303 The transparent connoisseur 2: More Rembrandt core

Although the discussion is still somewhat one-sided, Schwartz continues his attempt to correct certain misapprehensions on the part of his colleagues concerning the nature and extent of Rembrandt’s work as a draftsman. Here he compares the promise of the recent Getty exhibition with the compromise it delivers.

Schwartzlist columns 301 and 302 were written in preparation for a symposium on the drawings of Rembrandt in the Getty Museum in Los Angeles, held in conjunction with an exhibition entitled Drawings by Rembrandt and his pupils: telling the difference. The articles challenged two major premises upheld by the makers of the exhibition: that there are only 75 drawings whose attribution to Rembrandt is certain enough to allow them to serve as touchstones for further attributions; and that Rembrandt’s lifetime production of drawings was smaller than the 1,575 sheets catalogued by Otto Benesch. I argued that the so-called core group of drawings numbers about 170, and lifetime production at least 9,000 sheets.

The organizer of the symposium, Lee Hendrix, and the speakers at the symposium were kind enough to give me five minutes to present my ideas from the rostrum, to all the 200 or more attendants of the symposium. I cannot say that I was altogether content with the response. Peter Schatborn, the foremost expert in the field and one of the main authors of the exhibition, defended the short core list and did not respond to the more important question of Rembrandt’s way with drawings. In the corridors and by e-mail I was patted on the back by some colleagues, but no one as yet has put into print a reaction to my challenge.

What follows are post-symposium reflections on the situation. The issues might seem highly technical and of interest only to specialists, but I think they do matter. In the first place as a corrective to our understanding of Rembrandt, but also for the critical insights they offer into the techniques and practices of scholarship.

Transparency does not come naturally to the art connoisseur. Think only of how a problem lands on his or her desk. No one bothers a connoisseur, nor do they bother themselves, for an opinion about an unproblematic attribution. Where transparency reigns – say Géricault’s Raft of the Medusa, painted in public by the artist – the opinion of a connoisseur that the painting is indeed by the master is not worth much.

What makes the opinion of the connoisseur worth more is the absence of transparency. Say that a hitherto unknown drawing that corresponds with a detail from the Raft of the Medusa turns up in a sale in Australia. The question that makes all the difference in the value of the sheet is this: was the drawing made by Géricault in preparation for the painting or was it drawn after the painting by a later follower? For an answer to that question one would go the leading expert in Géricault’s drawings. Until last year, when he died at the age of 89, that was the formidable Lorenz Eitner. Typically, the auction house would send a photograph of the drawing to the expert, who would give or deny his blessing to the attribution. The reasons for accepting or denying Géricault’s authorship, which may or may not be stated explicitly, are less important than the judgment itself. If Eitner was willing to say that the sheet is by Géricault, it would fetch more on the block – much, much more – than if he isn’t.

There are good reasons for connoisseurs to avoid being all too explicit in specifying their reasons for an attribution. Every positive statement invites contradiction from colleagues who see things differently. To define the features that constitute authenticity in the work of a particular artist reduces the market value of the connoisseur’s expertise and diminishes the mystique that is one of the most bankable assets of his stock in trade.

Nonetheless, there are intrepid and selfless connoisseurs who from time to time attempt to systematize their knowledge and techniques. One of those who did so was the German art historian Wolfgang Schöne, in his book of 1938 on Dieric Bouts. This put him in my estimation more than a notch above the most famous connoisseur of the twentieth century, Bernard Berenson, Mr. Mystique himself, who considered his mere opinion worth more than any scholarly argument. To my knowledge, there have never been comparative or even individual studies of the sustainability of attributions over time. My feeling is that Schöne’s have worn better than Berenson’s.

A move in the direction of transparency in the attribution of Rembrandt drawings has now been made, to their great credit, by Bill Robinson and Peter Schatborn. They do this in the catalogue to the exhibition Drawings by Rembrandt and his pupils: telling the difference, which was on display at the Getty Museum from December 2009 to the end of February 2010. This is from their statement on method, on p. 41 of the excellently designed, edited and printed catalogue. It is a restatement of a splendid method first broached in 1894 by the German art historian Woldemar von Seidlitz that since then has been honored by connoisseurs far more in the breach than in the observance.

The method for considering authenticity is as follows. First, assemble all of the relevant material, including drawings attributed to Rembrandt’s pupils as well as to Rembrandt himself. Next, identify a core group for Rembrandt and for each of the pupils that consists of signed drawings and studies for autograph paintings and etchings. (Sheets with inscriptions by Rembrandt also belong to this group, although these are less secure in certain cases.) Then, using explicit arguments and a rigorous process of comparison, introduce to the oeuvre drawings whose style and technique are consistent with those in the core group…. The attributions in Drawings by Rembrandt and His Pupils: Telling the Difference are the result of this carefully applied process.

This is beautiful, and one cannot fault the authors all that much if it is too good to be true. But it is. A first and methodologically fatal shortcoming in transparency is that the authors do not publish their core lists for Rembrandt and the fourteen pupils or associates included in the exhibition. That means that we must either take it on faith that their attributions were formulated as advertised, or else check the procedure ourselves. This is not a minor issue, since there is a huge discrepancy between the core list as maintained by the authors, counting some 75 sheets, and the far larger one of some 170 drawings that I compiled on the basis of Seidlitz’s criteria. See Schwartzlist 301, where I tell how I came into possession of the unpublished core list recognized by Martin Royalton-Kisch and Peter Schatborn and compare it to my own.

In Drawings by Rembrandt and his pupils: telling the difference, 46 drawings by the master are catalogued. Seized with irrepressible curiosity as I was to find out whether Seidlitz had finally carried the day, I went through the catalogue, noting the arguments advanced for each attribution. The first thing I noticed is that only eight of the drawings in the show are on the short core list, while another six are on the long list. From this one can conclude that the long core list is in fact already in operation, though unacknowledged. The major methodological question was: how did the authors justify their attributions of the sheets not on their core list? Did they really follow a strict procedure of comparison with core works or didn’t they? As an appendix to this column I attach a table showing all the references in the catalogue entries to comparative material relevant to the attribution of Rembrandt drawings. Here is the picture that emerges from the table.

In attributing to Rembrandt the 38 drawings in the show not on their core list, the authors follow their own emphatically formulated procedure only in fourteen cases. Nine more drawings are attributed to Rembrandt with no comparison at all. In five non-core-group cases, the comparison drawn by the authors is not to an established drawing by Rembrandt, but a painting or etching of his or a pupil.

In ten cases where another Rembrandt drawing is cited to back up an attribution, the drawing concerned is not on the core list and does not meet the authors’ definition of a solid basis for a Rembrandt attribution. This group and the nine drawings given to Rembrandt with no further argument detract from the authors’ claim that were following a “carefully applied process.” If they were, they did not share the evidence for it with their readers.

What seems to be happening is that alongside the rigorous comparison with the core group, implemented in fourteen attributions, quite another process, a less systematic one, is also taking place, on a larger scale. That is business-as-usual connoisseurship of the classic is-she-or-isn’t-she variety that the authors said in their statement they were leaving behind for something better. Although they fall short of their admirable aim, the authors nonetheless do contribute toward transparency in another, intermediate way. That is, they put into words their reasons for giving a drawing to Rembrandt or an artist in his circle.

That brings me to a final point that also deserves attention in a critique of connoisseurship. Consider the following comparisons between Rembrandt and his pupils:

“Lievens’s bravura linearity,… while displaying compelling mastery of draftsmanship, does not convey the nuances of the sitter’s ruminative psychological state to the degree encountered in [two of Rembrandt’s drawings]” (p. 51).

“Characteristic of Flinck are the rather uniform, flat areas of hatching between the figures, while the hatching in Rembrandt’s drawing is sketchier and more tightly integrated into the composition” (p. 71).

“Compared to Rembrandt’s drawing, Bol’s brushwork is excessive, and his deployment of light and shade undisciplined” (p. 85).

“Rembrandt liberally drew with his pen over sections that had already been washed. In his pupil’s work [Gerbrand van den Eeckhout], the washes lie on top of the lines, and the shapes appear more two-dimensional” (p. 119).

“Yet the pen strokes in the pupil’s work [Jan Victors] are somewhat less assured and more uniform than those in Rembrandt’s drawing” (p. 133).

“This makes [Fabritius’s] David less expressive than Rembrandt’s figure” (p. 142).

“Confronted by Rembrandt’s bold, imaginative departure from the letter of the text, Van Hoogstraten characteristically reverted to it” (p. 153) and so forth, down to

“De Gelder assimilated Rembrandt’s technique, but only superficially” (p. 235).

The exhibition was structured to make judgments of this kind visible. The 46 drawings given to Rembrandt were matched by 45 drawings by associates and named and unnamed pupils. On the left in the display cases one saw a drawing given to Rembrandt and on the right a similar drawing given to another artist. Like the quotations above, the texts all pointed out deficiencies and only deficiencies in their work compared to Rembrandt’s. A number of qualifications come to mind; I spare you the complete sentences belonging to each. Tautological, circle reasoning, tunnel vision, tendentiousness, leading questions, foregone conclusions, hero worship. However criticism of them is formulated, these overdetermined comparisons lead you to suspect that in positing them the authors were setting themselves up. That is an implicit danger in connoisseurship that can be countered by the rigorous application of rules like those of Seidlitz, even when they lead to conclusions that go against the grain. The authors of Telling the difference have not gone that far.

This column is entitled “The transparent connoisseur 2.” The first item in the series appeared on 30 March 2002, when I promised that it would “be continued at irregular intervals, as occasion demands.” I cannot say that occasion has not demanded a new installment for the past eight years. If I’m running behind, this comes partly from trying at one and the same time to make the critical points I feel have to be made while avoiding querulousness. I will risk the latter danger by placing on the Schwartzlist website in the coming time installments 149 and 154.

Exhib. cat. nr. Benesch nr.
Comparisons to Rembrandt drawings
Other comparisons
1.1 7 None Painting Br. 423
1.2 (Rembrandt or Lievens) Not in Benesch Compared but mainly contrasted to 1.1
2.1 39 Ben. 41
2.2 38 None
2.3 20 None
3.1 137 None Etchings B. 28, 192
4.1 128 None
5.1 100 None Paintings of the Passion
6.1 120 Ben. 141 and other drawings of the model
7.1 423 None
8.1 99 Ben. 97, 100, 140, 180
9.1 757 Ben. 442
10.1 327 None
11.1 606 Ben. 500a, 538
12.1 538 Ben. 500a, 759
13.1 97 None
14.1 140
15.1 416 Ben. 294A, 418
16.1 321 None
17.1 1103 Ben. 1101-02, 1174
18.1 180 Ben. 97, 100, 423, 425
19.1 161
20.1 519 None
21.1 500a
22.1 516 None Etching B. 105
23.1 1039 Ben. 643
24.1 909 Ben. 913
25.1 1264 Ben. 1333, C41
26.1 575 Ben. 161r
27.1 684 Ben. 361 Painting Br. 361 (no longer given to Rembrandt)
27.2 310 Ben. 137 (contrasted), 457, 590, 769
28.1 876 None
29.1 1267 None Etching B. 234
30.1 887 Ben. 886 Copy in Braunschweig
31.1 902 Ben. 904, 905, 906
32.1 1368 Ben. 1045A
33.1 980 Ben. 913, 914
34.1 860 Ben. 902, 904
35.1 1282 Ben. 1227
36.1 920 None
37.1 1296 Ben. 1295, 1297 Etching B. 222
38.1 835 Ben. 1296
39.1 988 Painting Br. 416
40.1 1057
41.1 1122 Ben. 1142
43.1 1142 Ben. 1122, 1179, 1180

Appendix. Table of the Rembrandt drawings in exhib. cat. Drawings by Rembrandt and his pupils: telling the difference (Getty Museum 2009-10), with reference to the comparisons made in support of attributions. Numbers in red are on the short core list maintained by Peter Schatborn and Martin Royalton-Kisch, those in blue on my own long list, which includes all items in the short list.

© Gary Schwartz 2010. First published on the Schwartzlist on 16 April 2010; posted on TheArtsFuse on 17 April 2010.

Respond to Gary.Schwartz@xs4all.nl