Skip to content

Pastels at the Louvre

10 June 2018

cvrEveryone reading this will now know that the Louvre’s long-awaited pastel exhibition has just opened (until 10 September), and I thought some preliminary impressions might be helpful. I have not yet had an opportunity to study Xavier Salmon’s catalogue raisonné of the collection whose publication coincides with the exhibition, and which doubles as a catalogue. I have seen it, but will confine this post to observations about the exhibition only. [See this later post for detailed comments.]

Anyone with the slightest interest must go to see the show. Most of the reviews that appear will inevitably focus on the great works in the exhibition, and tell us why La Tour and Perronneau are important. And rightly so – but all my readers know that already, or at least know that I think so. The music critic doesn’t have to take up space explaining why a Beethoven late quartet is important (perhaps a solo sonata would be a better analogy in this case), but launches straight into a discussion of the performance, not the piece. Which I shall do – after of course noting that a show which includes 20 pastels by La Tour and four or more by each of ten more artists (Vivien, Carriera, Lundberg, Chardin, Perronneau, Boze, Ducreux, Labille-Guiard, Vigée Le Brun and John Russell) cannot but be a triumphant success (which in those terms the exhibition certainly is). What could possibly go wrong?

Curatorial performance has many dimensions. First is getting people through the door. Standard practice is to arrange all sorts of enticements – inviting prominent specialists (or even sociétaires des Amis du Louvre) to attend a vernissage might be one, while failing to organise a scholarly colloquium to discuss findings seems rather more important (unless there is one to which I also haven’t been asked). Neither the title “En Société” (apparently an afterthought, with unfortunate resonances with the title of the recent Rijksmuseum show High Society) nor the bizarre graphic immediately outside the exhibition seem likely to draw in many passers-by or give any intimation of what delights await:


A story?

Far more important of course is the “hang”. Here there are again many aspects. First of all, what is the logic or narrative? This exhibition is hamstrung by its association with a book whose own structure and compass are curious. While beautifully produced, intellectually it is essentially an update of Geneviève Monnier’s catalogue from 1972: so we follow the division into seventeenth and eighteenth centuries, with artists then ranged in alphabetical order. Of course that wouldn’t do here, so we start off with some chronological progression – Vivien, Carriera, La Tour etc. – but thereafter non-specialists are likely to get lost: confused by a wall which ends with Regnault, in an Empire frame, but dated 1765, while behind have been Hoin, Greuze etc.

This could easily have been dealt with by the most obvious of solutions: print the date (whether known or estimated) prominently on the labels (or cartels in French). But there are no dates systematically given on the cartels. This is inexplicable. Sometimes they are discovered within the text. Sometimes they are given, but are simply wrong. How in the world a cartel labelling two pastels by Vivien as his “morceaux de réception à l’Académie royale… le 28 juin 1698” got past checking I don’t know. There are other such howlers in the wall texts (further examples are discussed below): Louis Vigée, we are told, continued to use pastel after La Tour and Perronneau – although he died more than 15 years before either. None of this assists in communicating a coherent story to the public.

Another astonishing hole in the project is the complete absence of any explanation about pastel as a medium. There are no heuristic aids to tell visitors what pastel is, how it is made, how it differs from natural materials, or how drawing with coloured chalks evolved during the seventeenth century into painting in pastel. I would be personally sympathetic to this approach if I felt it marked the maturity of public interest in the medium, but I am surprised that a museum such as the Louvre felt it an appropriate level at which to present the subject. But even within its own terms the compass of the show, and that of the catalogue, share Mme Monnier’s definition of pastels as complete works, distinct from sheets with touches of pastel, as set out in the Avertissement on p. 31 of the catalogue – but then go on to confuse by including (but not exhibiting) sheets by Deshays and Natoire.[1]

Physical description

Instead of dates, however, the cartels focus in obsessive detail on certain matters of construction. We told for example that one is in

Pastel sur quatre feuilles de papier gris-bleu assemblées à joints couvrants marouflées sur une toile imprimée d’une couche de préparation de couleur rouge-brun tendue sur châssis

while another is

Pastel sur plus de treize feuilles de papier bleu raboutées à joints couvrants marouflées sur toile tendue sur châssis

This is information of interest to specialists, but not I think to the general public, and is far better restricted to the catalogue. Sizes would probably be of more interest, but are not given. It would also be helpful if the information were consistently presented, and matched with other scientific descriptions – there seems for example little agreement on whether paper is blue or grey compared with recent publications, or whether there are 13 or 12 sheets on a particular work. What is of significance (unless you merely wish to evidence the curator’s close inspection) is where the joins are, which pieces have been isolated, why and when (for example, are the heads done on separate sheets and pasted into larger works where working in situ would have been awkward?). None of this is presented, in the exhibition or catalogue, although maps showing the joins turned the current Getty show (Pastels in Pieces, to 29 July) into a far more interesting report. Another particular point is the references to “gouache” which are probably simply wet pastel (whether applied with the tip of the pastel moistened, or ground into dust, mixed with fluid and applied with a brush). Again many of the pastels that have this are not so described, while others are.


The opportunity has not been taken to explore the frames in similar detail. This is to be regretted, as many are original and of very great interest (I may write more about this later). Others are later Louvre frames of Empire style which are to be expected. But there have been a number of less satisfactory recent additions. I don’t know why the Bartolozzi is in a Kent frame, a style that went out of fashion in England when Russell was born (perhaps this is less obvious to a French audience). A particularly unfortunate intervention is with the Perronneau Mlle Huquier, which formerly had an elaborate spandrel with curved corners which neatly concealed the tear in the lower left corner. That has now been removed and replaced by a bright straight-edged slip which serves only to reveal that the frame never fitted. This is a case for reframing completely if we want to see the whole pastel and enjoy it as Perronneau originally intended (many of his original frames were very modest and were widely changed c.1900 for more prestigious ones).


What would also be of great interest is to have comments on the condition and losses which these works have endured. While the catalogue goes into meticulous detail on recent interventions, it rarely provides explanations as to why we have misread images (the nun’s nose is perhaps an example, J.46.2183). There is nothing in the exhibition, and little on a first glance at the catalogue, which reveals scientific examination of these works – none of the spectroscopy or other scientific analysis which the Rijksmuseum for example have applied to their pastel collection and which might allow us to detect the presence of fixatives or later interventions with anachronistic pigments.

The catalogue also informs us that 11 restorers have worked on the collection for six years, which perhaps explains the obsession with descriptions of the physical construction. Of course we all want to see these works preserved to the highest standard, but this is a surprising amount of intervention not all of which I think it fair to say has been equally successful, but this isn’t the right place for a detailed discussion.

But one intervention in particular raised my eyebrow. The debate about what to do with old glass has been raging for years (you can find more about glazing pastels in my Prolegomena, §§ IV.15 and V.9): it is more fragile than modern replacements (and so too risky to travel) but its appearance is prized by connoisseurs. For reasons that escape me for a collection that is not supposed to travel, there has been a fairly systematic campaign not of removing the old glass, but of putting a second sheet of Mirogard behind it. While Mirogard is definitely preferable to acrylic alternatives (Optium is particularly popular, but has many drawbacks), the double glazing solution seems as dubious as the wares normally sold under that name. The idea misses the point of what connoisseurs value – the integrity of the original object, the assembly itself being part of the work of art. It is indeed a curious interpretation of the ICON duty to “conserve cultural heritage [as] reliable evidence of the past.” Whatever the theory (or deontology as the French might put it), there are practical objections. The installation may require deepening the rebate, and will certainly result in a considerable increase in overall weight, putting unnecessary strain on the frames and increasing vulnerability to shock. But the most obvious point that this exhibition makes plain (particularly because of the positioning of many of the works) is that the assembly results in bizarre double reflections from the lighting equipment. Mirogard’s principal fault is that it reflects white light as green. With the double sheeting you see each spot reflected as two, slightly separated ghost images, one white, the other green. It’s a weird effect, and once noted very disturbing. It shatters any illusion of being in the dix-huitième.

Lighting and hang

Readers of this blog will know that I am not a great fan of directional rather than ambient lighting for pastels: raking light can cruelly expose any conservation problems (including after restoration). Here we must praise the fact that the colour temperature has been kept down (avoiding the harsh colour distortions resulting from some equipment), but the lighting and the points made above on reflections and raking light take me to what I think is one of the most serious criticisms of the hang. The exhibition includes 115 pastels (not the 120 promised – see below) plus four drawings. Of these, twenty-four have been skied – hung as a second row, at a level at which only someone as tall as the curator could see them.

This was an extraordinary decision. Nor were the works concerned confined to the weaker examples: they include three of the very finest pastels in the show, La Tour’s Maurice de Saxe, Perronneau’s “Bastard” and one of the Chardin self-portraits. They are the ones that suffer most from the raking light and reflection problems. Even dirty glass (e.g. greasy streaks on La Tour’s Lemoyne) is painfully evident under these conditions (a good many of the pastels evidently recently bore sticky labels, approximately 1×5 cm, in the top left corner of the glass, the residue of which has not been cleaned properly). But housekeeping aside, it is a real shame that pastels of this quality that have not been visible for years (and presumably won’t again for another generation) should be exposed where they cannot be seen.

Double rows in displays are not unheard of. In many ways this show sites itself intellectually with the great exhibitions of the past, the famous Cent pastels of 1908 or that of 1927, and it is true that the latter had a wall of Perronneaus in two ranks. But compare these hangs for elegance and symmetry:

Display2018Perronneau paneau 1927

The current hang is dense, crowded and simply untidy. What a pity.

Wall colour

But nothing to the second and gravest issue with the presentation: the choice of wall colour. The second part of the show has a sort of crushed raspberry hue: it’s not unfamiliar in the Louvre, but I can’t say I like it much. M. Salmon’s previous choices, such as the crimson for his Versailles show, were far better. But it is the colour for the first rooms, and the final one, which I find the most baffling. Images on social media do not capture it well: cameras find it hard to locate the precise hue somewhere between light sage and mustard. I don’t know if this is the colour Germain Bazin called “vomis d’ivrogne”, but that is a more precise description than any I can muster without feeling queasy.

This isn’t just a matter of design. What colour you paint the walls can have a transformational effect on the pictures you put against them, particularly when, as with pastels, their whole effect depends on colour. Balance, harmonies and the very essence of a picture can be destroyed. Those of you who recall the great Chardin exhibition at the Grand Palais in 1999 will remember just how magically these great self-portraits came to life: here they lie struggling for breath. If Oscar Wilde and his wallpaper were fighting a duel to the death, this greenery-yallery would surely have hastened his demise.

The cartels

For many visitors who do not have the catalogue to hand, or cannot afford it, the cartels are the opportunity to tell the story, choosing something that will draw people’s attention to the significance of what they should find when they look at the work. Many of the cartels are banal and unhelpful. Others are hardly original. On Valade, all they can think of saying is:

Valade fut avec La Tour et Perronneau l’artiste qui, entre 1751 et 1769, exposa au Salon le plus grand nombre de portraits peints au pastel.

This comes straight from Ratous de Limay (1946), p. 81:

Valade fut, avec La Tour et Perronneau, l’un des académiciens qui, entre 1751 et 1769, exposèrent le plus de portraits au pastel aux Salons du Louvre.

Some errors

I should perhaps highlight a number of mistakes in the cartels in the hope that they can be changed. They should have been reviewed by someone familiar with the subject. Apart from those noted above, there are some issues with names, foremost among which is the reference to “Jean-Baptiste Siméon Chardin”, a well-known chestnut. Pierre Rosenberg sorted this out in 1979, but a quick reference to the online Dictionary of pastellists will remind you that Chardin’s names were simply Jean Siméon. The erroneous Jean Baptiste Siméon arises from an error in his inventaire après décès (18.xii.1779, AN mc/rs//921), but was the subject of a notarial deed of rectification (4.iii.1780, AN mc/lvi/248). If M. Salmon has found evidence that overturns this, he should publish it. There are other problems with names (I won’t dwell on the examples, which include Lenoir and Coypel, proving my contention that the hyphenation convention for forenames adopted by M. Salmon always results in inconsistency). Séguier has an acute. L’Effroi needs a capital E if the title is to be meaningful. Some of the foreign names are given in French, others not, some sort of. (For Fredrik I, born von Hessen-Kassel, you could try German, Swedish or French, but adopting the English Frederick seems odd.)

Nor are dates any more successful. Mmes Labille-Guiard and Vigée Le Brun were indeed admitted to the Académie royale at the same time – but 1783, not 1774. M. Salmon has adopted my specific identification of the Rozeville couple (I am pleased to see how much of my work he has drawn on throughout the catalogue), but has decided to assign them new dates. But a few seconds on the Archives nationales website would show him that Marie-Angélique did indeed die in 1762, not 1787 (perhaps if he had explored the genealogy further he would have realised that her daughter-in-law was also a Colignon). Similarly it would not take long to discover that Couturier de Flotte died in 1780 (in Paris on 9 February), not as stated (an error drawn from secondary sources). (Incidentally the inventory number is RF 1697, not RF 1967.) Again a proper examination of this family would have revealed that his daughter Marthe-Lydie-Olympe Couturier de Flotte (1768–1836) married, in 1788, Jean-Pierre Dussumier (1761–1802), so the Louvre donor was far more likely to be from the Dussumier de Fonbrune family than the (as far as I am aware unrelated) Poussou de Fontbrune family. Perhaps as much attention should have been given to the provenance of this collection as to the conservation details. An analytical index of collectors would be interesting, but there is none here, as there was none in Monnier (readers can always resort to my index of collectors).

Rosalba JF avec singe Louvre 4798Another disappointment concerns the Rosalba little girl with a monkey, the future marquise d’Havrincourt, née Antoinette-Barbonne-Thérèse Languet de Gergy. It would be useful for the cartel to tell us that this is the “ritratto della figlia dell’Ambasciator di Francia” recorded in the artist’s diary on 13 May 1725. (My annotated transcription of Rosalba’s diaries is here.) But as she appears to be about 8, not 2, M. Salmon cannot do that as he is under the widespread impression (floating round on the internet) that she was born in 1723. But in fact, as I explained in a previous post on this blog, she was actually born in Regensburg on 6 June 1717.

Another surprising comment is on the Le Brun pastel of Louis XIV which, we are told with confidence, is the model for the frontispice for Colbert fils’s thesis, which Véronique Meyer, the great specialist in these matters, has specifically challenged in her definitive study, Pour la plus grande gloire du roi: Louis XIV en thèses (Rennes, 2017, p. 189). Even if M. Salmon wants to sustain his view, it seems odd to flatly state it without discussion.

The Louvre’s collection

What then of the Louvre’s collection viewed outside the context of the presentation in this show? Readers will know from my earlier blog that the Louvre has not always had the most enlightened acquisition policy. Let’s turn to the numbers in that context.

As mentioned above, there are 115 pastels in the exhibition. The no-shows appear to be among the recuperated works which the museum holds on trust for the victims indefinitely. Two of these (by Perronneau) have recently been handsomely installed in the newly opened cabinet de pastels in Orléans, and it is unsurprising that they have not come (although cartels were prepared assuming they would):

Orleans MNR

Photo mBA Orleans, social media

Among the other disappointments are the Labille-Guiard of Catherine-Flore Pajou and the disputed M. d’Albespierre. But the cartels for the MNR pieces are bizarre: the information is appropriate for a catalogue, but these look like legal documents, with “comme de” heading even when they are thought wrong. For example we cannot tell from the cartel whether M. Salmon does or does not agree with the attribution to La Tour of Carlin (REC 8; my J.758.138).[2]


Restitution would be better served by explaining these works rather than setting out legal arguments which are of course freely available online. And perhaps they could have been integrated into the main hang.

Of the 115 some 75 come from just 11 artists. But almost all of the works are French – hardly surprising, and entirely justifiable in terms of the dominance of France in the eighteenth century. Of course, although billed as pastels of the seventeenth and eighteenth centuries, this is essentially an eighteenth century show: the Vivien group at the beginning are all close to 1700, and there are really only three purely seventeenth century works in the exhibition. The final three, by Prud’hon, all have a definite dix-neuvième aesthetic and look completely out of place in the show. Whether they are actually made after 1800 cannot be determined from the cartels as no dates are given. Prud’hon appears in the Dictionary of pastellists before 1800 because most of his pastels were made in the eighteenth century, and all of them are there presented together.

What is more remarkable is the nationality. Accepting Lundberg (in Paris for 37 years) and (more hesitantly) Carriera as effectively French, practically all the works are French school apart from the group of four Russells. M. Salmon is aware of this point, and hopes to rectify it by future acquisitions, “for example, of Cornelis Troost, Anton Raphaël Mengs or Hugh Douglas Hamilton”. Indeed. One might add Vaillant, Ashfield, Hoare, Cotes, Copley, Luti (why wasn’t his self-portrait shown?), Fratellini, Tiepolo, Rotari, Schröder and Roslin. So despite the certain victory in the exhibition stakes for a show with 20 La Tours, one has to concede that the Met’s 2011 show had a better balance in terms of giving an overview of pastel as a European phenomenon. In fact there might have been an argument for omitting the Russells and making this the definitive study of French pastel in the eighteenth century.

Except that is isn’t. There’s another gap right at the heart of this project, and while it would have been far easier to plug it is much less conspicuous and easily overlooked. This is that the story of pastel in France in the eighteenth century is far deeper than just the top names. While other countries had talented individuals (some of the names above), France had a system which led to a great many pastellists capable of producing wonderful examples. Few of them are household names, but the single example of the recently acquired Lenoir pastel makes this point clearly. To it one could have added (even on a very limited budget) examples of gorgeous work by Hubert Drouais, Glain, Lion, Pougin de Saint-Aubin, Davesne, Saint-Michel, Hall, Capet, Mme Gault de Saint-Germain, Berjon etc., not to mention better and more typical examples by Allais, Bernard, Hoin, etc. (Of course many of these artists are uneven, and most of their work is not of Louvre quality; but the examples that are should be embraced and promoted.) And although the Louvre is already rich in works by Vivien, all are in the vein of his official portraits: several recent examples, most notably the abbé Lalouette (J.77.248, now in Stockholm) which I discovered recently, would have provided a glimpse of the other side of his talent.

Instead these are the only pastels in the show that have been purchased since Monnier (a Ducreux autoportrait was received by legacy in 1985):

  • Liotard, Mme Tronchin 1982
  • Perronneau, Tassin 1985
  • Greuze, L’Effroi 1986
  • Hoin, Tete 1987
  • La Tour, auto à l’index, 2005
  • Vigée Le Brun, Jules de Polignac 2007
  • La Tour, préparation dite de Mme de Pompadour 2008
  • Russell, Bartolozzi 2008
  • Lenoir, Lekain 2013
  • La Tour, Mlle de La Fontaine Solare 2014
  • Vigée Le Brun, duc d’Orléans & Mme de Montesson 2014

The message (with the exception of the Lenoir, itself not a typical work) is more of the same rather than a conscious attempt to rebalance. The most recent examples, the two Vigée Le Brun pastels which I first discovered in 2013 and which I first published (although M. Salmon does not consider this worth reporting in his bibliography), are rather weak repetitions and arguably not really of Louvre quality (this was evident in the Vigée Le Brun exhibition in 2015, where the better version of the duc made the point, and again today where each of the four Labille-Guiards comprehensively trumped the Vigée Le Bruns.)

I cannot pass over in silence one of the ironies in the hang, where the Louvre’s sole Liotard is placed between Valade and – yes – Perronneau. Was this a subtle allusion to the very French view of Liotard of one of M. Salmon’s precedessors, that “Ses pastels, tant vantés par ses contemporains et ses compatriotes, n’égalent pas le moindre ouvrage d’un élève de Perronneau”? (You can of course find the reference in the Liotard article in The opportunity to discuss this is not however taken.


Since the Louvre collections have been the subject of vast research it is hardly surprising that there are relatively few problems of attribution for the works in the show. I will mention only a few here. Of course like everything else in this blog I offer a personal opinion only.

Inv. 24780 & 24781 /J.173.873 & J.173.874. Le petit dénicheur & La petite oiselière: I am surprised that the pastels Monnier catalogued as copies of Boucher have now been elevated to “attributed to” him.

RF 29662 & 29661/ J.47.1124 & J.47.1125. The Rozeville couple (mentioned above) are here attributed to Frey on the basis of a vague compositional similarity to the Jacquemart-André pastel by him which M. Salmon admits is in a poor state of repair (while astonishingly considering that the Louvre pastel “a conservé toute sa fleur”). But comparing the face of the Louvre pastel with one of another Frey in better condition (J.329.133, identified by Laurent Hugues, left; Mme de Rozeville is right) shows why the technique is completely different from Frey’s whatever the compositional similarity:

Salmon dismisses my proposed attribution to Lefèvre on the basis that his work is less “psychological”. Judge for yourselves whether there is (as I suggest) a similarity of facial expression (both the Louvre pastels share rather bovine, dim demeanours), of composition and of technique with the pastel by Lefèvre signed and dated 1743 (J.47.12, right; M. de Rozeville is left):

M. Salmon also provides no account of the social situation of these clients. Frey worked for the court; Lefèvre for a Parisian clientele, including people just like M. de Rozeville, who was an avocat au parlement de Paris.

RF 4241/J.103.126. M. Salmon has previously published the pastel of Nicole Ricard as by Lenoir, several times. I’m glad he’s retreated to École française, noting merely similarity with the Boston pastel by Lenoir (who in fact has a completely different technique – as you can now see for yourself as the pastel hangs immediately opposite the Louvre’s new Lenoir). My attribution to Allais in 2012 remains I believe far more plausible. M. Salmon rejects this on the basis that Allais’s technique is more graphic, less modelled. But those are precisely the reasons for my attribution, together with the characteristic treatment of the hair and the use of black chalk in the passementerie, as is evident to some degree from the other pastel by Allais[3] in the exhibition, but perhaps more clearly in this example signed and dated 1741 (J.103.221):


Much as we owe to the Louvre and to the many people involved in so large a project for the opportunity to see these wonderful treasures, I think it will be clear that I should have favoured a more accessible and collaborative approach in presenting it to the public. You must of course see it for yourselves.


[1] The Avertissement goes on to justify the exclusion of almost all reference to my website Pastels & pastellists ( on the basis that M. Salmon’s bibliographies do not cite dictionaries – despite the fact that he does cite, for example, Audin & Vial’s Dictionnaire…, and has extensive reference to Ratouis de Limay’s Le Pastel en France, 1946, which is nothing other than a dictionary with a few of the longer articles placed in the front of the book. Everything in that book will be found included or corrected in my “dictionary”, which has 15,000 reproductions in place of Ratouis de Limay’s 100 – and a great deal of information about the artists whose work M. Salmon catalogues which I suggest might well be of interest to his readers.

[2] In the catalogue M. Salmon makes it clear he does not think it can be by La Tour. He is right; neither do I. In the 2006 Dictionary and online until now, I listed it among French school, noting a possible attribution to Louis Vigée. Because it was originally referred to as by La Tour, I have a brief cross reference in the La Tour chapter to the main entry. M. Salmon does not cite the real entry, but does cite my La Tour chapter where the cross reference is placed. Anyone reading this page of his book would conclude that I think the work is by La Tour, and that M. Salmon is correcting me when in fact he is following me. (Postscript: I have now moved the main entry from French school to attributed to Vigée.)

[3] REC 9/ J.103.186. I first reattributed this work to Allais, which had traditionally been attributed to Heinsius. I inserted a cross-reference from Heinsius to Allais in the Heinsius article. Bafflingly M. Salmon cites the cross reference but not the entry in the Allais article, making it appear that I retain the Heinsius attribution.


From → Art history

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: