The painting shows the Madonna seated frontally on an elaborate, curved, two-tier, wooden throne of circular plan. She is supporting the blessing Christ child on her left arm according to the iconographic tradition of the Hodegetria. Mary is wearing a red mantle over an azure dress. The child is dressed in a salmon-colored tunic and blue mantle; he holds a red scroll in his left hand, supporting it on his lap. In the upper corners of the panel, at the height of the Virgin’s head, two medallions contain busts of two archangels [fig. 1] [fig. 2], with their garments surmounted by loroi and with scepters and spheres in their hands.
It was Bernard Berenson (1921) who recognized the common authorship of this work and Enthroned Madonna and Child and who concluded — though admitting he had no specialized knowledge of art of this cultural area — that they were probably works executed in Constantinople around 1200. These conclusions retain their authority and continue to stir debate. Of the various alternative proposals expressed thus far, only those that have considered them the products of a thirteenth-century Roman or Venetian painter have been definitively abandoned. Berenson’s opinion of the purely Byzantine figurative culture of the two panels still commands wide support, even though the dating of the paintings is in general placed slightly later, c. 1250 or within the second half of the thirteenth century. George Martin Richter’s view that the master who painted them was a Greek active at the time of the Norman kings in Sicily remains isolated. But a Sicilian origin is still supported by those scholars who follow the opinion of Viktor Lazarev that the Kahn and Mellon Madonnas were painted in Sicily during the last quarter of the thirteenth century, by artists who had come from Byzantium or by their local disciples. Some authorities on Byzantine art doubt, however, that the two Madonnas in the National Gallery of Art could have been executed by masters of Byzantine origin. In their view, also endorsed by some experts of Italian medieval painting, the characteristics of the two panels presuppose patrons with liturgical needs, iconographic precedents, and cultural traditions consistent with those of the Latin West. Further, they have asserted that the two paintings probably were executed in Tuscany. Other scholars have supported the thesis that the two panels were painted in Cyprus or in Thessaloniki.
Opinions thus vary widely regarding the cultural origin of this painting and the Enthroned Madonna and Child, but significant convergences can be ascertained on some points. It seems generally recognized, for example, that the Kahn Madonna is the earlier of the two images and that it was painted by an artist trained in Byzantium, although possibly at work far from his homeland, and for patrons of Western culture. The close stylistic affinities between the two panels are also commonly recognized, even if the Mellon Madonna is often indicated as the work of a different hand. In more recent years, however, various scholars have again proposed their common authorship.
The sudden appearance of the two panels in tandem and their common provenance from a small Spanish town have been considered strange; indeed, their alleged provenance from a church in Calahorra is authenticated, as far as I know, by no document. On the other hand, the fact that at least in 1912 the two paintings were reported together on the Madrid art market supports a Spanish provenance or at least makes it probable that they were purchased in Spain, perhaps even from the same owner. It would seem strange, in any case, to invent a provenance from Calahorra for paintings allegedly executed by Cimabue or by Cavallini. Various reasons have been adduced to cast doubt on whether both panels could have been executed by the same artist. Roger Fry was the first to emphasize the different structure and pictorial treatment of the two thrones. Many years later, Hans Belting cited the results of technical examinations by Ann Hoenigswald in support of his thesis that “most of the analogies which link the Mellon Madonna to the Kahn Madonna must . . . be interpreted as a deliberate attempt by the artist of the former to imitate the latter or a similar work.” Several scholars have accepted Belting’s conclusions, although it is fair to say that, in order to exclude the use of different technical procedures by the same painter, one would first need the evidence of systematic analyses — hitherto unavailable — of the pictorial technique of a particular painter in various phases of his career. How can we exclude a priori the possibility of an artist having developed, or having experimented with, new solutions, or different technical procedures, in the course of his career? While we may freely admit that there are some technical differences in the execution of the two Madonnas in the Gallery, we cannot infer from this that they are the work of different hands. Still less can any such conviction dispense us from a careful stylistic analysis of the two images.
The claim is made occasionally that modern restoration drastically altered the Mellon Madonna. A comparison of photographs taken c. 1920 with those taken in more recent times proves indeed that modern conservation measures have been extensive. Yet many parts of the painting that are still clearly legible disclose close resemblances with Enthroned Madonna and Child. The affinities concern not only the formal treatment of such details as the faces of the two images of Mary [fig. 3] [fig. 4] or the delicately poetic interpretation of the main personages, but also the high, sustained qualitative level of the execution of both panels, the skilful use of chrysography to articulate and give relief to the bodies, and some peculiar incongruities in the perspective of the thrones. In the Kahn Madonna the artist proposes a rectangular wooden throne with a very elaborate structure and tries to make it illusionistically credible, not only by the three-dimensional modeling of the individual components but also by making the right rear leg visible through the perforations. But he fails to remove incongruities in perspective: while the throne itself is seen from the left, the four legs are represented as if seen from the other side. In the Mellon Madonna, apart from the painter’s difficulty in convincingly rendering the foreshortening of the circular seat, it may be noted once again that the two front legs are seen from opposite directions: one from the left, the other from the right. It is difficult to think that all this is merely the consequence of the painter’s effort to follow his model, not least because the Mellon Madonna is decidedly not a copy of the other. The two images are in fact different in type and perhaps also in function. Their common provenance seems to suggest that they were executed for the same patron but belong to two different phases of the artist’s career and were intended to satisfy different needs.
Of the comparisons hitherto proposed, the one indicated by Otto Demus, of the Madonna of the Deesis in the southern gallery of Hagia Sophia, Istanbul [fig. 5], a mosaic probably executed immediately after the Byzantine reconquest of Constantinople in 1261, seems particularly significant in support of the Constantinopolitan culture of the artist of the two paintings. However, other images too, whether in the form of frescoes, panels, or miniatures, seem to suggest a similar origin of the two Madonnas, the execution of which is unlikely to have been much later than the Deesis in Hagia Sophia. None of the works hitherto compared with the Byzantine panels in the Gallery in fact seems so close to them as this mosaic in Hagia Sophia, both in facial type and in the extreme delicacy of the modeling; this close kinship represents, in my view, the confirmation — despite frequent denials — of the common authorship of the two paintings.
It is understandable that some decorative forms used in the panels — for example, the elaborate incised decoration of the halos of the Kahn Madonna and the type of throne — should have suggested the artist’s contact with Tuscan figurative culture. But closer analysis will show that the decorative motifs of the halos, as well as the wooden thrones of the Washington paintings, are well known in Byzantine art. The genre of the Madonna and Child Enthroned on a monumental scale, particularly popular in thirteenth-century Tuscany, was also well known in contemporary central and southern Italian painting and in that of the Venetian and Adriatic area. Influenced, in its realistic elements, by icons imported from the various centers of Byzantine art and by products of so-called Crusader art, these large images were at times commissioned from artists of Byzantine origin and training: this seems to me the case of the Madonna in the Basilica of San Nicola at Bari. It seems plausible to suggest a similar origin for the Kahn and Mellon Madonnas, though these are distinguished from the icon in Bari by the particular accomplishment of their execution and by their stylistic kinship with works having at least a probable Constantinopolitan origin. In conclusion, therefore, while admitting that the origin of these paintings will probably long remain a bone of contention, the present state of our knowledge suggests that they were produced in a workshop culturally bound to Constantinople in years not long before or not long after 1261, the presumed date of the image they most closely resemble, the mosaic in Hagia Sophia.
Miklós Boskovits (1935–2011)
March 21, 2016