great-grandson), whose monograph of 1909 and in particular its reissue in 1923 became the standard work on the artist until Dominique d’Arnoult’s infinitely deeper catalogue raisonné appeared in early 2015 (but with a 2014 colophon). But every time I go back to Tourneux I discover things I had forgotten he knew, such is the unfairness of art history.

In fact, as regards Perronneau’s death, Tourneux relied on documents located by Nicolaas de Roever (1850–1893), the archivist for the City of Amsterdam who had more than a little interest in art, and was the co-founder of the journal Oud-Holland. Here is what Tourneux wrote:

Au cours d’un troisième séjour de l’artiste en Hollande, le 19 novembre 1783, le sieur Jean Martens se présenta devant le secrétaire de la ville d’Amsterdam et déclara que le sieur Jean-Baptiste Perraunot (sic), sans profession spécifiée, âgé de quarante-deux ans (sic), demeurant sur le Heerengracht, près de la Leidschestraat, était décédé « par suite de fièvre ». Bien que le défunt demeurât dans un quartier fort riche, il n’eut le lendemain que le convoi des pauvres, et fut enterré à la Leidschekerkhoff, cimetière situé près de la porte de Leyde. Lorsque l’aimable et regretté M. N. de Roever me communiqua le résultat de ses recherches dans les archives municipales dont il avait la garde, je crus à une erreur de transcription, en ce qui concernait l’âge du défunt; mais M. de Roever me confirma et me prouva plus tard de visu que le registre portait bien un 4 et un 2. Ainsi, et jusqu’à la minute suprême, tout ce qui a trait à la personnalité du peintre devait rester entouré de mystère et de confusion, et si l’allégation du registre des décès n’était pas démentie par d’autres actes non moins authentiques que celui-ci, elle serait de nature à justifier la méprise de Nagler, dont j’ai parlé au début de cette étude.

Avant d’expirer aussi loin des siens, dans quelque chambre d’auberge, Perronneau avait pu, du moins, confier ses dernières volontés à un Français de passage à Amsterdam et le charger de les transmettre à sa famille. Fils du compositeur languedocien qui, dans la pastorale de Daphnis et Alcimadure, avait devancé nos modernes félibres, Mondonville fils, né en 1748 à Paris, où il est mort en 1808, n’a pas été traité par les répertoires biographiques avec la même faveur que son père, et le récent Supplément, ajouté par M. Arthur Pougin au Dictionnaire de Fétis, se contente de nous apprendre qu’à dix-neuf ans Mondonville fils avait composé six sonates pour violon et basse et qu’il se faisait parfois entendre dans les concerts. Entre temps, il crayonnait volontiers, comme l’atteste un croquis à la mine de plomb du désert d’Ermenonville, daté du 19 juillet 1786, et annoncé il y a quelques années par un catalogue de librairie, et il ne se refusait pas à prêter en 1782 au Salon de la Correspondance fondé par Pahin de La Blancherie, le portrait de sa mère, peint par La Tour (collection Eudoxe Marcille), ainsi sans doute que celui de son père, possédé aujourd’hui par le musée de Saint-Quentin.

Chargé verbalement par Perronneau de ses dernières instructions (ainsi que l’atteste l’acte de partage de la succession), Mondonville avisa la veuve du peintre et l’Académie royale de la perte qu’elles venaient de faire. L’Académie ne s’émut guère de la nouvelle: de l’aveu même de Renou, rédacteur du procès-verbal, on « oublia » de notifier le décès de Perronneau à la séance du 20 décembre 1783, et ce fut seulement à celle du 10 janvier 1784 que la mention en figura au registre.

As you can see from the passage, Tourneux must have been working on Perronneau well before the date his articles appeared, as de Roever had died several years before. But the elements of a good story are all here: a pauper’s funeral (think Mozart), a fever (if you continue thinking Mozart your imagination may be running away with you), and a chance encounter with a passing Frenchman who happened to be connected with Perronneau’s great rival.

So how much of this picture is right? By the time it was retold by Vaillat & Ratouis de Limay, Mondonville had disappeared, but the two register entries were presented in transcriptions of the original Dutch (perhaps from Tourneux’s notes). There was a ludicrous attempt to reconcile the “42 Jaren” with the age of 68 – or actually 67, since that is what is meant by the words “dans la 68e année de son âge” that appear in the Académie register soon after in the text, rather than the erroneous heading “âgé de 68 ans” that was added later (an error V&RdL perpetuated by opting for 1715 rather than the eight times more likely 1716 in the title of their book): the solution proposed was that the body that was buried was not that of the pastellist, but that of his young brother Jean-Baptiste-Henry, the subject of a (to V&RdL lost) painting exhibited by Perronneau in 1746 when the boy would have had to be 5 to be 42 in 1783.

Needless to say d’Arnoult debunks this silliness, correctly (I think) identifying the 1746 painting as the one in the Hermitage showing a boy who was evidently more than 5, and inferring J.-B.-H.’s age from the “âgé de 25 ans” when he died on 7 April 1755. That would put his date of birth to between 8 April 1729 and 7 April 1730 (rather than the “1730 ou 1731” that appears in d’Arnoult’s pp. 208 and 365 or the 1731 on p. 367), but in fact the 1755 parish record judiciously adds the words “ou environ” after the age. As it happens I have found a conformed copy of J.-B.-H.’s baptismal record: he was born 19 June 1730, and so died at the age of 24 and five-sixths. And it will surprise no one to learn that his parrain was his older brother, the artist.

Of course there is still an untidiness in that the Hermitage boy looks rather younger than 16, but that probably just means the painting had been done earlier. D’Arnoult suggests that an estimated age of “une douzaine d’années” and her 1730/31 assumed birth put the picture in period 1744–46: but I leave you to decide which of chronology, biology, arithmetic or language is being stretched. (The point is not merely pedantic: a 1742 date would transform our understanding of Perronneau’s early career, while an age of 15 would add to the issues of the ages of his sitters.)

D’Arnoult doesn’t mention another intriguing theory that was published some time ago, and which caught my fancy for a time. In an article[1] about a Perronneau pastel (you can find it in the Dictionary at J.582.1231) then thought to depict Belle de Zuylen (d’Arnoult is good at rooting out such fantasies, but less convincing in proposing an alternative identity), Paul van den Boogaard came up with an ingenious explanation of the 42 J beside the diagnosed “fever”, suggesting that it should be read as 42 g, for 42 degrees Celsius – the temperature that might have been recorded for a fever. My initial reaction was to wonder whether they had the technology, but Holland was quite advanced in the use of clinical thermometers for measuring fevers, Dr Boerhaave (the subject of a portrait by Troost, right, in the eponymous museum in Leiden) having pioneered the practice. D’Arnoult however was unable to find the register with the entry, and (as we shall see) Boogaard plainly hadn’t seen it either.

D’Arnoult also drops the Mondonville story, but provides the documents which were Tourneux’s only basis for it. Although there was a posthumous inventory of Perronneau’s estate shortly after his death (10 janvier 1784, his date of death being cited as 20 novembre 1783), seven years later another document was prepared liquidating the estate. (The 1791 document still gave Perronneau’s date of death as 20 novembre, suggesting that his widow also had never seen the death certificate, may not have known the cause of death, and perhaps had no more information from Mondonville than his bill.) In the 1794 document, the effects Perronneau had with him in Amsterdam at his death were included, having been omitted from the inventaire. In this liquidation (which we know Tourneux had seen), we find the paragraph stating that

Perronneau avant de mourir avait chargé verballement de ses dernières intentions un S. Mondonville

and going on to list the expenses this gentleman incurred in settling the Amsterdam funeral expenses which d’Arnoult discusses in some detail – without however telling us who Mondonville was.

It would of course be fascinating to find that Perronneau’s last contact was with the son of the subjects of two of La Tour’s most important pastels (left is the Art Institute of Chicago’s version of the mother, née Anne-Jeanne Boucon: J.46.1423): just imagine the fictional possibilities for such a final conversation, whether down the pub or not. All the more so when you know that Mondonville fils, as Tourneux has him, was not merely an amateur draughtsman, but a pastellist with an entry in the Dictionary. And that his mother, also known as an amateur artist (as well as the titular La Boucon in one of the most gorgeous pieces ever written by Rameau – who you will recall from my last post was with La Tour when Mme de Graffigny met them in 1748):

was too a pastellist, as we find from her father’s estate inventory. And the link with pastel goes back to that father, Étienne Boucon, who was not merely a patron of the arts, but close enough to Crozat that Rosalba mentions him in her diary and even had lunch with him. Indeed just before Perronneau’s death Mondonville fils, by then in possession of his parents’ pastels, exhibited them at the Salon de la Correspondance, as recorded in the Nouvelles de la république des lettres et des arts for 19 juin 1782:

But there is no evidence that Mondonville fils was in Amsterdam, and I think it highly probable that the S. Mondonville mentioned in the Perronneau document was actually his cousin, Martin Cassanéa de Mondonville, whose presence in Amsterdam is attested by the baptism of three children at the French church (Catholic) between 1779 and 1783:

I can find little further information on him (there is a genealogy here), but his mother and sister had recently returned[2] from Moscow by mid-1783 (and may have been in Russia when Perronneau was). He appears from the few records we have to have associated with French Catholic expatriates and visitors, including Marc’Antonio Missoli, an important dancer and choreographer. And of course he was also the grandson of Étienne Boucon.

But what about the cause of Perronneau’s death? Reporting de Roever’s reading of the register as “koorts” (fever), as entered by a certain Jan or Jean Martens (of whom Tourneux, V&RdL and d’Arnoult tell us nothing), d’Arnoult nevertheless introduces the possibility of the eruption of the Laki volcano in Iceland which was widely thought to account for many deaths around Europe in 1783/84, citing the paper of Thordarson & Self[3], which does indeed provide a hugely detailed assessment of the environmental effects of the eruption. D’Arnoult suggests that the large numbers of deaths it caused might account for the chaotic circumstances of the burial.

The Laki event of course has been the subject of many investigations of more or less rigour, from the initial contemporary comments of Lalande (Ducreux’s pastel, Versailles, right), who tried to dismiss the dry fogs that appeared in June 1783 as the result of hot weather and rain, to the multiplicity of studies attempting to apply modern science to remote events where the data are not what one would gather now. Lots of famous people died that year, and one document even suggests that Laki caused Leonard Euler’s death (as it happens, from a brain haemorrhage).

So it’s time to have our own look at the evidence and see if anything can be added. Here is the 20 November cemetery burial record, which is easy enough to locate in the Stadsarchief Amsterdam:

And here is the entry from the Amsterdam death tax register known as the Gaarders Archief, which was indeed a lot harder to track down:

But to answer some of these questions, you need not only the line concerning Perronneau, but the adjacent entries. Here is the whole page covering three days in November:

From this you can see immediately that Boogaard’s ingenious idea is simply wrong. Everyone in the register gets an age. It’s just a mistake.

What about the mysterious Jan Martens? Was he a reliable witness? Well, he should have been: Jan Martens (c.1737–1808) was a heelmeester (surgeon) in Amsterdam, in practice from at least 1761, as his name appears in the list of members of the Amsterdam committee for medical supervision:

By 1784 one source even refers to him as the “beroemde” (famous) Stads Chirurgijn. (Martens’s address, by a curious coincidence, is the same as that – “op de Prinsse Gragt, bij de Leijdse KruijsStraat” in 1735 – of one of Liotard’s brothers, Daniel-Louis, whose birth in Geneva was known to Roethlisberger & Loche, but not the fact that he settled in Amsterdam where he was a monteur de boîtes; I have been unable to establish if he was still there when his brother came to Amsterdam in 1755.)

As for the Gaarders Archief entry, the Dutch system at that time did indeed levy taxes on death registrations. They were based on assessments of income and capital, divided into five classes. The wealthiest, those with capital in excess of 12,000 guilders, paid 30 guilders; the lower rates were 15, 6, 3 and for the poorest, nothing. Based on the total numbers in the Gaarders Archief for the second half of 1783, only 0.3% of deaths fell into the first class, while 84% fell into the fifth, or “pro deo”, category. In view of the link to assessed taxable income and capital, it is unclear to me how the tax would have been imposed on visitors, particularly those without close family members in a position to make declarations of assets or income (perhaps some kind Dutch archivist familiar with these records could assist, but my attempts to find examples were unfruitful before my attention span expired).

The second baffling thing in these entries is the phrase “Is Gehaalt” which appears in the cemetery register. D’Arnoult merely translates literally “a été enlevé”, while V&RdL expand as “on a cherché le corps, sans frais funéraires l’enterrement a eu lieu.” On its own the words are wildly ambiguous; perhaps again a kind Dutch archivist can assist. But a perusal of the entire cemetery register shows that the phrase is used very rarely. The nearest example I could find was two months before. Since 84% of burials were exempt from tax it hardly seems likely that V&RdL’s construction of the words was correct. Could it mean that the body wasn’t buried at all, but reclaimed – perhaps by Martin de Mondonville, to be buried elsewhere? If so I have found no trace of it in any Amsterdam church or cemetery, under any misspelling of the artist’s name; and this seems very unlikely.

Another question concerns the use of a cemetery rather than a church. D’Arnoult wonders whether his being Catholic in a Protestant city restricted his choice, but in fact there were plenty of Catholic churches available; there was even a French Catholic chapel (as opposed to the Walloon church for French Huguenots) where Mondonville’s children had been baptised. A rather different suggestion occurs in the biographies of the famous bookseller and publisher of Rousseau, Marc-Michel Rey, who was also buried in the Leydse Kerkhof, five days after his death (incidentally Gaarders Archief and history agree that this was due to a chest complaint, “borstziekte”, long before Laki had erupted), on 13 June 1780, “met vier Koetzen” – with four coaches, evidently an ostentatious display. One source[4] suggests that his choice of cemetery rather than church meant that he was suspected of atheism – something of which no one accuses Perronneau.

We should also consider the basic geography. Perronneau – as d’Arnoult notes – was staying in a wealthy area of town, on the Herengracht. You can see his lodgings marked in green in the contemporary map of the city (confusingly north is about 7 o’clock) at the top of this post; they stood on the spot occupied by the modern corner house (no. 396) as seen in this photograph of the Herengracht, looking roughly northwards (in a house very similar to the old one, bearing the date 1665, still standing two doors away, as can be seen in the drawing by Caspar Philips, left). The map also shows the residence of Dr Martens (in blue) and the location of the cemetery (in red). As the documents attest, Perronneau was ill enough to confide in Mondonville, so he probably was able to consult Martens, as a local doctor, before his death. And from his lodgings to the cemetery was but a short distance, by canal.

So what about the cause of death? Again the vital document is the Gaarders Archief: but not just the entry for Perronneau alone (where I confess the handwriting is unclear, and “koorts” looks more like a number of other possible words), but all the adjacent entries. I have tried to reconcile these with burial records in the various churches, although I confess I ran out of patience fairly soon. But one salient fact was that while Perronneau was buried the day after he died, in all the other cases I matched up the delay was between 2 and 5 days.

Armed with the Gaarders Archief figures, I found that there was indeed a significant increase in deaths in Amsterdam immediately after the eruption. Taking just the pro deo monthly totals, July 1783 showed a 66% increase on the average of the previous ten months, so this looked very plausible.

I turned not only to the paper about the Laki eruption d’Arnoult cites (Thordarson & Self 2003), but also some other studies – notably one specifically addressing mortality attributable to it.[5] This sets out to apply the work of Thordarson & Self to England, where detailed county-based mortality figures allow a sophisticated statistical analysis of the epidemiology. No one doubts that the effect of the disaster on Iceland was devastating, with up to 20% of the population dying (but more from hunger than the direct effects of inhalation), but were there more deaths in total in more populous countries like England?

The paper contains some useful cautions: “it is important not to confuse coincidence with cause”, although the evidence for the Laki eruption causing the dry fogs throughout northern Europe in June–July 1783 was strong. However “adverse health effects from exposure to volcanic gases are generally acute in nature, so any impacts would be expected to occur and be noticed during the period of contact with the gas.” In other words Perronneau would have suffered more when he was in Bordeaux that summer and during the journey to Amsterdam (although of course, following his trips to Poland etc. he could have been expected to be weakened generally). But Witham & Oppenheimer note that many other causes – epidemics, the extreme weather conditions of a very hot summer in 1783 followed by a very cold winter – are as likely to explain any increases in mortality figures in countries like England. Specifically they note that the lack of uniformity in different counties’ figures suggest local, rather than global, explanations are required.

One of the difficulties Witham & Oppenheimer faced was that the English figures gave no individual cause of death. The Amsterdam tax registers however do – most (but not all) of the entries contain some cause, however accurate the diagnoses may be. Perronneau’s “koorts” (for that is I now think how the word must be read) was in fact fairly constant as a proportion: about 11-14% in each of the months I counted. I found few instances of “borstziekte” or anything that looked closely related to inhaling sulphurous particles. But what I did find, rather to my surprise, was a sudden epidemic of measles (“Masel” for mazelen) coinciding with (rather than I think being caused by) the eruption: in July 1783 34% of the deaths were from this cause. But by November, the numbers were back to normal (513 pro deo deaths recorded in the full month: Amsterdam had had to cope with more than a thousand in a single week during some seventeenth century plague epidemics), so there was no likelihood of disorder at the cemeteries.

Amsterdam as a busy seaport was of course vulnerable to infectious diseases, and it seems to me that a more plausible explanation of Perronneau’s death and burial was that he had a contagious fever which caused sufficient concern, perhaps even alarm, for the state surgeon to be called in, and for him to order the immediate burial in the nearest possible location. In the absence of any family to certify income and pay the tax, a pro deo burial was inevitable. That notwithstanding the fact that Perronneau had with him in Amsterdam possessions (including 20 pictures) worth over 4000 livres, which alone would have taken him into the chargeable bracket if regarded as taxable capital.

But there is another aspect of all this which strikes me as more interesting and in a way sad. It is the statement in the 1791 liquidation that

Il faut observer que ledit Inventaire [the posthumous inventory taken in 1784] ne contient aucun effet à l’usage personnel de feu S. Peronneau. Ces effets avoient eté par lui emportés à Amsterdam….

All his personal possessions removed? Taken together with the fact that his widow remarried – or tried to remarry, as I discussed before – less than two months later, with Gertrudian alacrity, one can only wonder whether there is not a very simple explanation for the profound melancholy expressed in so much of his work.

Aschenbach? Not really; more Rameau than Mahler. And certainly not Agatha Christie, if that is what you wanted.

Notes

[1] Paul van den Boogaard, “An unknown portrait of Isabelle de Charrière (1773)”, Cahiers Isabelle de Charrière, 6, 2011, pp. 62-66.

[2] This is from Roberte Marchand’s biography of his uncle.

[3] “Atmospheric and environmental effects of the 1783–1784 Laki eruption: A review and reassessment”, Journal of geophysical research, cviii, 2003. This cites a report in the Odense Adresse-Contoirs Efterretninger, no. 36, 18 July 1783 confirming that the sulphuric gas cloud reached Amsterdam on 11 July, while another report commented on the sulphuric odour noticeable in late June. (Arnoult 2014, p. 183 reports this date as 3 November with no other source.)

[4] P.-J. Kapteyn: see Karl Rudolf Gallas, “Autour de Marc-Michel Rey et de Rousseau”, Annales de la Société Jean-Jacques Rousseau, 1922, p. 76.

[5] C. S. Witham & C. Oppenheimer, “Mortality in England during the 1783-4 Laki Craters eruption”, Bull. Volcanol., 2005, 67:15–26.

Advertisement