Rembrandt takes a giant leap backwards

Picture: National Gallery of Art

This is a weak picture that crudely imitates Rembrandt's late style. But in the early twentieth century it was one of the celebrated Rembrandts in the collection of Rodolphe Kann, who was a sensitive connoisseur with a sensational collection. It was sold by Duveen to the Widener Collection, and was exhibited by the National Gallery in Washington as a Rembrandt until at least the 1960s. What did they see in it? It's an odd failure of connoisseurship. People whose sense of quality was normally acute lost their critical faculties when confronted with any dark, broadly painted Dutch portrait. 

More recent connoisseurs have pruned the canon, restoring Rembrandt's oeuvre to a smaller and more consistent core. Rembrandt's style evolved and he had talented students and followers imitating his work closely, so there is unlikely ever to be consensus. Works at the edges will continue to be debated, and views will likely shift to and fro. But the radical pruning was necessary and brave. It was necessary because our appreciation of Rembrandt's genius was blunted by the admission of so many bad pictures to his oeuvre. And it was brave because new attributions are much more popular than de-attributions. No one likes to be told that the picture they've loved as a Rembrandt isn't any such thing. And reputations are made with new discoveries, not by righting others' mistakes. There is a connoisseurial bias towards big name attribution that takes courage to resist. 

The process of redefining Rembrandt was of course contested. The work of the Rembrandt Research Project has been especially controversial, and I doubt anyone would now defend their initial approach, which downgraded some unquestionably excellent and authentic pictures. I'll return to the RRP when I review the imminently forthcoming final volume of their Rembrandt Corpus in a future post, but there were plenty of other connoisseurs re-defining Rembrandt at the same time. The famous Man with a Golden Helmet was downgraded to Rembrandt school by the museum that owned it, and plenty of other scholars were undermining traditional optimistic attributions. But the pendulum seems to be swinging again. The final version of the Rembrandt Corpus under the leadership of Ernst van der Wetering adds dozens of new Rembrandts (or rather gives them back to Rembrandt), many of which will continue to be contested. 

I'm excited by the debate (I'm geeky like that). I think the arguments have been fascinating and productive, even if the conclusions are necessarily open-ended. I love the Rembrandt Corpus, despite sharing the consensus view that many of its early conclusions were bonkers. And I think the slimmed-down version of Rembrandt's oeuvre is more plausible than the earlier bloated catalogues of variable pictures. Connoisseurship is an ongoing process of re-assessment, not a magical insight into who painted what. Our sense of quality changes, as well as our sense of authorship. Pictures now seen as bad were once valued as masterpieces. The process of re-defining Rembrandt has been productive and I think we now have a clearer view of the master. But it's not a final view, and I look forward to the debate over van der Wetering's re-attributions.

My views on this are almost exactly the opposite of those recently expressed by Bendor Grosvenor in the FT and on his blog. I disagree with him on specific attributions, but I disagree more fundamentally on his reasoning. He starts by turning history on its head. Where most think that Rembrandt scholarship went awry when early twentieth century connoisseurs started attributing every dark Dutch daub to Rembrandt, he thinks the revisionists were the ones at fault, needlessly questioning authentic pictures. But a glance at some of those old catalogues shows that too much was accepted too readily. The RRP undoubtedly went to far in its scepticism, but it was going in the right direction.  

The worst argument Bendor makes is that Van Dyck painted more pictures in a shorter life than are attributed to Rembrandt, implying that there must be more Rembrandts out there. We have no idea how much time Rembrandt spent goofing off, wandering the countryside with his sketchbooks, or hanky-panky with the maid. We don't know how much he procrastinated, or how much time he spent teaching, or even how many pictures have been lost. And above all we have to judge how much of the master must be present before a studio picture can be called autograph - how much do those same-y English Van Dyck portraits have to be by the man himself for them to be called Van Dyck? Two connoisseurs could agree about the division of labour but still disagree about the label ('Van Dyck' or 'Van Dyck and studio' or 'studio'?). This kind of reasoning is no basis at all for expanding an artist's corpus. And it is question begging; what about the tiny oeuvre of some of Rembrandt's pupils? It's as much an argument for expanding their oeuvre as it is for expanding Rembrandt's. Of course he could have painted more, but we can only establish what he painted by examining pictures, not by speculating about how productive he might have been. 

Picture: Metropolitan Museum of Art

A specific picture that Bendor thinks is by Rembrandt is the so-called Auctioneer in the Met (above). He is wrong to say that it was downgraded simply because of x-rays. It was downgraded because of visual weaknesses that are not solely the result of its poor condition and are quite evident without x-rays. The Met's own catalogue notes that the x-ray discoveries 'merely underscore what is obvious from an examination of the painting's surface, namely that the work is an impressive but superficial adaptation of Rembrandt's mature style" (Liedtke Dutch Paintings p. 762). The Art and Autoradiography catalogue that disputed its attribution to Rembrandt in 1982 itself went beyond analysis of the x-rays, noting features of brushstrokes and use of paletteknife that create technical effects rather than define form. Met curator Walter Liedtke elsewhere describes it as "facile in conception as well as technique":

"The young man's sleeve and cuff, for example, are painterly passages brushed out with little thought for texture, modelling, anatomy (where is the elbow?), or for the complexities of light, and this approach corresponds with that found in other areas: the flat effect of hair framing the face, with an opaque stroke below the hat , and the face abruptly divided into light and dark, all of which, if compared with the Aristotle, looks like a schematic reproduction (like a photocopy), a simplification of a Rembrandt idea." (Walter Liedtke 'Some Paintings not by Rembrandt in the Metropolitan Museum' in Görel Cavalli-Björkman (ed) Rembrandt and His Pupils Nationalmuseum Stockholm 1993, p. 124)

I agree. 

On Art History News Bendor rather cheekily criticises the NG and Met for not changing their attributions on the basis of Wetering's judgment, suggesting they believe that "they know more about Rembrandt than van der Wetering". Museums employ curators to make informed judgments about presentation, including attribution, based on scholarly assessments but not deferring mindlessly to the biggest celebrity expert on a given artist. It should not be suggested that their decision to take a different view is a slight on the reputation of van der Wetering. The National Gallery of Art in Washington even decided to go against some of the views expressed by the (very eminent) scholar who wrote the catalogue of their collection of fifteenth century Italian paintings. For what it's worth, I think the Met and NG are right and van der Wetering is wrong in the cases that Bendor cites. But my bigger objection is to the presumption of deference to authority. Bendor starts with a bias towards expanding the corpus, then implies that we should defer to an expert who is expanding the corpus. The reasoning is spurious, and the consequence militates against healthy debate. 

Of course the pictures are the main thing. Does it matter if Rembrandt painted the Man with a Golden Helmet? Yes, I think it does. We want to know something of the creative mind behind the picture, and we want to understand Rembrandt through the pictures he actually painted. But attribution isn't everything. It's a wonderful picture, whoever it's by. The point I'd like to argue most forcefully is that we should be more tolerant of uncertainty. We should of course be respectful of expertise; van der Wetering's view is worth more than mine. But as an amateur Rembrandt enthusiast I have my own view of the man, and I've made my own judgments about many of the controversial attributions. I'll happily argue the toss without expecting to be taken seriously. Thinking about Rembrandt attributions focuses my thoughts around quality and technique and shapes my view of the man. It's just not the case that half a century of idiots have lost sight of the true Rembrandt, and by deferring to van der Wetering and Bendor Grosvenor we'll get him back. It's an ongoing dialogue. Bendor taking the opposite view from me on certain attributions is all part of a welcome debate. But to my mind his comments about attributions at the NG and the Met give too much weight to the authority of a single expert, whereas we should rather welcome informed dissent. In my view it is a brave and wise decision to leave these pictures unattributed rather than display them under a famous name.

13/10: Bendor Grosvenor responds:

Not for the first time, you have misinterpreted what I have said. I don't know why. Perhaps its necessary in order to justify a 'Grumpy' approach. Which is a shame, as it often gets in the way of the more acute and well put observations you make. But in this case you were – if you’ll forgive the mixed metaphors – both tilting at windmills and shooting the messenger. That said, I’m glad you've amended your earlier version of this post.

As to the remaining points, unfortunately you infer far too much of what I might believe from an article in the Financial Times which, by necessity, is word limited, and which is directed at a more general audience. In fact, if you listen to the podcast of the same piece, which is not subjected to a sub-editor's scissors, you will note that I do not, as you say, "turn history on its head" when it comes to the number of genuine Rembrandts that might exist. I make it clear that there were many dud Rembrandts in the 600 or so catalogued by the likes of Bredius et al, and that much weeding out had to be done. My point that the revisionism had, by the early '90s, gone too far in the opposite direction is now a pretty widely held view. So actually you and I agree on this. 

You further misinterpret me by devoting a whole paragraph to my single line comparison, meant only to convey to the general reader an idea of how productive or not 17th Century artists were, on Van Dyck's output. Only a fool would think that one can establish attribution by any other method than, 'examining pictures', as you put it. Surely you know that I would not suggest any other method, not least because I have something of a track record of proving authorship by ‘examining pictures'.

On the ‘Auctioneer' and 'Man with a Golden Helmet', it is the case that X-ray evidence was a significant, and probably major factor in their downgrading. You cite criticisms of the picture made some time after the period I was discussing. As to the specifics, what you see as visual weaknesses in the ‘Auctioneer' strike other viewers, such as Ernst van de Wetering, and for what it’s worth myself, as lively and impressive passages. So there we must agree to disagree. As to other criticisms of the picture (the visual impact of which, despite what you might say, really is affected quite significantly by its condition), these seem to me to focus too much on searching for weaknesses in small areas, rather than an appreciation of the whole. 'Where is the elbow', Walter Liedtke asks? Evidently, under the baggy cloak.

Yes, on AHN, I do 'rather cheekily' criticise the Met and the National Gallery for not changing their attributions. It seems odd to me that van de Wetering's view is good enough for many museums and national institutions (just earlier this year for the National Trust, for example) but not for the Met or the NG. Not even to upgrade to ‘Attributed to’, or ‘Possibly By’, or ‘It has been suggested’. No, they remain firm on ‘Follower of’. In fact, for almost every other artist, museums do indeed defer to the view of whoever happens to be the pre-eminent scholar of the day, especially if they are writing a catalogue raisonné on the scale of the 'Corpus'. But not it seems for Rembrandt. 

And that is my broader point of course - that Rembrandt scholarship/connoisseurship has become so bafflingly disputed that nobody is really sure what a Rembrandt looks like any more. You say you object to this 'deference to authority' - but take your statement to its logical conclusion, and it means that you object to any sense of expertise at all. It is not 'spurious reasoning' to believe that a highly regarded individual who has spent literally decades studying the work of one artist, Rembrandt, more closely than anyone ever before, might know more about that artist than generalist curators, no matter how talented and scholarly those curators are (which I do not doubt) and to, on balance, seek to trust the opinion of that individual, in this case Ernst van de Wetering, more on what is and is not a Rembrandt. 

You say that it 'does matter' if Rembrandt painted a certain work, but go on to say that 'attribution isn't everything'. Perhaps this is where you and I disagree the most. I think knowing who painted what is the first essential building block of all art history. I also think it is possible to try and establish authorship in the majority of cases. So where you say we should be 'more tolerant of uncertainty', I say no, let's not, because in a way that represents giving up. I believe we should always try and find out who painted what, first because we can, and second because when the average gallery visitor goes to admire a painting, they like to look at the label and know the basics; subject matter, artist, date. Everything else follows from that basic information. I say that the first responsibility of a gallery is to try and provide visitors with these answers. 

Finally, let me tell you what really ‘militates against healthy debate’ in these matters. It is not, as you say, me writing a post on Art History News, or swinging a lamp on the history of Rembrandt connoisseurship for the benefit of Financial Times readers, or for that matter introducing many of my readers to your website. Rather it is, to pick two very recent examples which, for the sake of discretion I cannot be more specific about (but which you can surely recognise), one major international art gallery not letting me film in front of a painting because they don’t want to discuss its attribution publicly, and another major international art gallery attempting to prevent me from illustrating a piece with one of their paintings, because they disagreed with what I and others had to say about the attribution. You say such institutions are being ‘brave’ to leave pictures unattributed. I say they are refusing to engage with wider scholarship, to say nothing of protecting their own reputations.