A question of attribution

The forthcoming sale of a purported Caravaggio is set to exceed £100m, but its provenance is far from clear

Art
Doubts and gaps: The Toulouse Caravaggio, “Judith Beheading Holofernes”, to be auctioned this month (©Hubert Fanthomme/Paris Match via Getty Images)

What price a line in a letter? In the case of a soon-to-be-auctioned painting supposedly by Caravaggio the answer is likely to be millions. The picture, showing Judith beheading Holofernes, was found in a Toulouse attic in 2014 during a routine auctioneer’s visit. Despite some water damage, the painting was in good condition and was most certainly Caravaggesque, replicating a bona fide version of the same subject c.1598-1599 in the Palazzo Barberini, Rome.

The French government placed a 30-month export ban on the picture on the grounds that it was “a very important Caravaggio marker, whose history and attribution are to be fully investigated”. Neither the Louvre nor any French museum took up the offer to buy the work for €100 million and last December the export ban was lifted. The picture, cleaned, restored, rechristened the Toulouse Caravaggio and dated to 1607, will now go on sale at the Toulouse auction house of the finder, Marc Lebarbe, on June 27. There is no reserve, but bidding will start at some €30 million, and it is estimated to fetch between €100-150 million. Should the picture hit that level, it will become the second most expensive Old Master painting ever sold, after the $450 million Salvator Mundi ascribed to Leonardo. The Toulouse Caravaggio is currently on a world tour, part of a marketing drive to drum up interest, taking in London, Paris and New York.

It is a painting that has split opinion. Keith Christiansen, of the Met in New York, and Nicola Spinosa, former Director of the Capodimonte Museum in Naples, are among the scholars who claim it is a genuine work. But many other experts dispute this (including two leading British specialists I spoke to and Mina Gregori, the grande dame of Caravaggio studies). When the Louvre examined the painting, it did not rule either way on attribution (though its failure to buy has been taken as some sort of judgment) and a gathering of Caravaggio scholars at a study day for the picture at the Pinacoteca di Brera in Milan in 2016 failed to reach a consensus regarding the hand of the artist.

Attributions of old paintings are a curiously imprecise business. Technical examination has shown several factors that place the image at least very close to Caravaggio: it is painted on the same type of canvas used for other paintings he made in Naples while on the run from a murder charge in Rome; the paint contains substantial amounts of calcium carbonate, again characteristic of his work at the time, as were the incised lines along the limbs, again present. Importantly, the picture has pentimenti, signs of modifications overpainted by the artist, that are traditionally held to show the painter changing his mind and that therefore can’t be the work of a copyist (though even the most proficient copyist might need to make adjustments).

The clinching evidence for the Caravaggio believers, however, is documentary. On September 25, 1607, the Flemish painter Frans Pourbus wrote to the Duke of Mantua from Naples describing two “bellissimi” works by Caravaggio for sale, one of which was a Judith and Holofernes. The pictures may have been for sale in the workshop of the artists and dealers Louis Finson from Bruges and Abraham Vinck from Antwerp, who knew Caravaggio. The Duke didn’t buy the pictures because they are mentioned in Finson’s will dated 10 years later in Amsterdam in which he bequeathed his half share of them to Vinck. The first of the paintings, the Madonna of the Rosary, was bought by a consortium of artists including Rubens and Brueghel and is now in Vienna. The fate of the Judith and Holofernes is, from this point on, unknown. To complicate matters further, there is another version of the Toulouse painting made by Finson and now in Naples.

The sellers have commissioned a handsome website devoted to their picture ahead of the auction. On it they state: “The authenticity of the rediscovered work . . . is well documented.” But this is not the case. There is plenty of suggestion, but simply no indisputable evidence to prove that their picture is the one seen by Pourbus in 1607. That line in Pourbus’s letter is the key to the painting’s success at auction.

The same suggestibility was at play too during the sale of the Salvator Mundi. A vital part of the pre-sale promotion stressed that it was a painting with a royal heritage. When, after Charles I’s execution in 1649, an inventory of his goods was drawn up for sale, one item, number 49, was a “Peece of Christ done by Leonard”. From this imprecise tag a link was made to the Salvator Mundi, despite it not carrying the branded mark of the Royal Collection (a “CR”—Charles Rex—surmounted by a crown) on the back. There is, though, a painting of Christ as the saviour of the world by one of Leonardo’s followers,  Giampietrino, now in the Pushkin in Moscow, that does bear Charles’s mark. Nevertheless, when Christie’s sold the Salvator Mundi in 2017 the catalogue essay dealing with its provenance was entitled “A peece of Christ done by Leonardo” and a possible link had become a near-certainty.

Of course, the documentary links in both cases, although far from copper-bottomed, could be correct: the paintings have enough about them to suggest they might be by Leonardo and Caravaggio. Indeed, a major art historical journal apparently has a piece ready to run after the Toulouse sale explaining why its writer believes the Caravaggio to be an autograph work. There are, however, more than enough stylistic doubts about the picture and gaps in its provenance that those suggestive scrawled lines can’t bridge for the sale to be a test of the faith—as well as the purses—of would-be purchasers.