3. Mimetophobia
One very general criticism of Vermeer's Camera and indeed of other authors who have argued for the painter's extensive use of the camera obscura, made insistently by several scholars, is that it is a naivety to imagine that the artist might have reproduced, in faithful detail, the exact appearance of the interior and furnishings of some actual house in Delft.21 This point is made, less as a reasoned argument, than as a dogmatic a priori assertion: "It is inconceivable that Vermeer would have worked like that; therefore he did not work like that." What we seem to have here is a serious case of the art-historical condition that Michael Kubovy has termed mimetophobia: 'the morbid fear of slavish imitation'.22 All sorts of arguments are put to support this contention. "Delft houses of the period did not have marble tiles on the floor." "The ceiling joists in 17th century Dutch houses did not run in the direction that Vermeer shows them." "The pictures contain perspective impossibilities, geometrical anomalies, or inconsistencies of lighting which could not have occurred in reality." And so on. This position might be strongly, even viscerally held. However, the balance of evidence in the case of Vermeer is against it.
It must be acknowledged that the question of the specific location of the room depicted by Vermeer in the ten paintings remains unanswered, and is perhaps unanswerable. Maria Thins's house and the Vermeer family inn 'Mechelen' - the two buildings in which it has been suggested that the artist might have worked - were both demolished in the 19th century , and there seem to be no drawings of their north-facing facades in which one might recognise the very windows of Vermeer's studio. Perhaps he worked elsewhere, in some building yet to be recognised in Delft, but it seems unlikely. Nor can we say, obviously, whether Vermeer's portraits are good likenesses. There are nevertheless three other kinds of evidence on which to assess the extent of Vermeer' s naturalism and fidelity to appearances.
The first is his rendering of the real items - maps, pictures, chairs - to be found today in libraries and museums. The second is the degree to which Vermeer's representation of other pieces of furniture, and architectural details, are typical of houses and styles of interior decoration of the period, even if we cannot identify the originals themselves. The third is the internal consistency with which Vermeer depicts both objects and effects of light. As an example of this last point, I show in the book how in some paintings by Pieter de Hooch, the regular grid of floor tiles does not meet up correctly behind obstructions such as pieces of furniture or standing figures. In Vermeer this never happens. 23 Vermeer's grids always join up perfectly between the various visible parts of the floor. This fact suggests either highly accurate preparatory perspective layouts, or very careful study of real floors - as I would argue, through the camera lens - rather than ad hoc perspective lines sketched during the process of painting. Indeed in most other respects the geometry of Vermeer' s perspectives is extremely precise and consistent.
Now I would certainly not want to claim that Vermeer is always and everywhere a slavish copyist of what was in front of him. This is evidently not the case, as I emphasise at several points in Vermeer's Camera.24 Some of the most obvious - although still relatively minor - departures from actual appearances are in Vermeer's copies of paintings by other artists, as for example the version of Jacob Jordaens 's 'Crucifixion' in 'Allegory of the Faith', where Vermeer has removed two figures, or the version of Dirck van Baburen's 'The Procuress' in 'Lady Seated at the Virginals', where van Baburen's picture is not shown to scale, and has been stretched vertically. (The version in Vermeer's 'The Concert' on the other hand is reasonably faithful.) Critics of my general thesis who leap triumphantly on one supposed instance or another in which Vermeer has departed from realistic appearances, as something which will bring the whole edifice of my argument tumbling down, are therefore missing the point. It is, rather, a question of the balance of facts in support of or against Vermeer's naturalism, the accumulated weight of evidence on either side. This weight, I contend, is overwhelmingly on the side of his naturalism.
Figure 6: Lions' head chair in the collection of the Prinsenhof Museum, Delft (left) compared with Vermeer's painted version in 'The Glass of Wine' (right).
Figure 7: Tapestry-covered chair in the collection of the Prinsenhof Museum, Delft (left) compared with Vermeer's painted version in 'The Concert' (right). Many of the details are laid out in the book, and there is not the space, or perhaps the need, to review them all again here. To select just a few instances: the chairs with sculpted lions' heads and the tapestry-covered chairs that appear in several pictures are identical to chairs of these designs in the collections of the Rijksmuseum in Amsterdam and the Prinsenhof Museum in Delft, down to the exact profiles of the turned legs and the shapes of the cut-out cross-members (Figures 6 and 7). The same is true of the very characteristic turned legs of Vermeer's table (Figure 8). As for the printed paper maps, Vermeer's versions are so accurate, down to every last cartouche and little sailing ship on the ocean - as James Welu has shown - that in several cases it has been possible to identify not just the map's designers and publishers, but the specific editions (Figures 9 and 10). 25 Not only can the book studied by 'The Astronomer' be recognised; the very page at which it is open is legible.26
Figure 8: 17th century table in the Rijksmuseum, Amsterdam (left) compared with Vermeer's painted version of a table with identical details in 'Woman Holding a Balance' (right).
Figure 9: Map of Europe by Joan Blaeu, 1659 (left) compared with Vermeer's painted version in 'Woman with a Lute' (right). The map is cropped at top and right in both cases.
Figure 10: Map of the Seventeen Provinces of the Netherlands by Claes Jansz Visscher (left) compared with Vermeer's painted version in 'Allegory of Painting' (right). (The ornamental borders showing views of Dutch towns are omitted in both cases.) What is more, as already mentioned, it is not just the geographical detail of the maps and the decorative detail of the chairs that is so precisely represented; their overall dimensions also correspond closely to the originals.27 The perspective reconstructions of the two types of chair give overall heights, and heights to the seats, that are within one or two centimetres of the museum originals. The reconstructed heights and widths of Vermeer's maps are, in all those cases where measurement is possible, within a few per cent of the surviving library copies.
Turning to the architecture of the room itself, the calculated dimensions - again as already described - are similar in all ten paintings. There are anomalies, certainly, as spelled out in the book.28 In two pictures, 'Girl with a Wineglass' and 'The Glass of Wine', the sizes of the half-open casements are smaller than in other paintings. (I argue that Vermeer has deliberately adjusted them for compositional reasons, since they are very close to the paintings' viewpoints and would, if unmodified, appear disproportionately large. One correspondent, Roger Hawkins, has made the further point that, had Vermeer indeed been using a camera obscura, the casements would have been close to the lens, hence difficult to bring into focus.29) In 'Girl with a Wineglass', where one would, by comparison with other pictures, expect a second further window, there is none visible. JØrgen Wadum has made much of this discrepancy.30 However it is just possible to make out, in both this picture and in 'Girl Interrupted at her Music', the pale vertical stripe which would correspond to the plastered brick pier between the visible window and the place where the second window would be. That is to say, the ghost of the missing window still haunts the dark corner of the room. In 'Girl Interrupted at her Music' it is possible that Vermeer may originally have included a shuttered window that has been painted over in the heavy restoration to which this picture has been subject.
In other details, there are variations in certain dimensions obtained from different paintings, largely because the surfaces in question are seen at very oblique angles and hence are more difficult to measure reliably. But the overall conclusion that any reasonable-minded observer would reach must be that this is the same room throughout. The main obstacle to this idea, on first sight, is the varying patterns of marble and ceramic floor tiles.
Walter Liedtke and others - citing a recent paper by C. Willemijn Fock - have objected that marble tiles would not have been found in the relatively modest type of Delft house in which Vermeer would have lived and painted.31 Fock's paper is in effect a riposte to Peter Thornton' s book Seventeenth-Century Interior Decoration in England, France and Holland, in which Thornton says that the Dutch particularly favoured floors in chequerboard patterns, made either of quarry tiles or of imported marble.32 Thornton' s argument was based however just on the evidence of contemporary genre paintings. Fock points out, quite properly, that a rather different picture is given by probate inventories, building specifications, plans, contemporary descriptions and archaeological findings. Specifically, she shows that marble floors, although they did exist in Holland at this period, were rare in burgher houses, and even there tended to be found in halls and corridors, rather than reception and living rooms. Painters, Fock says, were in the habit of introducing fictive marble tiles into interiors to emphasise perspective effects.
As it happens, this is close to what I suggest in Vermeer's Camera, where I do not argue for the real tiles of the room being marble. On the contrary, I hint that Vermeer's marble may indeed be imaginary.33 What I do show is that beneath all the different arrangements of black and white marble, and the smaller ceramic tiles, is an underlying grid that is identical throughout (Figure 11). The ceramic tiles have exactly half the edge-length of the marble tiles. It is in this grid that the dimensional consistency lies. I suggest that Vermeer may have traced the grid always from the ceramic tiles, which could have been the floor's real covering, and that he reproduced the actual appearances of these tiles in two or three early paintings; but that later on he simplified his task, and introduced variations in design, by substituting the larger fictive marble tiles (each occupying four grid squares). Naturally, once the grid was drawn, it would have been a simple matter to vary the pattern of blacks and whites ad lib.
Figure 11: The underlying grid which is common to all the patterns of floor tiles in Vermeer's interiors. The smaller ceramic tiles in 'The Glass of Wine' and 'The Girl with a Wineglass' are shown in (d). Each of the larger 'marble tiles', in other paintings, coincides with four of the ceramic tiles: as for example in (a) 'The Music Lesson', (b) 'Lady Standing at the Virginals' and (c) 'Lady Writing a Letter, with Her Maid'.
One reviewer, Ceri Shields, has drawn attention to the striking difference between Vermeer's treatment of the two types of tile, a difference which gives support to this hypothesis.34 The ceramic tiles are rendered very naturalistically, showing all the unevenness of the glaze and the scuffing and chipping of the tiles' edges; while the 'marble' tiles are quite schematic, mechanical, with the veining more suggestive of spilled liquid than polished stone.
A second criticism of a similar nature is that Vermeer shows ceiling beams, in three paintings, running across the picture from left to right; where in typical Delft houses of the period - it is claimed - they would have run in the perpendicular direction, away from the viewer and towards the far wall.35 The logic behind this argument is that these were terraced houses, with long brick party walls, and shorter walls containing the windows at the ends of the plan. It is one of these window walls that is always at the left in Vermeer's ten paintings. (When the windows themselves are not visible, the light still comes from this direction.) The beams, it is argued, would span always between the party walls. The argument is very plausible in relation to many Delft houses, which were narrow in plan, around 4 or 5 metres wide. Both 'Mechelen' and Maria Thins's house however are less usual in being closer to 7 metres wide. The reconstructed room is some 6.5 metres long, just about the right dimension - allowing for the thicknesses of the walls - to fit across the complete width of either building. Here, for good structural reasons, a different form of construction could be expected.
My colleague Marc van Leusen has consulted Delft architects and Dutch specialists in the history and restoration of old buildings on this question.36 Willem Weve, an architectural historian working for the Delft municipality, says that domestic construction was not in fact standardised in the city in the 17th century; but that the type of ceiling shown by Vermeer is one among several arrangements used in houses, and surviving examples can indeed be found. The timber members are indeed small beams, probably of pine, supported by a wall plate over the windows, as seen at top left in 'The Music Lesson'. They would have been relatively deep, so the floorboards which they support are not visible. It is likely, according to Weve, that the beams were supported at their other ends on a wall which would be on the right of Vermeer's pictures, but is always out of sight. Wilfried van Winden is a partner in the Delft architectural practice Molenaar and van Winden which specialises in restoration projects in the city. Van Winden' s own house in the centre of Delft, dating from the 17th century, has ceiling beams as Vermeer shows them.37
Wadum makes the argument that Delft houses of the period generally, and Maria Thins's house in particular, would not have been exactly rectangular in their geometry, because of the exigencies of their sites and inaccuracies in construction; and that had Vermeer been conscientiously tracing camera images of his actual room, this lack of rectangularity would be detectable in the paintings.38 According to Wadum it is not. Close examination of the 1830 map would indicate that the plan of the front part of Maria Thins's house (see Figure 4) - if it is indeed correctly identified here - was very close to a true rectangle, much closer than many of its neighbours. Nevertheless, when some of the relevant paintings are measured carefully, it is possible to find minor departures in the architecture from precise rectangularity.
If the ceiling beams visible in three pictures ('The Music Lesson', 'Allegory of Painting' and' Allegory of the Faith') are studied closely, they can all be seen to slope downwards somewhat, from left to right. This could be a result, either of the wall being seen at a slight angle, or - perhaps more likely - of the beams sagging between their points of support at left and right.39 If such an effect was observed in just one of the paintings, it might be put down to the vagaries of Vermeer's drawing or methods of measurement. The fact that it occurs in three cases is at least suggestive of the possibility that this is a real geometrical property of the room in question. Certainly, it would be unlikely to occur had the compositions been set up using standard geometrical perspective construction procedures.
