The Ghent Altarpiece by Hubert and Jan van Eyck, completed in 1432, ranks among the most important objects in art history. Composed of no less than twelve panels, it is both exceptionally large in size and complex in iconography and structure.
One major question and focus of art historical research is the altar’s genesis and – related to this – the original conception of its structure. Art historians have, for example, expressed doubt that all the panels belonged to one retable right from the beginning.
Most recently, Griet Steyaert deserves recognition for putting forward a new, thought-provoking hypothesis on the original arrangement of the panels (Griet Steyaert, “The Ghent Altarpiece: new thoughts on its original display,” The Burlington Magazine, February 2015, pp. 74-84.). Her article presents a stimulating and courageous approach to this intensely debated subject. It is thus unfortunate that her proposed solution should come along with a couple of problems. These deserve to be discussed, all the more so as it was none other than The Burlington Magazine, likely the most renowned journal in the field of art historic research, that published Steyaert’s article. My essay is a brief reply in the sense that I will not bring in any concrete art historical facts, let alone plunge into questions regarding whether the proposed reconstruction is technically consistent with the physical evidence left by the altar’s history. (This is not only beyond my expertise, but moreover the Ghent Altarpiece is currently being restored and analysis during this process will most likely yield new technical evidence.) I believe that a few logical and aesthetic considerations are sufficient to cast doubt on the likelihood of the suggested reconstruction.
Steyaert points out several “incongruities” (p. 75) in the altarpiece that led to her doubts about its current presentation (p. 75). First, the upper panels of the wings have rounded tops whereas the upper panels at the center of the retable have rectangular tops. This causes the spandrels of the central panels to peak out when the retable is shut, leaving them uncovered. Second, between the upper and the lower tier of the retable some of the vertical moldings of the frames are not continuous (neither opened nor closed). Third, while the lower tier of the opened retable shows one continuous landscape, the upper tier is disruptively split up into several disconnected spaces. Furthermore, the figures occupying these various spaces are of differing scales. In the center, in a room with brownish tiles we see God the Father, quite tall, with the Virgin and St. John at his left and right; continuing outward in both directions, in a different room with blue and white tiles, we find two groups of much smaller angels. These are framed, in turn, by Adam and Eve depicted in yet another space, stone niches, and at a different scale, medium in relation to the others. In addition, the figures in the upper tier are much larger than the figures in the lower tier.
These ostensible aesthetic inconsistencies cannot be denied. Indeed, other authors have drawn attention to them before. Erwin Panofsky, for example, suggests that the Ghent Altarpiece was assembled from various panel paintings – most of them uncompleted – found in Hubert van Eyck’s workshop after his death in 1426 (Erwin Panofsky, “Die altniederländische Malerei. Ihr Ursprung und Wesen”, Cologne 2001, Vol. I, p. 219.).
Steyaert offers a new solution to these incongruities, suggesting that the Ghent Altarpiece was originally a “three-dimensional, tower-like construction” (p. 74). Knowledge of this structure and the structure itself Steyaert presumes to have been lost in the artwork’s eventful history.
She explains that the differing scale between the figures of the upper and lower tiers in the opened retable was her biggest concern, and what brought her attention to tower retables in the first place (p. 77). Taking a closer look at these objects and their depictions in paintings and illuminations from the period, she discovered examples of a similar phenomenon: “combinations of large sculpture enclosed by painted wings and placed on top of an altarpiece with small figures” (p. 77). “Could it be” Steyaert wonders, “that the panels of the upper tier were originally arranged around a plinth to form a tower retable, not with the usual sculpture(s) but with monumental painted trompe l’œil representations of sculptures?” (p. 78) Her reconstruction of the Ghent Altarpiece is as follows: The lower tier remains basically the same, while the upper tier is folded around a hexagonal plinth. In contrast to typical tower retables, there is no sculpture on the plinth. This is made up for by the central Deësis whose three large figures serve as fake-sculptures, giving the appearance of actual objects on the plinth.
This is an interesting possibility to take into consideration. Yet it seems that most of the incongruities the author is concerned with are not solved by the suggested arrangement.
Steyaert’s model does not supply a solution for the lack of vertical continuity between the frames. In her construction the vertical frame moldings of the upper and the lower tier are still not aligned – neither when the retable is open nor when it is closed. Moreover, the upper and lower tiers’ outer-most frames would no longer form one axis, which is the case at present and itself quite a convincing argument for the retable’s current presentation.
The structure of the retable may lack vertical continuity, yet one of its most impressive features is the spacious horizontal unity of the Annunciation taking place in a room that stretches over four panels. Steyaert’s reconstruction would have these panels folded around three sides of a hexagonal plinth. Thus the front of the closed retable would only offer a view into the empty part of the room. The essential panels, with Gabriel and the Virgin, would seem partly withdrawn, creating a sense of restraint that would be at odds with the artistic confidence that seems to speak through this revolutionary composition.
Overall, the resulting incoherence in perspective caused by the distorted foreshortening of the floor tiles, along with the awkward kinks that would be inflicted on the back wall and the ceiling, illustrate just how Steyaert’s arrangement would turn the generous conception of this scene into what really could seem like odds and ends from Hubert’s workshop.
It seems clear to me that the suggested reconstruction contains even more instability and incongruity then the altar’s present configuration. More vertical elements are put in disorder and an artistically stunning and spacious horizontal unity is obscured.
Another incongruity that the new model does not in fact resolve is the differing scale of the figures. “It could be argued that the most disturbing element in viewing the Ghent Altarpiece as an integral composition is the scale of the large figures of the top tier [...]” (p.77). Since in tower retables large-scale figures are usually placed above small-scale figures, Steyaert believes that in her reconstruction the problem resolves itself.
But if she set out to resolve the incongruity itself (and she claims to have done so on p. 84), she has failed. The argumentation is flawed: Folding the upper tier of the Ghent Altarpiece around a plinth will not change the fact that figures in this tier are bigger than the others. The impression that the upper figures are “crushing” (p. 77) the lower ones will remain unless a forced shift in perspective were created by placing the upper tier at a distance of several dozen meters from the lower, which of course is not a sincere suggestion.
In fact, perhaps the problem doesn’t exist at all. Certain 15th-century artworks in which larger figures were positioned on top of smaller ones – mentioned by Steyaert herself – suggest that this “disturbing element” may only appear as such, if at all, to a 20th or 21st century observer.
In short, the conclusion to draw from the evidence of tower retables shouldn’t be: The Ghent Altarpiece will no longer be aesthetically irritating as soon as we turn it into a tower retable. – But: Maybe we should not regard the differing scales in the Ghent Altarpiece as aesthetically irritating in the first place.
The idea of adding a plinth to the retable also seems problematic. Steyaert gives several examples of tower retables featuring plinths on which a carved figure was placed. In her reconstruction of the Ghent Altarpiece, however, the plinth would be empty with the three figures of the Deësis serving as “trompe l’œil representations of sculptures” (p. 78).
First, none of Steyaert’s examples have an empty plinth, which makes the existence of such an object unlikely – although one could argue that the Ghent Altarpiece is an unlikely object in the first place.
Second, talking about representations of sculptures with reference to the Deësis is incompatible with what is actually depicted in the painting. In order to see what representations of sculptures in van Eyck’s case look like, it is enough to shut the retable and glance at the lower tier where John the Baptist and John the Evangelist are rendered as stone figures.
It seems unlikely that the figures of the Deësis were even intended, as Steyaert later claims, to remind the beholder of enamel work with the faces and hands of living beings (pp. 82-83). The van Eycks (or just Jan; we cannot be sure) were masters at simulating the visual and haptic impressions of diverse materials. If they had wanted to create the illusion of enamel, they probably would have painted the figures as actual enamel works – that is, with the appearance of shiny surfaces, in imitation of the material in question. It becomes evident, however, when looking at the paintings, that the figures are “alive” and dressed in fabrics and fur.
Third, speaking of trompe l’œil representations of sculptures with reference to the Deësis is even more problematic. A trompe l’œil painting deceives the eyes in the sense that it is not clear to the beholder whether he is looking at a painting or an actual object. An example of Eyckian trompe l’œil painting can again be found on the outer panels of the Ghent Altarpiece where the books of the prophets seem to protrude over the frame, thus obscuring the border between painting and reality. This phenomenon is however not present when glancing at the Deësis. Although diverse materials and objects are rendered in a highly illusionistic manner, making them appear almost real, the beholder is at no point fooled into believing that he is confronted with real objects or figures.
Fourth, Steyaert’s theory is problematic where it combines illusionistic painting with the three-dimensional plinth: “I believe that although the plinth was actually empty, the extremely convincing trompe l’œil paintings made it look as if it were not and as if the crown at the feet of God occupied the space between and in front of the three holy figures.” (p. 84)
It seems to me that the contrary is the case. The contrast between the three-dimensional plinth and the painting is likely to actually highlight that the crown is nothing but a two-dimensional painting. Furthermore, it’s not clear that the three figures of the Deësis are in need of this sort of stage, since within the painting they already are situated on solid ground, creating the illusion of a three-dimensional space. The addition of an actual three-dimensional platform ruins the illusionism so successfully created in the painting itself.
That an “onlooker of average height [...] would have been able to see the crown at God’s feet without obstruction,” as Steyaert states on p. 83, is only the case in her schematic rendering when the onlooker is positioned accordingly. A beholder who approached the painting to admire the smaller figures of the lower tier would soon find the upper tier hidden behind the plinth.
To what extent the plinth, as Steyaert claims, would make the recession of the landscape depicted in the lower tier “more convincing” (and why improvement of the artist’s work should motivate art historical theories) is difficult to understand.
Rearranging the Ghent Altarpiece according to the structure of a tower retable would neither solve the problem of missing vertical unity nor that of the figures’ differing scales. Indeed, as it turns out, tower retables give us historical evidence that casts doubt on whether the latter is a problem at all. Furthermore, the new model would split up what little (but impressive!) unity the retable offers at present and would spoil certain illusionistic effects. I believe that the aesthetic and logical considerations outlined here speak against the proposed hypothesis. Steyaert’s bringing tower retables into the debate may have the merit of causing a basic reconsideration of the entire discussion.