Research on one topic (Vermeer exhibitions) put Schwartz on the track of another (historical Rembrandt numbers). This year is the 100th anniversary of the publication, in 1923, of the most extreme highs and lows known to man for the count of paintings by Rembrandt. (Click on images to enlarge them.)
This year a century ago saw the publication of both outer limits of the number of paintings attributed to Rembrandt, by Wilhelm R. Valentiner (1880-1958) and John C. van Dyke (1856-1932).
The top figure was racked up by Valentiner. He took a running start that began in 1904. In that year the publisher Deutsche Verlags-Anstalt launched a series of monographs on great artists that became an art-publishing staple: Klassiker der Kunst (Art classics). The first two volumes were dedicated to the paintings of Raphael and Rembrandt, both written by the versatile classicist-art historian Adolf Rosenberg (1850-1906). In 1905 Rosenberg added a volume on Rubens, the fifth in the series. All were to go into multiple editions. By the time it became apparent that a new edition of the Rembrandt book was required, Rosenberg was nearing death. The second edition, dated 1906, reprints Rosenberg’s brief text on Rembrandt, but was otherwise entirely revamped, expanding the oeuvre from the 397 paintings in the first edition to an astounding 556. The volume has no author’s name, but it will have been composed by the young Wilhelm Valentiner, who in 1904 earned his Ph.D at Heidelberg with a dissertation on Rembrandt and his milieu. Valentiner did sign the third edition, of 1908, which jacked the corpus up to 597 paintings.
In that year Valentiner, on the recommendation of Wilhelm von Bode, the founding director of the Kaiser-Friedrich Museum in Berlin, was appointed to the department of decorative arts in the Metropolitan Museum of Art. When the Great War broke out, he returned to fight for Germany. In 1921 he returned to the States, to resume what became a sterling career in the American museum world. Rembrandt remained a central concern, and in 1921 he published, as the twenty-seventh volume in Klassiker der Kunst, a book of 102 Rembrandt paintings “rediscovered” between 1910 and 1920. The Rembrandts kept pouring in, and two years later, in the second edition of volume XXVII, with a subtitle extended to 1910-1922, he put the crown on his cataloguing with 115 paintings. Added to the 597 from 1909, this brings the tally up to 712. (On the title pages of the Klassiker der Kunst volumes, the publisher gives the number not of paintings but of illustrations, which also include details and comparative material; the numbers given for editions I-III are my own count. The 1921 and 1923 editions include numbered checklists.)
True, in some entries Valentiner expresses doubt. For example,
of these small panels, two of the twelve new Rembrandt attributions published in 1922 by Cornelis Hofstede de Groot, he says he “is not completely convinced.” Unexpectedly, three paintings in the first edition of Wiedergefundene Gemälde do not return in the second edition. For the record, since I do not think this has been observed before:
Below, out of sympathy for Valentiner, and to illustrate the dilemmas in trying to place potential Rembrandts, are the related paintings that were maintained in the second edition, Valentiner having found them just that much better. Performing subtle determinations of connoisseurship at this level is not easy.
Valentiner’s expansionist attitude fit into his career track as a paid advisor to American collectors and museums (he was on his way to becoming director of the Detroit Institute of Arts) in search of Rembrandts. This occasionally led to charges of conflict of interest, especially when as a museum curator he participated as a principal in a sale. On one such occasion, when he was on staff of the Met, he defended himself by saying that as a functionary in the decorative arts department he was free to trade in paintings.
The increase of 80 percent in the number of Klassiker der Kunst Rembrandts between 1904 and 1923 did not go unnoticed. There were quibbles, there was outrage. Most involved disagreements about individual paintings or certain chunks of Valentiner’s oeuvre, without disputing the basic assumptions behind it. One critic reacted differently. Six hundred miles east of Detroit, the head of the art history department of Rutgers College in New Brunswick, NJ, John C. Van Dyke, picked up the cudgels with an opening statement, signed in February 1923 and published that year by Charles Scribner’s Sons in New York, in which he put the very name Rembrandt between quotation marks:
So far back as 1883 my notes on the pictures in European galleries indicate that I was sceptical of the Rembrandt attributions and the Rembrandt tradition. [..] I could not understand the variety of ways in which Rembrandt handled a surface, why he found it necessary to use so many different styles and methods, why he was weak and strong, smooth and rough, cramped and flowing, all in the same year. I wondered if “Rembrandt” were not a cloak covering the work of many pupils. [… In 1911] I pointed out by name and number a great many erroneously attributed “Rembrandts,” and insisted that the whole Rembrandt oeuvre was a huge snowball that had gathered to itself the work of the school, and that every turn of the ball added new ingredients to the mass.
Following this conviction (which is eerily similar to the revelation Bob Haak experienced in 1956 and gave birth to the Rembrandt Research Project, which however treated problem paintings as “later imitations,” while Van Dyke correctly kept them in their proper time), Van Dyke set about bundling like works into distinct artistic personalities. Most could be found among the fifty-six pupils identified as such in the literature (actually, the designation “pupil” deserves quotation marks as well), but Van Dyke also assembled seven groups for which no name suggests itself, while giving other “Rembrandts” to fifteen artists who were not pupils. He starts the reconstruction of each grouping with a body of work generally assigned to that master. He would then comb the Rembrandt tradition, and would find paintings that belong more to his groupings than to each other.
To show how unafraid he was of where this system led, I illustrate one of his most daring corpuses, that of Gerrit Willemsz Horst. The four paintings in the left columns of the two pages he saw as solid Horst attributions. The four in the right columns he labels “Horst (given to Rembrandt).” They include two paintings of iconic Rembrandt status, the Danae in the Hermitage and Samson blinded in the Städel. When Van Dyke was finished, his “Pictures by the master” was left with only forty-eight paintings, on a list that however “does not pretend to completeness.”
712 to 48. Valentiner has fifteen times more Rembrandts than Van Dyke. As unlikely extreme as this ratio may be, and as unpopular today the two corpuses of both scholars, it remains a valuable reminder of the extent to which one’s idea of what a Rembrandt is depends on your chosen starting point, and of the weakness of the available categories and groupings. Today one can say that there is consensus concerning roughly three hundred paintings, with about another hundred in play. Some halfway up and/or down from the records set in 1923.
And the Vermeer ratio? In 1923 there was no competition for Cornelis Hofstede de Groot. In vol. 1 of his Catalogue raisonné of the works of the most eminent Dutch painters of the seventeenth century of 1907, augmented with two paintings he did not see until later, he recognized thirty-six paintings as Vermeers. Earlier this month the Rijksmuseum accompanied its Vermeer exhibition with a catalogue of all the Vermeers it accepts. They come to thirty-seven, including one painting that was not attached to Vermeer’s name until 1969. Brushing off some minor discrepancies, and leaving out the decimal fractions, the Vermeer ratio is 1:1.
What to say? If our numerous Rembrandt categories are too weak, could the single one for Vermeer be too strong?
© Gary Schwartz 2023. Published on the Schwartzlist on 25 February 2023.
The research for this column was conducted for a lecture on Vermeer exhibitions, from 1935 to 2023, that I gave on 20 February for the Royal Antiquarian Society (Koninklijk Oudheidkundig Genootschap; KOG), in the auditorium of the Rijksmuseum, which I was invited to deliver by the recently appointed president, Jan Teeuwisse. The lecture was filmed and is available on YouTube at https://www.youtube.com/watch?v=f1n5BV-iyBA. My request that the Dutch talk be subtitled in English, which apparently costs a lot of money, is under consideration. (Most of the column is an elaboration on part of the lecture that I had to abbreviate drastically.)
As for the possibility that our image of Vermeer’s oeuvre is overdetermined, the only serious dissident view that I know is that of Benjamin Binstock, in his article “The apprenticeship of Maria Vermeer,” Artibus et Historiae 29 (2018), pp. 9-47, and book Vermeer’s family secrets: genius, discovery, and the unknown apprentice, London (Routledge) 2009.
That this week the return to murderous savagery in Europe has persisted for a year, with no cessation in view, makes me more bitter than I can say, or even admit to myself. The Atlantic world has once more been plunged into a war of European making, for which the only chance of an ending that does not destroy Europe entirely, depends on the United States. An ending that, no matter when it comes, will leave hundreds of thousands of tragic deaths behind.
Responses in the Reply box below (these will be viewed by all visitors to the site) or personally to Gary.Schwartz@xs4all.nl are always appreciated and will be answered.
So will donations. Don’t be shy, send a donation.
Your donations help defray the costs of the Schwartzlist and encourage Gary Schwartz to write more columns.
Excellent.
Thanks, Steve. You and I seem to be all alone in the Comments section. I thought this subject would interest my colleagues.
One wonders: how many items, in these so different canons, have firmly documented provenance? What fraction involves connoiseurship?
Imagine if all such were labeled as “Consensual.”
Richard, if provenance is taken into consideration at all, it is only to belittle its importance. The Rembrandt Research Project was disdainful of all those royal and aristocratic pedigrees. All the information that counts, they said, is in the object itself. Period. Not that those hundreds of “rediscovered” Rembrandts had much going for them that way. They mostly emerged from obscurity, even the good ones, which there were.
Thank you in the name of sobriety and fairness for giving Van Dyke some well-deserved publicity: even though he went overboard in the opposite direction, his heart and mind were in the right place, and the man certainly had panache. The earlier catalogues, however much overpadded, had the virtue of brevity, a quality that has since been offset by the sheer number of catalogues devoted to Rembrandt’s oeuvre: has any other artist benefitted from nearly as many? (What makes the Rembrandt meme so irresistible?). The very bulk of the RRP’s stranded “Corpus of Rembrandt Paintings” should discourage further attempts, shouldn’t it?
That’s certainly true of the Corpus, Jean-Marie, but the funny thing is that when you start researching any individual painting by Rembrandt (all right, not the Tulp or the Night watch), it’s astonishing how little there is.
That is funny indeed, but not amusing. I can assure you, however, that the literature on Br. 431 is going to grow further (it is already featured at length on Wikipedia). One of my life goals is to go to Stockholm to check out Br. 430, which I believe is closely related and indicative of a special moment in 1631 (perpetuated by Dou and others).
That’s thanks to you alone, Jean-Marie, indebting us to you forever. Another, similar instance of an unjustly de-attributed Rembrandt is the oil sketch for the etched portrait of Lieven Coppenol in the Metropolitan Museum of Art. In 1996 a painting police detective, Louis Velasquez, wrote a tractate objecting to the treatment of the painting in the Met exhibition Rembrandt/Not Rembrandt. Although no one else had stood up for the attribution since Gerson questioned it in 1969, Velasquez won out. Walter Liedtke backtracked, and in 2000 Ernst van de Wetering wrote a lot about it in a catalogue on Rembrandt as a printmaker. In volume VI of the Corpus (2015) van de Wetering gives it a lot of space, only quoting himself and not mentioning Velasquez at all.
So all’s fair then.
Speaking of panache, one day Egbert Haverkamp-Begemann took his class to the Met and the Frick, delivering his views on authenticity – and the pressures brought to bear in deciding it – in front of their Rembrandts. It was like sailing with Tromp; the opposition suffered heavy losses.
Nice to imagine that session. Thanks, Richard.
Thank you, Gary, for kindly mentioning my Vermeer book. This sought to articulate closer *relations between Vermeer’s paintings*, specifically his gradual *painting-by-painting development*. That account was only possible after excluding seven “misfit” paintings, which conversely could fit as works by his apprentice, who based her compositions on elements from (combinations of) his paintings. The same approach potentially applies to distinguishing between Rembrandt’s paintings and those of his students.
As I’ve admitted, Ben, I have not yet allowed myself to delve into your work. But I was pleased to have the occasion here to signal your exceptional position in Vermeer studies.