La Tour’s second thoughts

So much of my work on La Tour has been unravelling and rejecting myth. Herodotus faced much the same problem with his sources, but eventually conceded “having condemned others’ opinions, I must now say what I think about these obscure matters.” The problem of course is in finding new, reliable sources of information – or else one is simply compounding the confusion. With an artist on whom so much scholarship has been devoted, entirely new sources are difficult to find. But sometimes crucial information has been hiding under our noses.

The legend about La Tour’s destruction of two of his masterpieces in a senile attempt to “improve” them is more than just a story: the evidence was shown to all in the Louvre exhibition in 2018. The sorry state of Dumont le Romain (left), and the even sorrier remnants of what was once Jean Restout (right), were bravely presented to an audience with a reasonable account of their confused history. You can find my version of this written up in the relevant entries in my online catalogue raisonné: Dumont at J.46.1681, Restout at J.46.2687. (Remember you find these by searching the J numbers in the search box on www.pastellists.com, opening the relevant pdf and going to the J number which is in a decimal sequence – so J.46.2787 is before J.46.279. Or you can go direct to the pdf from www.pastellists.com/LaTour.htm. You can also find a precis of the discussion below in §II.4 of my main La Tour article, www.pastellists.com/Articles/LaTour.pdf. As always the crucial contemporary documents are transcribed, with further references, in www.pastellists.com/Misc/LaTour_chronology.pdf.)

But here at any rate is a broad chronology of what must have happened.

At a session of the Académie royale de peinture et de sculpture on 25.v.1737 “le sieur Maurice-Quentin de La Tour, Peintre de portraits en pastel, aïant fait apporter de ses ouvrages” was agréé (provisionally accepted for membership). His set pieces for full reception were selected the following week: they were to be portraits of the academicians François Lemoyne and Jean Restout. Lemoyne committed suicide a few days later, and Jean-Baptiste Van Loo was nominated instead: but his departure to London and later return to his native Provence created a further hurdle, before La Tour eventually submitted Restout alone, in 1746, when he was finally reçu. Four years later he also presented the portrait of Dumont le Romain as a gift to the Académie; it is often erroneously described as a morceau de réception.

Some six years after that, in 1756, the Polish painter Tadeusz Kuntze copied both works. Although this has been in the Dictionary since 2015, the copies are not mentioned in other La Tour scholarship and their significance has escaped me until now. Tadeusz Konicz, dit Kuntze (Zielonej Górze 1727 – Rome 1793), was trained in Rome at the Académie de France, 1747–52, and stayed on to paint religious and allegorical pictures there in the tradition of Reni and Solimena. In 1756 he was sent to Paris where he made oil copies (all now in Wilanów) of artists’ portraits which had been acquired by the Académie royale (normally as morceaux de réception), including pastels by La Tour (Dumont le Romain, Restout) and Lundberg (Boucher, Natoire). He returned to Poland in 1757 before settling in Rome in 1759 and disappearing from our story. His copies were run of the mill, boringly but helpfully unimaginative.

A few years later attention focused on engravings of both La Tour pastels. Neither sitter had had a portrait engraved (the Cochin portrait of Dumont was engraved by Saint-Aubin only in 1770). The engraver Pierre-Étienne Moitte (1722–1780) – who also engraved La Tour’s portraits of Belle-Isle and of Jolyot de Crébillon – was agréé on 26.iv.1761, with Galloche, acting recteur, deputed to set him two subjects for his morceaux de reception. Evidently the La Tour portrait of Restout was one of those, but the other was never recorded. Six months later, in a session of the Académie where La Tour was present, the question of the format of the engraving was raised: La Tour’s portrait being deemed unsuited to the usual oval format (Moitte’s head of Crébillon for the Galerie française is no doubt what was in mind), the Académie decided that the whole portrait be engraved, but in view of the additional work required, this single engraving would suffice for Moitte’s reception. It was not however delivered until 1771 (although it must have been based on the pastel before its reworking already underway in 1769 – see below), for reasons unknown but one may speculate that La Tour’s dissatisfaction with his own work may have played a part in the delay. Nevertheless the engraving accurately corresponds to Kuntze’s 1756 copy of the original version of the pastel.

Separately the engraver Jean-Jacques Flipart (1719–1782) produced a full-length portrait after Dumont le Romain. (Again one may speculate that it was originally the other set piece for Moitte, but there is no evidence for this.) Flipart was agréé in 1755 but never reçu, and this engraving was not part of his Académie requirements. Apart from the La Tour, he engraved a self-portrait by Rosalba and a pastel by Vivien, but most of his work was not portraiture: he was best known for his genre pieces after Greuze. A 1772 Chasse au tigre, after Boucher (actually a leopard), is one of the few plates for which the engraver’s preparatory drawing survives (Paris, Drouot, Thierry de Maigret, 27.iii.2009, Lot 76).

The lettering on Flipart’s Restout includes the artist’s offices, and thus provides a terminus post quem for the plate (or at least the complete state), of 1768, when Restout was promoted to chancelier of the Académie. Unlike the Moitte, it is reversed from the pastel; and perhaps for this reason its departures from the Louvre work have gone unnoticed. It is equally possible that anyone comparing the print with the pastel would simply have assumed the alterations were the engraver’s fancy. That theory survives the discovery of what may be Flipart’s preparatory drawing in the Walker Art Gallery (again omitted from all La Tour scholarship to date) – a drawing which however is in the same sense as the pastel, reinforcing the suggestion that it may have been preparatory to the engraving; the lower part is unfinished. (The evidence of the Cleveland préparation is limited to the central fold in the turban, which matches far more closely the print than the Louvre pastel. The Restout préparation in Saint-Quentin, of stunning quality, with a knotted falling lock of hair as in the Kuntze copy, can however tell us nothing about the overall composition.)

But it is in both cases the almost exact match of the Flipart and Moitte engravings with the Kuntze copies that provides incontrovertible evidence of how the pastels looked in 1756 and during the 1760s before La Tour’s changes.

That changes were made is of course well documented. Shortly after Restout’s death in 1768 La Tour retrieved both portraits with the intention of “improving” them. Mariette mentions only Restout, while Diderot compounds the confusion when he interrupts his Salon de 1769 with an account of a visit to La Tour’s studio in which he suggests that La Tour is copying rather than altering the pastel of Restout he had borrowed:

Je sortais du Sallon; j’étais fatigué; je suis entré chez La Tour, cet homme singulier qui apprend le latin à cinquante-cinq ans, et qui a abandonné l’art dans lequel il excelle pour s’enfoncer dans les profondeurs de la métaphysique qui achèvera de lui déranger la tête. Je l’ai trouvé payant un tribut à la mémoire de Restout, dont il peignait le portrait d’après un autre de lui dont il n’était pas satisfait. O le beau jeu que je joue, me dit-il! Je ne saurais que gagner. Si je réussis, j’aurais l’éloge d’un bon artiste; si je ne réussis pas, il me restera celui de bon ami. Il m’avoua qu’il devait infiniment aux conseils de Restout, le seul homme du même talent qui lui ait paru vraiment communicatif, que c’était ce peintre qui lui avait appris à faire tourner une tête et à faire circuler l’air entre la figure et le fond en reflétant le côté éclairé sur le fond, et le fond sur le côté ombré; que soit la faute de Restout, soit la sienne, il avait eu toutes les peines du monde à saisir ce principe, malgré sa simplicité; que, lorsque le reflet est trop fort ou trop faible, en général vous ne rendez pas la nature, vous peignez; que vous êtes faible ou dur, et que vous n’êtes plus ni vrai ni harmonieux.

Diderot’s account at least offers an explanation of La Tour’s interest in a tribute to his recently deceased mentor. No such explanation can account for the assault on the Dumont pastel: the subject would live on to 1781.

The following year La Tour laid out the problems with the portrait of Restout in his long letter to Belle de Zuylen (5.iii.1770). The letter is too long to quote in full, but this is relevant:

C’est s’occuper de chimères, on ne fait ny tableaux ny poëmes tels que je les désire. Cette perfection est au-dessus de l’humanité; je l’éprouve actuellement: j’ay sur le chevallet le portrait de feu M. Restout, fait et donné à l’Académie en 1744; j’ay voulu depuis sa mort luy témoigner ma reconnoissance des grands principes de peinture qu’il m’a communiqué, en remaniant cet ouvrage. Après avoir fait cent changemens, on me dit « Quel dommage! » Il y avoit un mouvement qui se communiquoit à ceux qui le voyoient. Je suis encore après et ay changé jusqu’à ce jour; je ne puis dire quand il sera fini. On attend d’autres ouvrages faits anciennement, que j’ai eu en fantaisie de remanier; je les renverray si un compagnon de voyage arrive avant.

(Once again La Tour is confused about dates: his morceau de réception was presented to the Académie in 1746, not 1744.) But at least the letter makes it clear that what was under way was a “remaniement”, not a copying. The postscript disclosed that the Académie had required him to return the portrait of Restout, more or less as it was:

les regrets de l’Académie m’obligent de tacher de remettre le portrait de M. Restout à peu près comme il était. Voilà bien du temps perdu et des efforts in vanum. Mieux que bien est terrible! On ne se corrige pas, puisque j’ay tombé dans le cas plus de cent fois.

The pastels were presumably returned to the Académie soon after, or perhaps later. They were listed among the revolutionary seizures from the ci-devant Académie on 9.xii.1793, when they were inventoried in the Premier Garde-meuble with this note: “Ces deux tableaux sont perdus par l’auteur même qui, trop vieux, voulut les retoucher: on peut compter que les glaces.” In the 21.vii.1796 inventory, Phlipault noted that they had not been transported to the maison de Nesle with the other Académie pictures; the entry included the important note that by then they were “sans bordure”; if the glass too had been removed since 1793 that would have led to further losses beyond those inflicted by the artist.

But interesting though these verbal documents may be, they leave us completely ignorant of the visual issues which must be paramount in any art historical analysis. Further there is a limit to what scientific analysis alone can bring to this discussion: the use of multiple sheets (Restout we know is on 13 sheets of paper, Dumont on 5), repaired joins  etc. can offer little to tell us whether changes were made during the 1740s or thirty years later: no newly invented pigment or material is likely to be detectable.

What we can now say with confidence is that La Tour decided to make radical changes to both works for essentially aesthetic reasons. Those to Restout are less clear in view of the subsequent deterioration (so that what was deliberately altered and what has been damaged are sometimes irretrievably confused). Why round the corners of the canvas on the easel, unless La Tour had developed a dislike of linearity (or had noticed that the perspective of the plane of the canvas and that of the easel were incorrect)? Most curiously, the proper arm, visibly on a separate sheet on the Louvre pastel, seems if anything to be closer to the original than the rest of the work. The most important alteration is that the portfolio resting on the artist’s knees on which he has been drawing has been replaced by a far less ambitious work table covered in baize. The gesture of drawing, with the porte-crayon now resting on cloth, makes no sense. But the covering up of the sitter’s legs has transformed this three-quarter length portrait jusqu’aux genoux into a half-length image: what is lost in sense is gained in proximity: it is in the current vernacular up close and personal.

The changes to Dumont can be more exhaustively listed. Among the minor details, one notes that the original had a fuller background curtain, a rectangular palette with an oil reservoir, a larger group of brushes and a simpler table with no drawer, supporting different objects. The effect of these differences, notably in the table, is again, and even more dramatically, to change to viewpoint, providing a di sotto in sù perspective (unique in the œuvre) which served to make the portrait both more intimate and more reverential.

What is clearly happening illustrates La Tour’s problems with the viewpoint, one of numerous particular difficulties facing the portraitist on which he wrote at great length to Marigny in 1763:

Les gens délicats sont blessés d’un tableau dont le point de distance est près et n’a pas au moins vingt-cinq pieds. Partant de ce principe, quel embarras pour une vûe courte et foible, forcée d’être à deux ou trois pieds du modelle, obligée de se hausser et baisser à mesure, de tourner à droite, à gauche, pour tâcher d’appercevoir de près ce qu’on ne peut voir bien que de loin! Il faudroit être à ma place pour sentir les efforts que je fais pour mettre une figure et une teste ensemble dans les règles de la perspective. Les angles sont si courts que la personne qu’on peint de près ne peut pas regarder de ses deux yeux à la fois l’œil du peintre. Ils vont et viennent sans être jamais ensemble. C’est pourtant de leur parfait accord que résulte l’âme et la vie du portrait. De la naissent les inquiétudes qui occasionnent tant de changements qu’ils font passer le malheureux peintre pour fou ou tout au moins capricieux, fantasque; à la vûe de tant de difficultés l’humeur gagne l’artiste et, au souvenir de M. Coypel qui n’a pas rempli les intentions du Roi, elle s’aigrit et s’éloigne de beaucoup de choses telles que des devoirs, des bienfaisances, etc.

In his letter to d’Angiviller in 1778, in which La Tour argues at length as to why he needs the use of an additional logement in the Louvre, spelling out all the difficulties consequent to his perfectionism he mentions perspective. And the postscript reinforces this:

J’ay oublié qu’il s’agit du portrait de M. Retout [sic], que j’ay enlevé pour un mot de critique de feu M. Toqué: c’est un maître à danser. Ce mot et le désir de donner aux élèves l’exemple avec le précepte de la perspective qui manquoit dans mes portraits sont les causes funestes des peines infinies que je me suis donné jusqu’à present. Dieu et Monsieur le Comte me soient en ayde, j’en ay un très grand besoin.

It may be possible to read this as indicating that La Tour had not returned Restout to the Académie as he had reported to Belle de Zuylen, or perhaps that he had borrowed it again; once again La Tour’s correspondence baffles us today as much as it baffled Marigny and other recipients at the time. But the evidence of Kuntze, Flipart and Moitte tells us much of what we need to know, and hadn’t troubled to see until now. The distant monuments to Pompadour and de Rieux are dismantled for these friends.

Postscript 2.i.2023

Further evidence of the appearance of the Restout pastel is provded by the oil self-portrait of Johann Heinrich Tischbein, c.1752–55 (Schloss Wilhelmshöhe) which, as has been observed by Heidrun Ludwig (Burlington magazine, .i.2023, p. 77) was inspired by the La Tour which he must have seen while studying in Paris with Carle Van Loo 1743–48. It clearly refers to the earlier state of the pastel.

Advertisement