Friday, 28 September 2012

Leonardo da Vinci. The Isleworth Mona Lisa


In an extraordinary bout of promotion, the Mona Lisa Foundation has captured incredibly wide media attention through the announcement on Thursday that they are in possession of the “earlier version” of the Mona Lisa – the portrait of Lisa Gherardini, wife of Francesco de Giocondo. The announcement, ostensibly comes from a non-profit research foundation, but the directors of the Foundation are to be identified as belonging to the syndicate of owners.

David Feldman, the major stamp dealer who is a director of the Foundation, kindly arranged for me to be sent a high resolution image and a copy of the glossy, gilt-edged book, which contains their “proof” that they own the first version of Leonardo’s portrait. I have not seen the painting in the original, but some things are so clear from the image and from their mish-mash of suppositions in the book that seeing the original is most unlikely to change my present conclusions.

The book, apparently written for the most part by Feldman’s brother, Stanley, is as physically impressive as it is historically slippery. There is no sense of how to distinguish core evidence, evaluate sources and construct arguments methodically. The piles of unstable hypotheses, stacked one on another, would not be acceptable from an undergraduate.

They (he?) says there must be a first Mona Lisa – the evidence shows this. Let’s cut back to basics. We now know, courtesy of the annotation by Agostino Vespucci in his edition of Cicero’s Letters to Friends, that the painting was underway in 1503. Vespucci, who knew Leonardo, mentions her “head” and that the painting was incomplete.
The next possible mention is the travel diaries of Antonio de Beatis, who visited Leonardo's French residence in the service of the Cardinal of Aragon. Antonio noted three pictures, one of which was of “a certain Florentine woman portrayed from life at the instance [instanza] of the late Magnificent Giuilano de’ Medici”. This might be the Mona Lisa , though Antonio’s precision as a source is questionable. He says that Leonardo suffered paralysis in his right hand and that we “cannot expect more good things from him”. Leonardo was left-handed. If the portrait is the Mona Lisa, it is possible that Giuliano, whom Leonardo served in Rome 1513-16, expressed interested in obtaining the portrait.  

In any event, the next really solid reference is in the 1525 list of the possessions of the cunning Salaì, who had obtained some key Leonardos that were in the master’s possession at his death. “La Gioconda” (i.e. the wife of Francesco de Giocondo) is recorded in the list designed to facilitate the division of the late Salaì’s possessions between his two sisters.  The best of Salaì’s Leonardos, including the Leda and the St Anne, entered the French Royal Collection at an unknown date, presumably during the lifetime of Francis I, Leonardo’s patron.

Where is the evidence for an earlier version of the younger Lisa? The most straightforward explanation consistent with the evidence is that there was one autograph portrait, never handed over the commissioner but retained (like other paintings) by Leonardo himself. We know that he was notably slow painter, and the physical evidence in the Louvre painting – different modes of handling and crack patterns – favours an extended period of execution. The painting may not even be quite finished now.

The book claims that none of the evidence of scientific examination indicates that the Isleworth picture is not by Leonardo. Nor does it show that it is not by Raphael. Even this ineffectual claim, with its double negative, is not justified. The infrared reflectogram and X-ray published on p. 253 do not reveal any of the characteristics of Leonardo’s preparatory methods. Leonardo, as the infrared images of the Louvre painting show, was an inveterate fiddler with his compositions even once he had begin to work on the primed surfaces of his panels. The images of the Isleworth canvas have the dull monotony that would be expected of a copy.

The carbon dating of the canvas on p. 246 produces a date band (broad as ever for carbon dating) that effectively ends in the early 15th century! Either the technique had gone awry or the linen was in existence at least 100 years before  the painter used it.

I see lots of dossiers of “scientific evidence” attached to purported Leonardos. It often seems enough to have the texts with the data, diagrams and images to “prove” the authenticity, whether or not the they actually tell us anything that actively supports Leonardo’s authorship.

When we come to look really carefully at the “Isleworth Mona Lisa”  it is evident that the copyist has failed to understand significant details and the suggestive subtlety of Leonardo’s image.  I could give a big list, but here are a few:
1) Lisa’s dress, as revealed by the gathered neckline in the Louvre painting, consists of a miraculously thin, translucent overlayer with thicker opaque cloth underneath. The copyist does not understand this structure and renders it lamely;
2) the spiralling veil over her left shoulder, rendered by Leonardo with depth and diaphanous vivacity, is transformed into a series of dull stripes of inert highlight;
3) Lisa’s  hair has that characteristic rivulet pattern in the Louvre painting, but is rendered in a routine manner in the Isleworth picture;
4) the veil beside Lisa’s right eye floats over the sky, rocks, water and her hair with extraordinary delicacy, with its meandering edge marked with a minutely thin, dark border – but not in the Isleworth version;
5) the folds of draperies in the latter are hard, routine and show little sense of the folding processes that are apparent in the Louvre painting;
6) the mid-ground hills / mountains in the Isleworth picture are painted in a thick, heavy-handed and opaque manner, with none of the optical elusiveness of Leonardo, and none of his living sense of the “body of the earth”;
7) the island on the left of the painting is truly bad – a literal blot on the landscape. There is no logic to the reflection and no other sign of the water that is responsible for the reflection;
8) the head in the Isleworth picture has been conventionally prettified in stock direction of the standard Renaissance image of the “beloved lady”. The idea, in the book, that Renaissance portraits of mature women can be used as accurate registers of the their actual age is misguided.

Everything points to the Isleworth painting being a copy. There is a comparable copy – island and all – in the National Museum in Oslo. Another is illustrated on p.199. There are families of copies of the Mona Lisa. This family of three is not the best.

And, on this flimsy but noisy basis, the Mona Lisa Foundation has the world-wide media jumping to attention. Any Leonardo story is mega-news. It is this phenomenon that is really notable in the current episode of Leo-mania. Leonardo would have been pleased. He was certainly looking for enduring fame.


  1. Dear Professor Kemp. Many thanks for this rational outline. I particularly enjoyed the eighth point - using a stylised portrait for any historical extrapolation can only be done if there are other sources to verify the artist had gone to pains to depict the sitter accurately.

    Immediately, the example of Isabelle d'Este comes to mind - whom Pietro Aretino mocked as disonestamente brutta in her older age - boldly instructing Titian to paint her in her youth. Her actual age and appearance may or may not have been suggested by the portrait which now survives in the form of a Rubens copy.

    Leonardo's portrait was highly popular and influential in its day, that it generated drawings (even by Raphael for example), and inspired other copies is not surprising, a phenomenon still present in the world of art and fashion to this day.

    How self-serving owners can distort this into "research" is always perplexing.

    Many kind regards

  2. 6) the mid-ground hills / mountains in the Isleworth picture are painted in a thick, heavy-handed and opaque manner, with none of the optical elusiveness of Leonardo, and none of his living sense of the “body of the earth”;
    7) the island on the left of the painting is truly bad – a literal blot on the landscape. There is no logic to the reflection and no other sign of the water that is responsible for the reflection;

    These are the qualities that stood out to me. I have heard that Leonardo's interpretation of the landscape came from his exposure to some Chinese silk paintings that were brought to the Vatican. The landscape, seen in that light, makes the other landscape look truly clumsy and clearly not conceived in the same way.

  3. Good to read your clear remarks.
    May I suggest something that puzzles me:
    If the Mona Lisa of Paris originally had eyebrows, as suggested by Lumiere Technology, then it is dubious that this Isleworth version has the same identical absence of eyebrows. If both versions had eyebrows in around 1510 then would the process of disapearrance have been identical in different atmospheres and conditions of storing?? It is hard to believe. Besides your remarks, I suspect very much that this Mona Lisa is a copy of the original of Paris and was executed when the Mona Lisa of Paris had lost his eyebrows... Also, it is a bit difficult to believe that the incredibly precise Leonardo would have failed to reproduce identically the shape of the face between the early version and the version of Paris. The faces are too different.

  4. This is just to add that the instructive Prado copy, proved to have been painted in parallel with Da Vinci (same pentimenti etc...), has kept her eyebrows...

  5. Thanks all for comments. I don't know about the Chinese paintings in the Vatican - but definitely worth thinking about.
    The issue of the eyebrows and the copies warrants a serious study, but most of the copies are not available in high res images. The Prado version, which may originate close to Leonardo, even in his studio, does indeed have pencil-thin eyebrows. What is needed is a systematic study of "families" of copies (like a family tree) of the kind that Thereza Wells undertook for our book on the "Madonna of the Yarnwinder". It may be that the presence / absence of eyebrows would be one of the main criteria. It would be a big project and involve some less than thrilling paintings. Any volunteers?

  6. Very interesting Dr. Kemp. I was just wondering about the two columns/pilars that are seen in Raphael's copy/sketch (circa. 1504 I think). Why aren't those columns present in the Louvre Mona Lisa? or is it possible that Raphael just added those columns out of his own imagination? Thanks in advance!

  7. Many thanks. The Louvre painting has not been cut down at the sides and the shafts of columns were always very slight slivers. Almost all the copyists were uncomfortable with this and made the columns wider.

    1. Thanks for the prompt reply Dr. Kemp. So, is it your view that there never were two versions of the Mona Lisa or do you feel that there may have been two versions? also, what do you make of Pascal Cotte's suggestion that there is an earlier version of the Mona Lisa underneath the Louvre version? Thank you! (May I also add that I watched a documentary about 'La Bella Principessa' and was very impressed by the work you did to attribute it to Leonardo)

  8. I can easily follow your argument but in my opinion the argument has more to do with those exploiting the portrait for reaping capital gain than dealing with the work itself. My major concern is with the present deceptive consortium of owners who are obviously exploiting smearing this meritorious portrait no matter of the artist responsible.

  9. My personal experience has led me to the conclusion that many attributions of art dealing with Leonardo's time are hinged on a certain amount of speculation. And that understanding that during his time it was not unusual for pupils and followers to participate in major works - Verrocchio's Baptism of Christ confirms that story.

  10. And so a certain amount of caution needs to be taken when attributing works to Italian masters. This is the case with the Gioconda portrait. There are things wrong with this painting. Drawing in the perspective from the column bases establishes a vanishing point in turn designating the horizon line in basic perspective geometry. This situates the horizon atop of the model's head - most unusual especially as other elements tend to suggest other horizon lines are in play-the landscape suggests two more horizon lines. One passes through the sitter's eyes, this being the most common when dealing with portraiture from this period, and another passing under her chin. With Leonardo's adamant pedantic approach to composition this appears an unlikely construction from him.

  11. Then there is a definite lack of shadow thrown from the column base on the left side of the balustrade - where Leonardo despises anyone who omits shadow in the rendering of the composition. And then there is the knot pattern that weaves across the bodice of the dress. If we look closely at the knot patterns on the Bella Principessa the motive appears to be repetitive but on closer examination the knot construction is different on each individual knot constructing the motif. And this is one example of many, possibly distinguishing the master's work from pupils. You may check this for yourself. And because Leonardo was fascinated with these types of motifs, tying them to his own name in sketches, this concept is not that far fetched.

  12. Other contemporaries used these types of patterns, such as Raphael, but through my research I've never discovered any one as meticulous to achieve such a unique result. Unfortunately the Gioconda picture has a repetitive knot construction in the motif - not really pointing to the master. In the end when comparing the Gioconda painting with other works from his followers there appears room for the work to be contributed to by others, such as Melzi but at the end of the day no-one will ever know the truth. And this is not only the case of attributing an artist to a work of art but also identifying the model.

  13. After consulting the sculptured bust and medallion from Romano, of Isabella d'Este, and the profile drawing from Leonardo of Isabella d'Este there is strong argument that she is a strong contender as being the model in the Gioconda picture. Coming back to the Isleworth painting - there is a misconception of how this painting should be scrutinised. If this canvas is perceived as a study for a portrait there are a number of determining features that present themselves. On close observation the background was obviously worked into the composition after the portrait had been painted and so any discussion on this naive addition is pointless. The heavy impasto that literally outlines the portrait and the overpaint on the columns is indicative that this was all an afterthought. The background in the development of this study was not the purpose of this exercise.

  14. The perspective is part of the construction. Drawing in the perspective from the column bases establishes the vanishing point. It hits the model in the corner of her mouth as if this invisible arrow points to the smile. This means that the horizon line cuts through her mouth. Leonardo uses these types of tricks and cryptograms to insinuate the identity of the sitter. As there is a relationship between between Giocondo - Gioconda (smile) the insinuation is instructive. As mentioned earlier the horizon line normally cuts through the eyes placing the model at the same level as the artist. Out of all the copies and version on the 'Gioconda' theme there is only one other work where this horizon phenomenon is present. It is present in Raphael's sketch with has often been associated to Leonardo's work on the Mona Lisa's portrait. And the other interesting observation is that this sketch includes the columns! The shadows on the balustrade in the Isleworth portrait are present and coherent with the composition.

  15. Yes there are a number of observations, such as the hands that appear not to be fully rendered to perfection neither the layering of the robes, but the face is beautifully constructed and worthy of note is that Leonardo in treatise whilst explaining how to paint flesh tones on canvas mentions a list of pigments to be used which corresponds to the list of pigments present in the face of the Isleworth portrait recorded by Dr Küln in 1977. This illustrates that Leonardo did experiment with canvas. The canvas is identical in character to the drapery studies attributed to Leonardo in the Louvre collection. The last observation is the knot pattern along the bodice of the dress. The motif is made up of individual constructed knots forming the motif. All in all one can not easily dismiss this work as 'uninteresting' but in the context of being a study, where Kenneth Clark was convinced that there is a missing link between the Gioconda picture and the Raphael sketch, makes this work worthy of note. But what is sure is while this portrait remains in the hands of greedy rogues it is uninteresting to pronounce it as anything of interest. Dr. Kemp please condone David Feldman and the Mona Lisa Foundation but refrain from passing judgement on a work, which is as equally open to debate and speculative argument as the Gioconda in the Louvre when attributing it to one master craftsman. Perhaps the word workshop should be added for a better comprehension of the period's practices.

  16. thank you for your attention