Jan van Eyck
NG 186 
Portrait of Giovanni(?) Arnolfini and his Wife

Lorne Campbell, 1998, updated September 2021

Extracted from:
Lorne Campbell, The Fifteenth Century Netherlandish Schools (London: National Gallery Publications and Yale University Press, 1998).

Oil on oak panel, 84.5 × 62.5 cm,
painted surface 82.2 × 60 cm

Signature

Johannes de eyck fuit hic / .1434.
(Jan van Eyck has been here. 1434.)

Provenance

The first recorded owner was Don Diego de Guevara, a Spanish nobleman who was brought up at the Burgundian court, spent most of his life in the Low Countries and died in Brussels in 1520.¹ His coat of arms and device² were on the wing panels, mentioned in 1516, 1523–4, 1556–8 and 1700 but subsequently lost. He gave the picture to Margaret of Austria (1480–1530) and it was the first item listed in the inventory of her paintings taken, in her presence, in her palace at Mechlin on 17 July 1516: ‘a large picture which is called Hernoul le Fin with his wife in a chamber, which was given to Madame by Don Diego, whose arms are on the cover of the said picture; done by the painter Johannes.’ In the margin, it was noted ‘it is necessary to put on a lock to close it: which Madame has ordered to be done’.³ In a later inventory of Margaret’s possessions, taken at Mechlin between 9 July 1523 and 17 April 1524, it was described as ‘another very exquisite picture, which closes with two shutters, where there are painted a man and a woman, standing, touching hands, done by the hand of Johannes, the arms and device of the late Don Diego on the said two shutters, the name of the personage being Arnoult Fin’.⁴ It was then in the seconde chambre à chemynée of her Cabinets.

It passed with many of Margaret’s pictures into the collection of her niece Mary of Hungary (1505–58). She left the Low Countries in 1556 and went to Spain, where she died in 1558. In an inventory of her effects taken shortly after her death were listed the paintings that she had brought from the Low Countries, including ‘a large panel, with two doors with which it closes, and in it a man and a woman who take each other’s hands, with a mirror in which the said man and woman are shown, and on the doors the arms of Don Diego de Guevara; done by Juanes de Hec, in the year 1434’.⁵ Mary’s paintings passed to her nephew Philip II (1527–98), King of Spain. In the summer of 1599, Jakob Quelviz, who came from Leipzig, visited the Alcázar of Madrid and saw there in the Salle chiqua: ‘an image where a young man and young woman are joining hands as if they are promising future marriage; there is much writing and also this:
Promissas fallito (sic) quid enim promittere laedit
Pollicitis diues quilibet esse potest.’⁶ The inscription mistranscribed by Quelviz is from Ovid’s Ars amatoria, Book I, lines 443–4:
promittas facito, quid enim promittere laedit?
pollicitis diues quilibet esse potest(See that you promise: what harm is there in promises? In promises anyone can be rich.) Though Quelviz did not remark upon the mirror or the signature, he must certainly have been referring to NG 186; the inscription from Ovid was mentioned in the next description known, in the inventory taken after the death of Charles II (1661–1700), King of Spain. The portrait was still in the palace at Madrid and was listed as ‘a picture on panel with two doors that close with its wooden frame gilded with unburnished gold, some verses from Ovid written on the frame of the picture, which is of a pregnant German woman (una Alemana preñada) dressed in green giving her hand to a youth (mozo) and it appears that they are getting married by night and the verses declare how they are deceiving each other and the doors are of wood painted with marbling, valued at 16 doubloons’.⁷ In 1794, it was in a Retrete in the Cuarto del Rey in the Palacio Nuevo at Madrid and was described as ‘one vara high by three quarters of a vara wide, a man and a woman holding hands, Juan de Encinas inventor of oil painting, 6000 reals’.⁸ A vara is a Spanish yard of 84 cm. The picture is presumed to have left Spain during the Peninsular War.

The next recorded owner was James Hay (died 1854), a Scottish soldier who served with the 16th Light Dragoons in the Peninsular War and was present at the Battle of Vitoria in 1813.⁹ The Duke of Wellington recalled in 1814: ‘The baggage of King Joseph after the battle of Vitoria fell into my hands, after having been plundered by the soldiers; and I found among it an Imperial containing prints, drawings and pictures.’ The Imperial was part of Joseph Bonaparte’s coach and was found to contain many paintings from the Spanish royal collection.¹⁰ It is possible that James Hay took NG 186 from Joseph Bonaparte’s baggage at Vitoria; or, at any rate, that he obtained it when he was serving in Spain. According to the dealer Chrétien‐Jean Nieuwenhuys (1799–1883), writing in 1843, Hay, ‘having been dangerously wounded at Waterloo, was moved to a private house in Brussels and lodged in a room decorated with this picture. During his long convalescence, it attracted his attention several times; he admired it so much that, in the end, he acquired it. He kept it until 1842 …’.¹¹ Nieuwenhuys did not disclose how he learned this romantic story but perhaps he had it from Hay himself, who may have tried to sell the picture to Nieuwenhuys and who would have been anxious to impress upon him that he had acquired it by honest means. Hay was certainly at Waterloo (18 June 1815) and was so desperately wounded that he could not be moved from the battlefield for eight days. He was then taken to Brussels to convalesce.¹²

The story that, during his convalescence, he fell in love with the painting does not entirely square with the fact that, immediately after his return to England, he attempted to sell it to the Prince Regent. Sir Thomas Lawrence sent to Carlton House on 10 October 1816 ‘A Painting in a Gilt Frame – Subject Two Portraits Male & Female Joining Hands – the Female Dressed in Green – The Male in Black with a large Round Hat on &c – 33½ Inches by 23¾ by John Van Heyk this Picture is sent for The Regent’s inspection – it was Painted by John Van Heyk – the person who first discovered the art of Painting in Oil Colours – Returnd –’. Included in the inventory taken at Carlton House in December 1816 as ‘Portraits of a man and his wife. John van Eyck’ (measurements given as 2 ft 8 in by 1 ft 11 in), it was in ‘The Middle Attic next the Prince’s Bed‐Room’. A note in the same inventory recorded that ‘This Picture was returned to Sir T. Lawrence April 25th 1818’.¹³ Hay later entrusted it to Dr James Wardrop (1782–1869), who lived in Charles II Street. He noted that it ‘was sent to me by Col. James Hay (Queen’s Bays) to keep for him during his absence. It was hung up between two windows in a Bed‐room where it remained, about 13 years, during this period, it was seen by many visitors, none of whom deem’d it worthy of their notice. During the 13 years Col. Hay was absent from London, & I never saw him again until he asked me if it could be convenient for me to send for his picture which he did accordingly. A few weeks afterwards, I saw to my surprise in the British Gallery Exhibition Col. Hay’s picture, & a very short period afterwards Mr Seguier called on me, mentioning that as I was a friend of Col. Hay’s who was then in Ireland would I communicate to him that he (Mr Seguier) had recommended the Trustees of the National Gallery to purchase the Van Eck picture & he was authorised to offer for it six hundred pounds. This sum was accepted by Col. Hay.’¹⁴ James Hay’s will gives no indication that he ever took any further interest in pictures.¹⁵ The ‘British Gallery’ exhibition was the British Institution exhibition of June 1841, where the painting attracted some notice,¹⁶ and on 2 May 1842 ‘the extraordinary merit of the Picture by Van Eyck, its perfect preservation, its extreme rarity and the moderate price at which it may now be obtained viz^t 600 Guineas’ induced the Trustees of the National Gallery ‘to recommend it without hesitation … as a valuable [page 177][page 178]addition to the National Collection’.¹⁷ Though the portrait was in the Gallery by June 1842,¹⁸ Hay had still not been paid in February 1843.¹⁹ The portrait was first exhibited, under glass and in a case, in March 1843 and immediately attracted crowds of visitors.²⁰

Exhibitions

BI 1841 (14); London 1945 (not catalogued); ‘An Exhibition of Cleaned Pictures (1936–1947)’, NG 1947 (22); London 1975 (13); London 1977–8 (not numbered); London 1998 (not catalogued); ‘On Reflection’, NG 1998 (not numbered).

Versions

Though there is a temptation to imagine that any work of art showing a mirror, a chandelier or a couple with a dog must be influenced by NG 186, many fifteenth‐ and sixteenth‐century artists do seem to have had some knowledge, direct or indirect, of the double portrait. In the Heinrich Werl presented by Saint John the Baptist (Madrid, Prado), dated 1438, by a follower of Campin (see p. 404),²¹ the convex mirror reflects the room and the figures of an approaching friar and novice: it is so reminiscent of the Arnolfini mirror that there must surely be some connection between the two paintings. In Petrus Christus’s Holy Family (Kansas City, fig. 2),²² the whole interior, including the bed, the high‐backed chair, the chandelier and the fruit on the window‐sill, appears to have been suggested by the interior in NG 186. A German painter of about 1490, probably working in or near Cologne, produced a double portrait (fig. 3) so strongly recalling NG 186 that, once again, he must have had some knowledge of van Eyck’s painting.²³

The illuminator Loyset Liédet, who moved from Hesdin to Bruges in the 1460s and worked there until 1478, seems to have owned a copy, perhaps a drawing, after NG 186 and to have referred to it constantly when he was preparing designs for the enormous numbers of miniatures executed in his workshop.²⁴ In the five‐volume Renaut de Montauban, which Liédet illuminated for Charles the Bold and for which he was paid in 1468–70, the miniature on fol. 202v of the first volume (Paris, Bibliothèque de l’Arsenal, MS 5072) is at least reminiscent of NG 186.²⁵ In the four‐volume Histoire de Charles Martel, which Liédet illuminated for Charles the Bold and for which he was paid in 1472, at least four miniatures may be classified as versions of NG 186. In the first volume ( BR , MS 6, fol. 423), where Charles Martel returns in disguise to his palace,²⁶ the couple standing by the bed are taken from van Eyck’s portrait. In the third volume ( BR , MS 8, fol. 7: fig. 4), where Charles the Bold(?) visits a scribe,²⁷ the brush, the chandelier, the convex mirror, the oranges, the string of beads and the prominent inscription on the rear wall (here Charles the Bold’s device) are all from NG 186. In the same volume, fol. 108v shows visitors being received at the court of King Pepin;²⁸ the antechamber on the right, where two courtiers are playing chess, comes from the same source. In the last volume (MS 9, fol. 7: fig. 5), where Ludie seeks revenge,²⁹ Liédet has signed his name prominently on the wall above the window: he was clearly inspired to do so by the example [page 179][page 180]of Jan van Eyck’s signature in NG 186. In the Demandes et Responses en Amours (Herzog August‐Bibliothek, Wolfenbüttel, Cod. Guelf. 84.7 Aug. 2°, fol. IIIv: fig. 6),³⁰ the couple and the interior are dependent on NG 186. In the Roman de Gérard de Nevers or Roman de la Violette ( BN , MS fr. 24378), the miniature on fol. 120, Gérard receiving the gift of a falcon (fig. 7),³¹ is taken from NG 186. In a manuscript of the Recueil des histoires de Troie sold at Sotheby’s on 13 July 1977 (60), the miniature of Juno sending Hercules to Egypt³² is once again an adaptation of the composition of NG 186.

The National Gallery picture was accessible in some form to artists who manufactured ‘historical’ portraits. The woman in the London portrait seems to have served as a basis for several images of Bonne of Artois, Philip the Good’s second wife (she died in 1425). There are painted versions in Berlin (fig. 8),³³ Cadiz³⁴ and Lisbon;³⁵ another was in the Haas collection in Detroit;³⁶ a mid‐sixteenth‐century drawn version is in the ‘Recueil d’Arras’ (Arras, Bibliothèque municipale, MS 266, fol. 62).³⁷ All show the lady wearing a slightly simplified version of the headdress seen in NG 186 and a V‐necked rather than a round‐necked dress, which, in the Berlin and Cadiz paintings, is pink rather than green. The woman in NG 186 seems also to have been adapted as a ‘portrait’ of Walburga of Mörs, who married in 1454 Philippe de Croÿ, Seigneur de Sempy. The original appears to have formed part of a series of full‐length portraits of members of the Croÿ family: they are known only from drawings in Rubens’s Costume Book.³⁸ An engraving published in 1600 of Ansegisel, Duke of Brabant, and his wife Saint Begga (died 693) may be partially based on NG 186.³⁹

A double portrait signed ‘Godefridus Ioannis fecit 1581’ was mentioned by Weale as a ‘curous imitation’ of NG 186 but there are in fact no close resemblances.⁴⁰ In contrast, Sánchez Coello’s double portrait (Prado) of the Infantas Isabella and Catherine, the daughters of Philip II of Spain, painted in about 1575 (fig. 9), is clearly based on van Eyck’s painting, then in Philip’s collection.⁴¹

It has been suggested that Velázquez saw NG 186 when it was in the Spanish royal collection and that it inspired him to paint the mirror in Las Meninas (Prado).⁴² Since van Eyck’s picture has been on exhibition at the National Gallery, it has attracted the attention of a great many artists and it is impossible to list all the works of art which it has inspired. Holman Hunt’s drawing of the Lady of Shalott (Melbourne), [page 181]done in or shortly after 1850, and Ford Madox Brown’s Take your Son, Sir (London, Tate Gallery), painted in the 1850s, are among the most interesting.⁴³

Engravings

The earliest prints of the portrait were published in the Illustrated London News of 15 April 1843, p. 258, and in Felix Summerly’s Handbook for the National Gallery, London 1843, p. 49.

Technical Notes

The portrait, cleaned in 1942–3, is very well preserved. There are small losses along a zig‐zag scratch across the mirror and minor flake‐losses in the side and back walls and in the bedcover; there is a slight damage in the ribcage of the dog. Some of the ultramarine‐containing paint in the woman’s sleeve and underdress, in the man’s tabard and in their reflections in the mirror may have deteriorated and become greyish. Many of the retouchings have discoloured and become a little obtrusive. There are two layers of varnish: the lower is somewhat physically degraded and is yellower than the upper layer.

The panel, which measures 84.5 × 62.5 cm, is made up of three oak boards, laid vertically, vertical in grain and radially cut. They are similar in size but not quite regular in shape, the widths at the top edge being, from left to right, 20.95, 20.4 and 20.9 cm, at the bottom edge 20.8, 22.0 and 19.7 cm. Each of the two joins is reinforced with three dowels. At its centre the panel is about 17 mm thick; the edges are bevelled at the back. Though the bevels are not completely regular, the thickness at the middle of the top edge is 11 mm. An old vertical split in the central board was about 26 cm long and extended from the top edge to the mirror frame. It is held at the back with four wooden buttons, placed there in 1940. The unpainted edges, regular in width, survive on all four sides. Slivers of oak, horizontal in grain, adhere to the upper and lower unpainted edges and must be fragments of the original frame. The largest fragment begins below the dog’s left forepaw and extends about 7 cm to our right.

The reverse (fig. 10) has been coated with various protective layers. All were applied when the panel was in its original frame, so that the reverse, like the obverse, has unpainted edges, which vary in width between 1.2 and 1.4 cm. The reverse side of the original frame must have been the same size as the obverse side. The lowest protective layer [page 182]consists of two distinct and thick applications of true gesso (that is, calcium sulphate), in which fibres of vegetable origin are embedded. Above this are a light brown layer, which is X‐ray opaque and which contains earth pigments, black and a lead compound, presumably lead white, and, over this, a layer of black or very dark brown paint containing verdigris as a drying agent. Since the gesso was presumably applied after the picture had arrived in Spain, none of these layers is original. No trace of any original ground or paint has been discovered but it is, of course, possible, though it seems unlikely,⁴⁴ that the reverse was decorated by Jan van Eyck and that his decoration was removed in Spain before the gesso was applied. The various protective layers are much damaged. On the reverse is written in chalk the number 10(?)3.

Because an X‐ray‐opaque layer is present on the reverse, it is impossible to make informative X‐radiographs (fig. 11). On the obverse, there is a chalk ground. The underdrawing, clearly visible in infra‐red reflectograms (figs 1, 12, 13, 15),⁴⁵ is executed in a liquid medium and applied with brushes. No priming has been detected. There is nothing exceptional about the pigments or media used. Finely ground natural ultramarine is present, for example in the woman’s blue sleeves and underdress, in the man’s tabard and, with black, over(?) the red glazes on the hangings of the bed. The underlayers for the bed‐hangings consist of combinations of vermilion, red lake and a red‐brown earth. In a sample taken from the tester of the bed, the medium has been identified as linseed oil. In a sample from the woman’s green dress, the surface verdigris glaze is in a medium of heat‐bodied linseed oil to which some pine resin has been added. Two opaque underlayers are present, containing verdigris, lead‐tin yellow and lead white.

In the underdrawing, the figures, the bed and the chest are all outlined. Van Eyck has used complex systems of hatched brushstrokes to indicate areas of shadow. In flat areas, such as the back wall and the floor between the figures, the strokes are roughly parallel and tend to slant upwards from left to right. In other areas, such as the bolster and the bedspread, the directions of the brushstrokes vary to give a suggestion of volume, while the intensity of overlapping lines indicates the depth of the shadow. A similar system, but on a much finer scale, is used in the underdrawings for the heads (figs 12, 13). The drawing in the green drapery is simpler and employs a more formulaic system of fold lines and short hatching strokes, though in the outline of the hem there are bolder, seemingly coarser, lines than are seen elsewhere in the underdrawing.

No underdrawing can be found for the oranges, the beads, the two pairs of pattens, the dog or the chandelier, which is painted on top of the hatched underdrawing and the paint of the rear wall. Numerous changes occur in both figures. The man’s hat was drawn a little higher and the shape was altered slightly during the course of painting, when the brim was enlarged over the wall and the crown was extended sideways. His eyes were drawn higher than they were painted and [page 183]looked more to his left. His nose and mouth were drawn above the painted features (fig. 12). The woman’s eyes were drawn lower and her gaze has shifted towards the man (fig. 13). The very fine, delicate hatching which indicates the shadows in her face corresponds to the drawn eyes; the eyelids and sockets of the painted eyes are not drawn. The man’s head seems to have been changed at an earlier stage, as the new positions of all the features are drawn and the hatching gives form to the second head. His right arm and hand have been altered, the hand having been further to our left and its palm exposed. Both his shoulders have been moved down. On his left, there are two drawn contours above the painted contour, which is also drawn. Three distinct versions are visible of each of his feet: a drawn outline; a painted foot that was painted out; and the final painted foot. The drawn feet are further apart than the painted feet. The skirt of his tabard was shorter. The woman’s hat was drawn lower; her left hand was angled downwards; the thumb of her right hand was drawn parallel to his thumb; and his fingers curled further round her hand.

The drawing shows a larger, octagonal mirror (fig. 15). The bench was higher and red paint from its back lies beneath the grey wall under the mirror, while the high‐backed chair was painted over the wall and was not underdrawn. A brush, larger than the painted one and hanging parallel to the wall, was drawn higher and nearer the mirror. The carpet was drawn shorter and has been extended across the underdrawing of the floor to meet the hem of the green dress. The chest had a more ornate base in the underdrawing. The crossbar of the window seems to have been drawn lower and the shutters were drawn to match its lower position. The top of the window may have been visible in the underdrawing.

The portrait is relatively quickly and freely painted. Van Eyck has blotted the green glazes on the woman’s dress, perhaps with his fingers; in order to soften the tonal transition between the floor and the shadow of the dog’s hind leg, he has smudged the contour of the shadow with a finger or thumb, which has left a recognisable print. A sgraffito technique has been used to define the handle of the brush and the highlights on the little twigs(?) which form the brush itself (fig. 16). The top of the sconce on the extreme right of the chandelier consists of a dark brown contour line and a highlight; the middle‐tone is the paint of the wall behind. In many places, for instance the fringes of the tester, the paint is worked wet‐in‐wet. Occasionally, for example in the stained glass of the window, van Eyck has painted at such speed that his brush has strayed or trailed across contours. He appears to have dropped a small brush, with blue paint on it, onto the mirror‐frame. The signature was evidently painted with a long narrow brush, the point of which had a slight tendency to divide. Some of the cadels are made up of several brushstrokes. The speed and economy of van Eyck’s technique are perhaps best admired in the amber beads (fig. 14) or in the reflection of [page [185]][page [186]]the chandelier in the mirror. His immense skill is everywhere apparent and the complexity of which he was capable, when he chose, is most clearly observed in the chandelier itself.

Description

The couple are in a brick‐built house (the bricks are visible through the open window); it is summer, for the cherry tree outside the window is in fruit. Since only the top of the tree is visible, the room must be above ground level. Its two windows and doorway are reflected in the mirror; the doorway is directly in front of the couple, who are standing between the two windows in the centre of the room. The upper parts of the visible window are glazed with clear bull’s‐eye glass and blue, red and green stained glass but the lower parts are not glazed: this was usual, even in palaces. A railing, which is reddish‐brown and appears to be wooden, crosses the lower part of the opening and is secured into the polished wooden casement. No comparable railing has been found. Wooden lattices were more common.⁴⁶ The window has six interior shutters, which are wooden. They are attached to the wall with metal hinges and, on their outer faces, they are studded with metal. The shutters on our left are rebated on our right and so can be closed under the shutters on our right. The upper edge of the lower shutter on our left and the lower edge of the upper shutter are rebated, so that the middle shutter can be secured across them to the metal catch in the mullion, which can therefore hold all six shutters in place.

The ceiling and floor are wooden. Bare wooden boards such as these are not often found in representations of interiors, where floors are usually tiled or inlaid with semi‐precious stones. In two roughly contemporary pictures, however, the Virgin is enthroned in rooms with floors of bare boards; around 1434 there may have been a brief vogue for wooden floors.⁴⁷ The ceiling is supported by at least one crossbeam, which is visible in the mirror and runs, above the heads of the couple, across the centre of the room. The interior walls are evidently plastered and painted grey. A large bed, apparently standing on a dais, takes up one corner of the room and is covered and hung with red woollen cloth – perhaps ‘scarlet’ and therefore expensive.⁴⁸ The rods supporting the tester and curtains are presumably iron and must be suspended from the ceiling.⁴⁹ The tester has a decorative fringe and the curtain behind the woman has been neatly gathered up into a bundle which, as the mirror reveals, hangs from the corner of the tester. One curtain was generally looped up in this way in order to display the bed, for elaborately draped beds were among the most costly and prestigious objects that anyone could own. The high‐backed chair may be an integral part of the bed: a chair which is clearly attached to the headboard of the bed is shown in the Kansas City Holy Family by Petrus Christus (fig. 2)⁵⁰ and in other paintings.⁵¹ On top of the elaborately carved chair‐back is a figure of Saint Margaret, praying and standing behind, or emerging from, the body of the dragon which is one of her attributes.⁵² On the arm of the same chair is a carved lion. A red cushion rests on the seat. A brush hangs on a purplish string from a nail driven into the wall next to the chair; it is seen in perspective and lies diagonally across [page 187]the chair. The lower part of its handle is dark brown, while the main part of the handle is reddish and bound with a dark grey cord. The rest, which is brown, may consist of fine twigs; they are tied round with a red cord. This brush is similar to one hanging behind the Virgin in a Campinesque Annunciation (Brussels),⁵³ and to various representations of hair‐brushes,⁵⁴ which appear all to be smaller than the brush here; it also resembles brushes depicted in the backgrounds of two sixteenth‐century Netherlandish portraits.⁵⁵ The brush in NG 186 is probably a clothes‐brush (kleerbezem): such brushes are often listed in inventories and in the fifteenth century seem to have been made of brushwood rather than bristle.⁵⁶

Below the mirror is a wooden bench with a footrest. It is draped in red cloth and two red cushions lie on its seat. Only in the most prosperous households was the furniture draped in this way.⁵⁷ On the arm is a carving of a monster with a grimacing human face, (?)lion’s ears and hoofs instead of hands; he wears a hat and a bib and is seated back to back with what is either a carved lion or another, similar monster. Benches were often decorated with carved lions;⁵⁸ the grotesque is perhaps less usual.

Beneath the window is a wooden chest. On the window‐sill and reflected in the polished wooden casement is an orange, its green stalk still attached. Three more oranges lie on the chest and all four are reflected in the mirror. The Spaniard Pero Tafur, visiting Bruges in 1438, ‘saw there oranges and lemons from Castile, which seemed only just to have been gathered from the trees’;⁵⁹ but oranges were very expensive.⁶⁰ The Lucchese merchants in Bruges traded not only in precious silks but also in oranges and other exotic fruits.⁶¹ Leaving four oranges scattered with studied carelessness on a window‐sill and on a chest was surely a way of indicating extreme affluence. In front of the chest and visible only in the mirror is a folding chair of a kind seen in many miniatures of palace interiors.⁶² The carpet, which is unusual in having no fringes, is apparently oriental, though the pattern cannot be related to any specific eastern tradition.⁶³ It is not particularly like the patterns on the carpets in van Eyck’s other paintings, where they decorate the steps of the Virgin’s throne. In manuscripts, carpets are found on the steps of princes’ thrones⁶⁴ or on the floors of palaces.⁶⁵ In fifteenth‐century Netherlandish art, oriental carpets are rarely represented outside Bruges; it seems that they were almost unknown outside Bruges and that, even there, they were uncommon.⁶⁶

It is difficult to find out about contemporary conventions and étiquette etiquette but one invaluable source is Les Honneurs de la Cour, written between 1484 and 1491 by Aliénor de Poitiers, the widowed Viscountess of Veurne. Her mother Isabel de Sousa had been lady‐in‐waiting to Isabella of Portugal, Duchess of Burgundy, and Aliénor had resided with her mother at the Burgundian court.⁶⁷ She was particularly interested in the conventions observed when ladies of various ranks were lying in. She noted that a noblewoman below the rank of countess normally placed only one carpet in front of her bed when she was giving birth; only a lady of the highest rank was permitted to carpet the entire floor of her lying‐in chamber.⁶⁸ Though in NG 186 van Eyck is unlikely to be making any reference to childbirth, the carpet in front of the bed may be an indication of social status.

Hanging from the ceiling is a brass chandelier with six branches: a kroon, lichtkroon or kroonluchter (the word kroon also means crown).⁶⁹ It is suspended from brownish‐grey cords, the knotted ends of which pass through the metal bar near the top edge of the painting. Other representations of comparable brass chandeliers, for example the Louvre Annunciation by a follower of van der Weyden (see p. 27)⁷⁰ and the Kansas City Holy Family by Petrus Christus (fig. 2),⁷¹ show that the cords would have been part of a mechanism for raising and lowering the chandelier and that the mechanism, consisting not only of cords but also of pulleys concealed in a metal box, would have been about as high again as the chandelier itself. It is reflected in the mirror and, if the reflection is to be taken literally, it is much nearer the mirror than the cross‐beam supporting the ceiling above the heads of the couple. The chandelier is therefore between them and the mirror and there is insufficient space between it and the ceiling to accommodate the system of cords and pulleys. If the chandelier were to be lowered, so that candles could be lit or put out, it could be grasped by the ring passing through the mouth of the lion’s head which decorates its base. One lighted candle burns in one of the sconces. Wax has dribbled over the foremost sconce on the right, which is now empty; the rearmost sconce on our right may contain the stub of an unlit candle.⁷²

Polished brass chandeliers, which gleamed like gold, were splendid and expensive objects, for brass was not easy to manufacture. An Italian visiting the Low Countries in 1517 was struck by the fact that ‘in all the churches … there are in the choirs branched chandeliers (arbori) and lecterns of brass …’.⁷³ In 1441 the Abbot of St‐Vaast at Arras paid the Tournai founder Michel de Gand the large sum of 53l. 16s. of Artois for ‘ung lamppier portant candélabre de laitton, lequel est mis et pendans au ceur de ladicte église’: it had been designed by Jacques Daret.⁷⁴ In 1459 the Abbot of St‐Aubert at Cambrai purchased a brass chandelier with five branches to set before a triptych just acquired from Rogier van der Weyden.⁷⁵ These may have been similar to surviving chandeliers which are larger and more elaborately branched than that in NG 186.⁷⁶ The metal chandeliers very occasionally seen in miniatures of palace interiors were possibly of silver gilt rather than brass but, even in palaces, most chandeliers were wooden.⁷⁷ The great halls erected at Bruges in 1430, when Philip the Good married Isabella of Portugal, and in 1468, when Charles the Bold married Margaret of York, were lit by wooden chandeliers.⁷⁸ Brass chandeliers were occasionally listed in inventories. Paolo Guinigi, Lord of Lucca, owned at the time of his expulsion from the city in 1430 ‘a brass chandelier with six branches, chased and with little brass bells and other decorations, hanging from the ceiling’ of the first chamber next to the great hall of his palace.⁷⁹ This was again rather grander than the chandelier in NG 186, which was perhaps more like an object described in a Rouen inventory of 1435: ‘a large brass chandelier with six branches which is said to be in the fashion of Flanders’.⁸⁰ A six‐branched brass chandelier was described in a Bruges inventory of 1469;⁸¹ at [page 188]Deventer in 1476 Clawes Timynck had owned three chandeliers but, as the materials from which they were made were not specified in the inventory of his estate, they were probably wooden.⁸² Another burgess of Deventer in 1479 had owned a ‘metal chandelier weighing eleven pounds’: a small object, perhaps made of iron rather than brass.⁸³ Philip of Cleves, Lord of Ravenstein, who died in 1528, seems to have had only one brass chandelier in his castle at Enghien;⁸⁴ while Everard IV de la Marck, grand mayeur of Liège, who died in 1531, appears to have owned only one ‘hanging brass chandelier’.⁸⁵ Though there may have been many brass chandeliers in churches, they do not seem to have been common in domestic settings. The chandelier in NG 186 is larger and more elaborate than those painted by Dirk Bouts in his Prado Annunciation and his Last Supper in the church of St Peter at Louvain.⁸⁶ Though not the grandest chandelier available, it is an impressive and expensive object. Arnolfini is unlikely to have had more than one such chandelier and he would have displayed it in his principal reception room.

Hanging from a nail beside the mirror are twenty‐nine semi‐transparent beads, which are probably amber and which are strung on a green cord with green tassels at both ends. The tassels, and therefore also the cord, are evidently silk. The beads are apparently prayer‐beads and similar paternosters of amber with silk tassels are described in the Burgundian inventories.⁸⁷ The dukes’ paternosters were decorated with gold or pearls; the paternoster in NG 186 is a less expensive, but still a precious, object.

The convex glass mirror is surrounded by a decorative red and blue border which seems itself to be convex and may be holding the mirror into its frame. The frame is ten‐sided and is perhaps of painted wood. Its edge is red and its surface, which is slightly shiny, is grey patterned with dashes of black. Ten roundels have been cut into its surface and in the roundels are ten scenes of the Passion of Christ. Each one has a red background and each is protected by a convex glass. The series begins at the bottom, with the Agony in the Garden, and continues clockwise round the frame with the Arrest of Christ, Christ before Pilate, the Flagellation, Christ carrying the Cross, the Crucifixion, the Descent from the Cross, the Entombment, the Harrowing of Hell and the Resurrection. Mirror frames were often decorated with religious images. A [page 189]fifteenth‐century Italian mirror frame exists which is octagonal; it is made of bone or ivory and is carved with figures of angels.⁸⁸ In the Burgundian ducal collection there was, for example, a mirror framed in silver gilt with, on the front, an enamel of Our Lady and her Son seated in a sunburst and, on the other side, a Coronation of the Virgin.⁸⁹ According to Saint Bernard, Christ’s suffering body was a mirror for the soul.⁹⁰ The Passion was therefore a particularly suitable subject for a painted mirror frame. The mirror in NG 186, though not decorated with jewels or precious metals, is an unusually elaborate object. Mirrors depicted in fifteenth‐century paintings and mansucripts manuscripts are almost invariably in plain round frames. In Bruges, nevertheless, the mirror‐makers and the painters belonged to the same guild and the mirror‐makers were permitted themselves to decorate with paintings the frames of their mirrors.⁹¹ The leading Bruges mirror‐maker was perhaps Michiel Langhe Jans, who was a guild official in 1425–6 and 1435–6.⁹²

By comparison with the convex mirror represented in the Saint Eligius(?), dated 1449, by Petrus Christus,⁹³ the mirror in NG 186 seems very large indeed. A convex mirror was manufactured by blowing a globe of glass and by introducing through the blowpipe a mixture of antimony, tin and resin or tar to coat the inside surface. The globe was then cooled and cut into circular mirrors. At Bruges in 1468 Jean Scallequin made two huge and fantastic chandeliers for the great hall in which took place many of the festivities following the marriage of Charles the Bold and Margaret of York. The base of each chandelier was set with seven mirrors, ‘the largest available’,⁹⁴ arranged in a rose pattern. According to Olivier de la Marche, the mirrors were one and a half feet, about 40 cm, in diameter and appeared to reflect ‘ten thousand men’.⁹⁵ They were evidently hemispherical. A mirror with a less pronounced curvature, such as that in NG 186, would necessarily have been much smaller.

The diameter of the mirror frame is approximately equal to the distance between the seat of the bench and the floor. If this distance is reckoned to be about 45 cm, the mirror glass is at least 28 cm in diameter and is therefore extremely large. It is, of course, impossible to say whether Arnolfini owned so big a mirror or whether van Eyck exaggerated its size. He included in his underdrawing an even larger octagonal object which was perhaps a negenoog or nine‐eyes: a name which, in the nineteenth century, peasants in the region north of Bruges gave to a round mirror surrounded by eight smaller circular mirrors.⁹⁶ This may have been a traditional design. It is conceivable that van Eyck drew a round mirror in an octagonal frame – a conventional negenoog – and exaggerated its scale; he would then have imagined a more elaborate, ten‐sided object, reduced the diameter of the glass to make it a more plausible size and substituted for the eight small mirrors the ten paintings of Passion scenes.

The mirror reflects the rest of the room and the space seen through the open door. This seems to be a passage, lit by windows on our left; another open window is at the back. Two men are entering. The man in front wears a blue chaperon with a long cornette and a blue robe. He appears to be raising his left arm (the right arm of his reflection) and is taking a step forward with his left foot (the right foot of his reflection). It seems that he is descending the first of two(?) stairs which separate the corridor from the reception room and that he has just arrived in the doorway; he and the principal subject seem to be exchanging greetings. The man behind him wears a red conical(?) hat and a red robe.

Above the mirror is lettered the inscription Johannes de eyck fuit hic/ .1434. It is decorated with cadels, like any other elaborate inscription. Jean Flamel’s ex‐libris inscriptions in the Duke of Berry’s manuscripts are wonderfully ornamental,⁹⁷ as are the titles of the knights on the stall‐plates of the Order of the Golden Fleece.⁹⁸ Ornamental inscriptions were considered appropriate wall‐decorations (fig. 21),⁹⁹ though the placing of an artist’s signature on a wall, rather than a device or a moral precept, is altogether exceptional.¹⁰⁰ The implication is that Jan van Eyck is the foremost of the two men seen in the mirror; the man behind him is perhaps a servant, for Jan, as varlet de chambre to the Duke, was entitled, at least when he was at court, to have his own varlet.¹⁰¹ It is thought that Jan included his own portrait as a small figure reflected in Saint George’s polished shield in his Virgin and Child with Canon van der Paele (Bruges).¹⁰²

The dog is clearly a household pet and has been recognised as one of a canine family from which the Brussels griffon has been bred.¹⁰³ The dog in NG 186 has brown eyes with vermilion secondary lights; its hair is grey‐brown, mixed from black, vermilion, yellow and white, but some of the strands [page 190]are nearly pure vermilion. It is not like the dogs found in paintings and miniatures of the Burgundian court, nor is it similar to those small, whitish dogs with long hair and long, drooping ears which are found in paintings by van der Weyden, Memling and their followers.¹⁰⁴

The room, often mistakenly called a bedroom, is in fact a reception room and the bed is an essential part of its furnishing.¹⁰⁵ In 1445, French ambassadors received at the Palace of Westminter were evidently surprised to find Henry VI in a richly furnished room without a bed.¹⁰⁶ In NG 186, the interior is expensively furnished but it does not rival in luxury the public rooms of the Burgundian palaces, for the floor is not tiled or covered in carpets, the walls are not panelled or hung with rich fabrics and there is no dresser with a display of valuable plate. It is a room in a commodious mansion where the owners are making a display of their wealth without risking criticism for imitating too closely the life‐styles of the court and high aristocracy. It seems at first that van Eyck has painted a real room exactly as he saw it; but a careful examination of the painted room and its reflection in the painted mirror reveals many inconsistencies. There is, for example, insufficient space for the chandelier, which, even at its present height, would seriously impede movement towards the back of the room. There is no sign of a fireplace and no possibility of siting one in an outside wall.

The two principal figures are richly dressed. The man wears a hat made of plaited straw dyed or painted black. His eyes are grey‐blue; his cheeks and nose are covered with tiny vermilion spots and he has no stubble. His outer garment is a tabard or heuque trimmed and lined with brown fur, possibly sable. The fabric of the heuque is probably silk velvet and it was once dark crimson‐purple: the pigments are red lake, ultramarine, black and white.¹⁰⁷ The doublet beneath is made of a thinnish patterned fabric, possibly satin damask; the design of arabesques and leaf shapes is in grey on black, with slightly purplish shadows. His cuffs are of silver braid on a purple base; his right cuff is tied with a scarlet lace which has silver tips. His hose and boots appear to be dark purple and the tops of his boots are defined by brown lines. He wears a gold ring on the index finger of his right hand. His discarded pattens have wooden soles and leather straps nailed to the soles. The straps are closed and appear to be joined with pins [page 191]passing through loops. Both pattens are spattered and stained with dirt; they are not new, for they are more worn on the insides than on the outsides of the feet.

The woman, who has brown eyes, has fair hair which is dressed in horns surrounded by small plaits. The horns are caught in intricately woven red nets and covered in a white linen veil which has been folded backwards and forwards on itself five times;¹⁰⁸ the edges are fluted. Around her neck she wears narrow gold chains. Her dress is of fine green wool and is trimmed and lined with white fur, perhaps ermine or pured miniver, the belly fur of squirrels. Van Eyck has apparently shown fur of a purer colour and finer texture than any white fur that was obtainable in his time.¹⁰⁹ The dress is very long and she has had to gather it up and hold it across her stomach. Though to a modern eye she may look pregnant, she is not: in van Eyck’s Dresden triptych of 1437, the virgin Saint Catherine is similarly posed and dressed, and is precisely comparable in shape.¹¹⁰ The ends of the woman’s hanging sleeves are decorated with dagging. The cloth has been doubled and cut into shapes like Maltese crosses; the edges are pinked to prevent fraying and to create a decorative effect. Three strips of dagging, each three crosses wide and several crosses high, are sewn in overlapping layers below the slit of her left sleeve. The exact structure of the dagging below the bags of the sleeves is difficult to discern but the crosses are still doubled and pinked.¹¹¹ The underdress, visible at the sleeves and hem, is blue and trimmed at the hem with white fur. The blue material may once have been patterned and was conceivably damask. Her sleeves are gathered at the wrists into bands of gold and pink braid. The cuff of her chemise is just visible at her left wrist. Her pink belt may be dyed leather and carries a golden decoration of dots and lozenges; at the centre of each lozenge is a small blue dot. She wears gold rings on the ring and little fingers of her left hand. Her discarded pattens are covered in red leather and have leather straps with brass(?) studs and buckles.

The clothes are not showily extravagant and neither the man nor the woman is making a great dispay of jewellery, but the fabrics, furs and braids would have been expensive. The man’s heuque would have been especially costly: if it is indeed velvet, it would have required an enormous amount of silk; it would probably have been dyed with expensive insect dyes; and the lining of marten skins would have cost a great deal. The dagging on the woman’s dress would have taken much labour and the fur lining, particularly if it is ermine, would have been very expensive indeed. The enormous train is an assertion of wealth and importance. In the Dresden triptych, Saint Catherine has a huge train because she is a princess.

The man’s hat and heuque are similar to those worn by some of the courtiers in the Festival at the Court of Philip the Good, known from copies at Versailles and Dijon. The lost original was painted in about 1430.¹¹² In the early 1430s, Philip the Good was buying velours plain en graine, perhaps similar to Arnolfini’s velvet, and heuques, some of which were black velvet and lined with sable.¹¹³ The woman’s clothes are not unlike those worn by some of the ladies in the Festival or those of Jacqueline of Bavaria in the copies after Lambert van Eyck’s lost portrait of 1432.¹¹⁴ Wooden pattens were in fashion at the Burgundian court in the 1430s and 1440s.¹¹⁵ It seems legitimate to conclude that the couple are wearing expensive and fashionable clothes.

The abundance of rich furs is perhaps surprising, since the cherries visible through the open window indicate that the season is high summer. The furs enable the couple to show off their wealth: they are magnificently prepared for the severe winter of 1434, ‘the great winter of frosts and many snows’.¹¹⁶

Contemporary observers would have responded to such a display of wealth by pricing the objects depicted. Christine de Pizan, writing in Paris in about 1410, was shocked by the effrontery of a merchant’s wife who ostentatiously exhibited her plate and expensive textiles. The carpets by her grandest bed were laid on the floor, so that people stepped on them; the bed‐curtains were of fine Rheims linen, valued at 300 francs; another linen cloth, as delicate as silk, was ‘all in one piece, without seams, made in a newly discovered way and very expensive …’.¹¹⁷ Christine implied, though she did not directly state, that the merchant’s wife was Italian and remarked that, in Venice, Genoa, Florence, Lucca, Avignon and elsewhere, rich women were less scrupulous in observing the conventions of rank than in Paris, where social hierarchies were more complicated.¹¹⁸ Someone with Christine’s acumen, looking at the painting of the Arnolfini in their chamber, might have been puzzled by the absence of a dresser and of plate but could not have failed to be impressed by the other indications of their great riches.

Displaying plate on a dresser was the most obvious way of demonstrating affluence. In NG 186, the absence of plate from a show of riches is as curious as the contradiction between the furred clothes and the discreet indication that, outside, it is summer. The painting seems so immediately convincing that it is easy to assume that van Eyck painted exactly what he saw, the Arnolfini in their chamber; but there are too many anomalies for that assumption to be justified. Obviously the signature and date did not appear on Arnolfini’s wall, even though it may have borne other decorative inscriptions; and there are other inconsistencies. The absence of a fireplace is disturbing; the chandelier cannot fit into the space it seems to occupy; the bed looks too short; and the mirror may be impossibly large and is unlikely to be a depiction of a real mirror. As we have seen, van Eyck may have based his mirror on one which belonged to Arnolfini – but may have made it larger and given it a more intricate and splendid frame. In the same way, he probably improved upon the appearance of the white fur on the woman’s dresses; he may possibly have invented the pattern on the rug because a real oriental carpet might have looked out of place in his portrait. Whatever happened, it is clear that Arnolfini and his wife did not inhabit a room exactly like the room depicted and that they did not own objects precisely like all those that furnish the room. According to his usual practice, Jan van Eyck has created a perfectly convincing semblance of reality but has altered things, presumably to conform to his aesthetic purposes and perhaps also to accord with Arnolfini’s aspirations. The painting cannot be taken as a literal record.

The Identity of the Couple

Another portrait which obviously represents the same man, more simply dressed and looking older, is at Berlin (fig. 23). It is generally agreed that it is also by Jan van Eyck and that it was painted a few years after NG 186. The board on which the Berlin portrait is painted is from the same tree as the panels of van Eyck’s Portrait of Baudouin de Lannoy (Berlin), perhaps of about 1440, and the Eyckian Stigmatisation of Saint Francis (Philadelphia).¹¹⁹ The Berlin Arnolfini would appear to have been in the Low Countries in the 1560s, when it was used as the basis for an imaginary portrait of Baldwin VII (died 1119), Count of Flanders. This was included in the Codex Iconographicus Flandriae, a manuscript written in Ghent in 1562.¹²⁰ The picture was subsequently taken to Italy and was in the collection of Ranuccio I Farnese (1569–1622), Duke of Parma, and his successors.¹²¹ There are no indications of the sitter’s real identity and the picture is of no direct help in identifying the couple in NG 186.

A few art historians have interpreted the inscription as meaning ‘Jan van Eyck was this man’ and have therefore identified the couple as Jan van Eyck himself and his wife Margaret.¹²² Their interpretation of the inscription is unacceptable. The portrait of the man, furthermore, does not look like a self portrait and is totally unlike Jan’s probable self portrait, NG 222; while the woman bears only a very superficial resemblance to Jan’s portrait of his wife, painted in 1439 (Bruges).¹²³

In the inventories of 1516 and 1523–4, the man was identified as ‘Hernoul le Fin’ and as ‘Arnoult Fin’. In the inventories of 1556–8 and in later records, no attempt was made to identify the couple, though in 1556–8 the coats of arms on the wing panels were still recognised as those of Diego de Guevara. Jan van Eyck seems normally to have written on the frames of his portraits the names of the persons represented; but in the case of NG 186 there may not have been any identifying inscription. Alternatively, the inscription was cryptically worded or difficult to decipher; or the original frame may have been restored when Diego de Guevara altered or added the wing panels. The identifications made in 1516 and 1523–4 may have been based, not on inscriptions by van Eyck, but on oral tradition. This could of course have been recorded on a paper label stuck to the frame in the time of Margaret of Austria but no longer there in the time of Mary of Hungary. In 1516 Diego de Guevara was still alive and the 1516 inventory was taken in the presence of Margaret of Austria, to whom he had given the portrait. Presumably ‘Hernoul le Fin’ and ‘Arnoult Fin’ reflect an identification made by Diego de Guevara, transmitted by him to Margaret of Austria but subsequently forgotten.

Diego de Guevara, the first recorded owner of the portrait, was an important patron of the arts;¹²⁴ he probably knew exactly who the couple were and could have been acquainted with them. He was a younger son of Ladrón de Guevara, Lord of Escalante near Santander.¹²⁵ He may have been born in about 1450¹²⁶ and was brought up at the Burgundian court.¹²⁷ His illegitimate son Felipe de Guevara owned two portraits of Diego and stated in about 1560 that one of them had been painted more than ninety years previously by Rogier van der Weyden.¹²⁸ If that is true – and there is no good reason to doubt the statement – Diego was in the Low Countries by 1464, when Rogier died. He is said to have been one of Charles [page 193]the Bold’s esquires at the Battle of Nancy in 1477 and to have flung himself over Charles’s dead body in order to protect it.¹²⁹ He was certainly one of Mary of Burgundy’s esquires at the time of her death in 1482.¹³⁰ His older brother Ladrón de Guevara is stated to have been in the service of Philip the Good, who died in 1467, and to have taken part in Philip’s wars against Charles VII of France (who died in 1461) and Louis XI.¹³¹ By 1471 he held an important and well‐paid position in the household of Charles the Bold;¹³² by January 1473 he was esquire and chamberlain and by September 1473 he was one of the maîtres d’hôtel.¹³³ Either Ladrón or Diego might have known the couple painted by van Eyck; they might very well have had acquaintances in common.

Crowe and Cavalcaselle in 1857 were the first to make a connection between NG 186 and the portrait of ‘Hernoul le Fin’ or ‘Arnoult Fin’ owned by Margaret of Austria.¹³⁴ ‘Hernoul le Fin’ could be a corruption of Arnoul le Fin, shrewd Arnold (and it may be noted in passing that a singer named Arnoldus de Fine was at Bruges in 1438 and 1449 and was employed as a scribe in 1452–3).¹³⁵ As Crowe and Cavalcaselle were once more the first to point out, ‘Hernoul le Fin’ and ‘Arnoult Fin’ are much more likely to be attempts to spell ‘Arnoulfin’. This was how the Italian surname Arnolfini was written, and doubtless also pronounced, in the Netherlands and in France. The Arnolfini were from Lucca but many of them came to trade in northern Europe. Crowe and Cavalcaselle thought that ‘Hernoul le Fin’ might be Giovanni Arnolfini, mentioned in the 1420s in the accounts of the Burgundian court.¹³⁶

Developing Crowe and Cavalcaselle’s observations, Weale in 1861 made the natural but – as it happens – incorrect assumption that this Giovanni Arnolfini was the same man as Giovanni [di Arrigo] Arnolfini, who died in Bruges in 1472. Weale concluded that the couple were Giovanni di Arrigo Arnolfini and his wife Jeanne or Giovanna Cenami.¹³⁷ Weale’s theory has been widely followed. Davies, however, was cautious and identified the pair as ‘Giovanni(?) Arnolfini and Giovanna Cenami(?)’;¹³⁸ while Dhanens argued that the man was Giovanni di Arrigo’s younger brother Michele Arnolfini. She believed that the painting represented a morganatic marriage and that Michele Arnolfini may have contracted a morganatic marriage.¹³⁹ She was, however, mistaken on both counts: the portrait does not depict a morganatic marriage;¹⁴⁰ and Michele’s marriage cannot have been morganatic, since his sons and grandsons were recognised in Lucca as the heirs of Giovanni di Arrigo and occupied prestigious positions in the government of the city.¹⁴¹

If ‘Hernoul le Fin’ and ‘Arnoult Fin’ are correctly interpreted as standing for Arnoulfin or Arnolfini, and if the identification was made on good authority, the man was a member of the Arnolfini family. During the second quarter of the fifteenth century, at least five of the Arnolfini, brothers and first cousins, lived in Bruges. In principle, van Eyck’s patron could have been any one of the five. Giovanni di Nicolao Arnolfini was residing in Bruges by 1419,¹⁴² became a burgess in 1442¹⁴³ and was still there in 1452.¹⁴⁴ His younger brother Battista di Nicolao Arnolfini, who died in 1446,¹⁴⁵ was mentioned in an Antwerp document of 1437¹⁴⁶ and was described in September 1439 as ‘merchant residing in Bruges’.¹⁴⁷ Another brother, Bartolomeo di Nicolao Arnolfini, who died in 1473, brought a lawsuit in 1449 before the Vierschaar of Bruges against his cousin Giovanni di Arrigo.¹⁴⁸ Michele di Arrigo Arnolfini was in London in 1436–7¹⁴⁹ and probably passed through Bruges on his way from and to Lucca. By 1449 he had settled in Bruges,¹⁵⁰ where he died in 1472.¹⁵¹ Michele’s elder brother Giovanni di Arrigo Arnolfini, the most successful and prosperous member of the family, spent much of his life in Bruges and died there on 11 September 1472.¹⁵² Each of the five will be considered here.

Giovanni di Nicolao Arnolfini

If patronymics are not given, it is not always easy to distinguish between Giovanni di Arrigo and Giovanni di Nicolao. The latter, often called Giannino or Jehannin, was certainly the older of the two: the diminutive form Giannino, frequently used by the Lucchese, should not be taken as an indication of youth.¹⁵³ His father Nicolao di Giovanni Arnolfini was a figure of some importance in Lucchese politics, served Paolo Guinigi, Lord of Lucca, and was closely associated with the Guidiccioni family.¹⁵⁴ Nicolao died between 1427 and 1430, before the expulsion of Paolo Guinigi.¹⁵⁵ The name of Nicolao’s wife has not been discovered; he had at least five sons. At Lucca on 29 March 1419 Nicolao Arnolfini appointed as his proctor Marco Guidiccioni, residing in Bruges, to emancipate [page 194]his son Giannino di Nicolao Arnolfini, also residing in Bruges.¹⁵⁶ As sons could be emancipated at any age, it is not possible to make deductions concerning Giovanni’s date of birth from the legal formality of his emancipation.¹⁵⁷ It seems to have been the custom to send young Lucchese merchants abroad when they were in their mid‐teens: they might previously have been instructed in the Flemish language, for at Lucca in 1453 Michele Guinigi owned a book on learning Flemish.¹⁵⁸ Biagio Balbani (1453–1523) was sent to Bruges in April 1468, just before his fifteenth birthday, and by 1473, when he was twenty, he was in partnership there with Francesco Micheli. Biagio married in 1483, when he was thirty.¹⁵⁹ Giovanni di Nicolao Arnolfini, who married in 1426,¹⁶⁰ was probably born in the last years of the fourteenth century.

It was clearly Giovanni di Nicolao who, described as Jehan or Jehannin, worked in close association with Marco Guidiccioni, an influential figure at the courts of Burgundy and Brabant.¹⁶¹ In 1421 Giovanni made disguised loans to the town of Bruges;¹⁶² in or shortly before 1422 he was involved in an attempt to sell a valuable collar of gold to Henry V of England;¹⁶³ in 1423 he sold to Philip the Good six tapestries of scenes from the life of the Virgin, which Philip then sent to Pope Martin V.¹⁶⁴ In 1424–5 the brothers Sandei, Lucchese merchants in Venice, went bankrupt, owing Giovanni 9,210 ducats; Marco Guidiccioni’s brother Aldibrando took responsibility for their debt.¹⁶⁵ Between 1424 and 1427, Giovanni, described as Marco Guidiccioni’s factor, regularly sold cloth of gold and silks to Philip the Good.¹⁶⁶ In 1425 he made a disguised loan to John IV of Brabant.¹⁶⁷ Guidiccioni was then away from Bruges and, during his absence, Giovanni was in control of his affairs. After Guidiccioni returned to Bruges in 1427–8, Giovanni was less frequently mentioned.

In 1426 Giovanni had married Costanza Trenta, an extremely well‐connected young woman. She was the only daughter of Girolamo, son of Lorenzo di Maestro Federigo Trenta. Lorenzo and his brothers were merchants of immense wealth and had considerable influence at the courts of France and Burgundy. Girolamo Trenta made his will in November 1423 and left Costanza a dowry of 800 florins or more, at the discretion of her grandfather Lorenzo.¹⁶⁸ On 23 January 1426 Costanza was thirteen and betrothed to Giovanni di Nicolao; she was clearly about to leave Lucca for Bruges. She renounced any claim that she might have, in the event of Giovanni’s death, under the laws of France, Flanders or England; she was to have rights of succession only under Roman law and Lucchese statutes. The Arnolfini agreed that, if she were widowed, she would be honourably escorted back to Lucca, with a sufficient company; her grandfather guaranteed that she would observe her undertaking and that, when she was eighteen, she would ratify it in due form.¹⁶⁹

Costanza’s grandfather Lorenzo Trenta was a notable patron of the arts. He owned a missal elaborately decorated with miniatures in the style of the Boucicaut Master, including a representation of himself and his family at Mass;¹⁷⁰ and he commissioned from Jacopo della Quercia tombstones for himself and his wife Isabetta Onesti and the great Trenta Altarpiece, dated 1422, for his chapel in the church of San Frediano in Lucca.¹⁷¹ It was apparently Lorenzo who was écuyer d’écurie to Philip the Good and whom, in 1425 or 1428, Philip sent to Lucca to recover jewels pledged by John the Fearless to Galvano Trenta.¹⁷² Galvano, who had died in 1421, was Lorenzo’s brother and had held in pawn not only Burgundian jewels but also many of the crown jewels of France.¹⁷³

Costanza’s mother Bartolomea was a Florentine, the daughter of Giovanni di Amerigo Cavalcanti.¹⁷⁴ She married as her second husband Arrigo di Lazzaro Guinigi,¹⁷⁵ whose brother Nicolao Guinigi was Bishop of Lucca from 1394 to 1435 and whose second cousin Paolo Guinigi was Lord of Lucca between 1400 and 1430. Bartolomea made her will on 8 February 1463.¹⁷⁶ Her brother Niccolò Cavalcanti may have commissioned Donatello’s ‘Cavalcanti Annunciation’, still in S. Croce in Florence; her sister Ginevra Cavalcanti married in 1416 Lorenzo de’ Medici, brother of Cosimo il Vecchio. It was on the occasion of Ginevra’s marriage that Francesco Barbaro wrote his De Re uxoria and presented a copy to Lorenzo.¹⁷⁷ Ginevra was the mother of Pierfrancesco de’ Medici (1430–76), who was consequently Costanza Trenta’s first cousin, and the grandmother of Botticelli’s patron Lorenzo di Pierfrancesco de’ Medici.

Costanza must have died shortly after her marriage. Her mother, writing from Lucca on 26 February 1433, mentioned that her daughter Costanza was dead. Since one purpose of the letter was to congratulate the parents of a child born thirty‐three months before, Bartolomea’s news may have been rather out of date. Costanza may have died long before February 1433.¹⁷⁸ Giovanni di Nicolao Arnolfini probably married again but no reference to a second wife has yet been found.¹⁷⁹

On 23 September 1426 Giovanni was one of several proctors appointed by Bartolomea de’ Bardi, widow of Galvano Trenta and wife of Giovanni’s brother Jacopo di Nicolao Arnolfini. The other proctors were Battista, a third brother, and Marco Guidiccioni; they were to make it clear in Paris, Bruges and elsewhere in France and Flanders that Bartolomea’s children by her first husband were under the guardianship of the Trenta family and that she took no responsibility for their affairs. On 1 November 1428 Giovanni, described as ‘Giannino Arnolfini’, witnessed a deed in Bruges.¹⁸⁰

The next fourteen years of Giovanni’s life are less clearly documented. In 1432 Giusfredo Rapondi and ‘Jehan Arnoulphin’, merchants residing in Bruges, sold precious textiles to the Burgundian court.¹⁸¹ This ‘Jehan Arnoulphin’ can be equated with the ‘Jan Aernoulphin’ who in 1436 was Giusfredo Rapondi’s factor. Since Giovanni di Arrigo Arnolfini was by 1436 in partnership with Paolo Miliani, Rapondi’s associate must have been Govanni Giovanni di Nicolao.¹⁸² During the 1430s, Giovanni suffered a mysterious reverse. Possibly it was the death (between 1436 and 1438) of his former associate Marco Guidiccioni,¹⁸³ perhaps it was the arrival of Giovanni di Arrigo Arnolfini but more probably it was the revolt of Bruges in 1437–8 that damaged the business interests of Giovanni di Nicolao. His brother Battista abandoned an active political career in Lucca to visit the Low Countries in 1437 and 1439, perhaps in order to settle Giovanni di Nicolao’s affairs.¹⁸⁴ Whatever happened, Giovanni di Nicolao decided [page 195]to withdraw from commerce but to remain in Bruges. On 22 November 1442, taking advantage of special terms laid down by Philip the Good, Giovanni paid 3 livres parisis and became a burgess of Bruges.¹⁸⁵ He promised never to engage in trade and was empowered to practise the ‘small burgess’s crafts’.¹⁸⁶

How Giovanni di Nicolao spent his last years is something of a mystery. He remained in Bruges, where on 31 January 1449 ‘Jehennin filz de feu Nicolas Arnulfin’ was one of three arbiters, all described as ‘marchans de Luques’, appointed to settle a dispute between Giovanni di Arrigo and the tutors of the only son of deceased Battista Arnolfini. Battista was Giovanni di Nicolao’s brother and had died in 1446; the dispute was over ‘merchandise and other matters’ but no details were given. Battista may have taken over some of Giovanni di Nicolao’s concerns but had not seen fit to make him one of his son’s tutors.¹⁸⁷ On 21 October 1452 ‘Jehennin Arnoulphin fils de Nicolas’ was once again an arbiter, one of seven ‘communs marchands de la nation de Lucques à Bruges’. They and the Lucchese consul had to settle a dispute involving Giovanni’s former associate Giusfredo Rapondi.¹⁸⁸

Giovanni di Arrigo Arnolfini

Always called Giovanni or Jehan, never Giannino or Jehannin, he was the eldest of the four sons of Arrigo Arnolfini, all of whom were described on 16 January 1419 as pupilli and who were therefore aged between seven and thirteen.¹⁸⁹ Giovanni di Arrigo must have been born between 1405 and 1409. His father Arrigo Arnolfini, who died in 1419, lived in considerable luxury in a large palazzo.¹⁹⁰ He married in 1399 Antonia Guinigi, first cousin of Paolo Guinigi, who ruled Lucca between 1400 and 1430.¹⁹¹ According to the chronicler Chastellain, Giovanni di Arrigo came to Bruges ‘a poor journeyman but made a fortune of 200,000 florins by selling silks to the ducal household and by farming the Gravelines toll’.¹⁹² ‘Poor’ is clearly an exaggeration: he came from a wealthy and well‐connected family but his vast fortune was of his own making.

Shortly before 29 September 1435, Philip the Good made the huge payment of 5,669 livres 10 sols de 40 gros, for cloth of gold, damask and satin, to ‘Jehan Arnoulphin le jeune’.¹⁹³ This must be Giovanni di Arrigo, called ‘le jeune’ to distinguish him from his older cousin Giovanni di Nicolao. No place of residence was given, perhaps because Giovanni di Arrigo had just arrived from Italy and had not yet set up a permanent establishment. By 1435 he was clearly very wealthy. It is not possible to say exactly when he arrived in the Low Countries. References of 1436–7 and later to a Giovanni Arnolfini, merchant in Bruges,¹⁹⁴ are apparently to Giovanni di Arrigo, recognised by that time as ‘the’ Giovanni Arnolfini.

From 1439 ‘Jehan Arnolphin’ was regularly mentioned in the Burgundian accounts:¹⁹⁵ from 1436 he was in association with another Lucchese, Paolo Miliani;¹⁹⁶ he gradually established himself as the chief supplier to the court of silks and cloths of gold and silver. In a period of twenty‐three months, from April 1444 to February 1446, he was paid 17,347 livres, approximately two per cent of the total expenditure of the recette générale.¹⁹⁷ In 1449, Philip the Good granted to him, for six years from April 1450, the right to farm the Gravelines toll, payable on all goods in transit between the Burgundian Netherlands and the English Staple at Calais.¹⁹⁸ In 1451 he acquired extensive premises near the Kruispoort in Bruges, from which he probably ran his business.¹⁹⁹ By 1454 he was councillor to the Duke;²⁰⁰ and in 1456 he had another grant, for a further six years, of the Gravelines toll.²⁰¹ In 1458 began his affair with Christina van der Wijck, who was later to accuse him of having raped her, to whom he promised mansions in Bruges and in Brussels and to whom he gave impressive quantities of jewellery and furniture.²⁰² From 1460 he was on friendly terms with the exiled Dauphin, then living at Genappe, south of Brussels.²⁰³ When the Dauphin became King of France as Louis XI in 1461, he made Arnolfini his councillor and Général des Finances in Normandy;²⁰⁴ by January 1463 Arnolfini had been knighted;²⁰⁵ and in April 1464 he became a naturalised Frenchman.²⁰⁶ For a time he remained on good terms both with Louis XI, who seems often to have dined in his house,²⁰⁷ and with Philip the Good, with whom he was ‘very familiar and friendly’.²⁰⁸ In 1465, however, when war broke out between France and Burgundy, Arnolfini attempted to return to Flanders but was arrested at Abbeville and was replaced as Général des Finances.²⁰⁹ The rest of his life he appears to have spent in the Low Countries and by April 1470 he was maître d’hôtel at the Burgundian court.²¹⁰

Giovanni di Arrigo did not marry until 1447. Between 1 April and 11 November 1447, Philip the Good purchased two silver pots ‘which my said lord presented as a gift from himself to Jehan Arnoulphin, merchant, residing in the said town of Bruges, on the day of his marriage’.²¹¹ Between 1454 and 1458, ‘Jan Aernulphijn’ and ‘Jane’ his wife joined the Bruges Confraternity of the Dry Tree.²¹² ‘Jane’ was probably Jeanne Cenami, who was certainly his wife by September 1464,²¹³ and who must have been the unnamed wife who, with Giovanni, was a member of the Bruges Confraternity of Our Lady of the Snow.²¹⁴ Jeanne was one of her husband’s executors²¹⁵ and died on 14 October 1480.²¹⁶ Though she had a pew in the church of St James,²¹⁷ she was buried with Giovanni in the church of the Rich Clares, where she endowed daily and annual masses.²¹⁸

In the numerous documents concerning Giovanni di Arrigo’s foundations in Lucca and in Bruges, there is no reference to any wife but Jeanne Cenami.²¹⁹ It seems reasonable to conclude that he married only once and that the woman whom he married in 1447 was Jeanne Cenami. They had no children, though he had two illegitimate daughters. He left Jeanne in control of his vast fortune²²⁰ and she made over certain rights to her nephew, Jean Cenami, who lived in France;²²¹ but Giovanni di Arrigo’s property passed eventually to his brother Michele’s descendants, who returned to Lucca.²²²

Giovanni di Arrigo Arnolfini died on 11 September 1472 and was buried in the Guidiccioni chapel, the chapel of the Holy Cross, in the church of the Rich Clares in Bruges.²²³ A magnificent memorial service was held at the church of San Romano in Lucca, with the flags of France and Burgundy much in evidence.²²⁴ The tomb in Bruges was apparently an [page 196]impressive one and his executors bought the patronage of the chapel from Marco Guidiccioni’s son.²²⁵ Church and chapel were devastated in 1571 but Arnolfini’s tombstone was recovered in 1613 and placed in the new church.²²⁶ It is now known only from a drawing made in 1715²²⁷ and a plan of the church made in 1775.²²⁸ Though large, it was simply decorated, with seven shields bearing the coats of arms of the Arnolfini and Cenami families.

Arnolfini left an elaborate will. No copy of the complete document has yet been found but extracts show that he made bequests on a princely scale.²²⁹ Daily masses were said in his memory in the churches of the Rich Clares and the Augustinians in Bruges.²³⁰ In his chapel in the church of the Rich Clares, there was, on the Sunday nearest the anniversary of his death, a sung requiem, with a full choir; on the anniversary itself, thirty requiem masses were said and alms were distributed to thirty poor persons.²³¹ He also left funds to endow daily masses in Lucca: at Santo Agostino, at San Romano (the Dominican church), at San Pier Cigoli (the Carmelite church), at the Franciscan church (the Franciscans declined to accept the endowment, which was transferred to the church of San Frediano), at the cathedral of San Martino and at the charterhouse of Farneta outside Lucca.²³² At San Martino, the daily mass was celebrated in the chapel of the Volto Santo and on every anniversary of Arnolfini’s death a ‘great’ mass was sung in the same chapel.²³³ He made a large bequest for the support of poor virgins in Lucca;²³⁴ there were doubtless other bequests to foundations in Bruges, Lucca and elsewhere.

Few of the works of art commissioned by Giovanni di Arrigo Arnolfini or for his foundations have been preserved. The missal used in his chapel in Bruges was still there in 1715²³⁵ but has since been lost; the great choirbook which he gave to Lucca Cathedral was cut up in the seventeenth century but has been partially reconstructed. It was written and illuminated in Bruges, apparently in the late 1460s.²³⁶ Whatever the interest of its content, the illumination seems unremarkable in quality. A copy of Le Livre de l’information des princes, a French translation of a text by Saint Thomas Aquinas, was transcribed in Paris for Arnolfini and finished on 4 February 1454; it is now in the Koninklijke Bibliotheek at The Hague (MS 76 E 20).²³⁷ The prologue is addressed to the young Charles the Bold, for whom the manuscript was evidently intended. The miniatures, by a Parisian illuminator, are not of very good quality. The first, on fol. 1, at the beginning of the Prologue, shows a kneeling man presenting a book to a young Knight of the Golden Fleece (fig. 24): they are no doubt Arnolfini and Charles the Bold but they are not, unfortunately, recognisable portraits. The kneeling man has brown hair, a very pale face and widely spaced eyes of uncertain colour; he wears a light red robe trimmed with brown fur, has a dark grey chaperon slung over his shoulder and carries a green purse at his belt. He is accompanied by a grey dog. It is perhaps worth stressing that Arnolfini’s tombstone was a very conventional one and that the choirbook and the manuscript which he presented to Charles the Bold are of no great aesthetic importance. There is no indication that Giovanni di Arrigo Arnolfini took much interest in the visual arts.

Michele di Arrigo Arnolfini

Michele was the youngest of the four sons of Arrigo, all of whom were pupilli, aged between seven and thirteen, on 16 January 1419:²³⁸ he must therefore have been born between about 1409 and 1412. He was in London in 1436–7 and probably passed through Bruges on his way from and to Lucca. Between 1439 and 1443 he followed a prominent, but very brief, political career in Lucca; by 1449 he had settled in Bruges, where he died in 1472.²³⁹ His wife Elisabeth, whom he may have married in about 1450, was mentioned with her husband in the mid‐1450s in the register of the Confraternity of the Dry Tree in Bruges.²⁴⁰ They had three sons and five daughters;²⁴¹ a younger son, Nicolao, was presumably the ‘Nicolaus Arnoulfin de Brugis’ who matriculated at Louvain university on 31 August 1475²⁴² and who was therefore probably born in or shortly after 1460. Michele’s son Giovanni Battista was by 1480 a ‘familiar’ of Pope Sixtus IV.²⁴³ The brothers Giovanni Battista and Nicolao di Michele Arnolfini married two sisters whose brother Giovanni Francesco Franciotti married in 1480 Luchina della Rovere, niece of Sixtus IV and sister of the future Pope Julius II.²⁴⁴

Battista di Nicolao Arnolfini

Battista was in Paris in the 1420s²⁴⁵ but had returned to Lucca by 1430, where he took a prominent part in politics and where he was Gonfaloniere in November and December 1432.²⁴⁶ In 1436–7 and 1438–9 he visited the Netherlands. Mentioned in an Antwerp document of 1437, he was described in September 1439 as ‘merchant residing in Bruges’.²⁴⁷ Returning to Lucca, he resumed his political career and was again Gonfaloniere in 1443. Having made his will on 13 September 1446, he died shortly afterwards. Battista married Antonia, daughter of Jacopo Micheli; they had an only child, Nicolao.²⁴⁸

Bartolomeo di Nicolao Arnolfini

Bartolomeo, who died in 1473, brought a lawsuit in 1449 before the Vierschaar of Bruges against his cousin Giovanni di Arrigo Arnolfini.²⁴⁹ Bartolomeo, who was Gonfaloniere of Lucca in 1452, 1456, 1462, 1466 and 1471,²⁵⁰ had to go to Rome in 1458 to settle a dispute between the Arnolfini and the Guidiccioni which was threatening the peace of Lucca.²⁵¹ Bartolomeo married Filippa Burlamacchi, who died in 1502. The available information on their children suggests that they married after 1450.²⁵²

Jan van Eyck and the Arnolfini

No document has yet been found to link any of the Arnolfini with Jan van Eyck. A Pietà by Jan van Eyck, however, was in 1480 in the possession of the widow of another Lucchese merchant, Paolo di Poggio;²⁵³ and Paolo had been in partnership during the 1450s with Giovanni di Arrigo Arnolfini.²⁵⁴ The only certain portrait of one of the Bruges Arnolfini is too schematic to be of much assistance: the miniature in Le Livre de l’information des princes cannot be used to confirm or reject the idea that the man painted by van Eyck is Giovanni di Arrigo Arnolfini.

The man in NG 186 was clearly prosperous and was evidently married to the woman beside him. The fact that Jan van Eyck painted a second portrait of the man (fig. 23), and the numerous changes in NG 186, presumably made with the couple’s approval, demonstrate sufficiently that they knew him well and consequently that they were not recent incomers from Lucca. Because NG 186 was never sent to Italy and because the Berlin portrait of the same man did not reach Italy until about 1600, it may be deduced that the couple remained in the Netherlands and that both portraits passed to heirs or legatees who stayed in the Low Countries. The man in NG 186 was therefore married by 1434, had lived in the Low Countries for long enough to get to know Jan van Eyck and left heirs who did not immediately return to Lucca.

Whereas none of the five Arnolfini recorded at Bruges necessarily fits this description, the most likely candidate is Giovanni di Nicolao. Giovanni di Arrigo, whose interest in the visual arts may not have been particularly strong and who was not necessarily in the Low Countries when the portrait was painted, did not marry until 1447; his heirs returned to Lucca. Michele di Arrigo, born between c. 1409 and 1412, seems to have been too young to have been the man painted by van Eyck; it is improbable that they could have got to know each other well before 1434. In any case, Michele seems not to have married until about 1450 and he had several children who returned to Lucca, presumably taking with them their father’s possessions. Battista di Nicolao and Bartolomeo di Nicolao merely visited the Netherlands. Even if they were married at the times of their visits, their wives are unlikely to have accompanied them; both spent their last years in Lucca and left children there who would have inherited their possessions.

Giovanni di Nicolao, in contrast, had lived in Bruges since 1419 or earlier and would have had every opportunity to become acquainted with Jan van Eyck well before 1434. Though he had married Costanza Trenta in 1426, she had died before, and possibly long before, February 1433. No proof has yet been found that Giovanni di Nicolao married a second wife, though it seems likely that he did. He remained in Bruges, probably died there and seems not to have had any surviving children. If Giovanni had no direct heir in Flanders, his property may have been sold after his death. The Berlin portrait, [page 198]showing the same person more simply attired and at a more advanced age, may perhaps support the identification of the man in NG 186 as Giovanni di Nicolao. The decline in his fortunes might explain why in 1434, in NG 186, he was shown arrayed in silks, whereas in about 1440, in the Berlin portrait, he was shown more soberly dressed in woollen clothes.

If ‘Hernoul le Fin’ is rightly interpreted as Arnolfini, then van Eyck’s couple may be tentatively identified as Giovanni di Nicolao Arnolfini and his putative second wife. He would have been in his late thirties in 1434. He may never have known Diego de Guevara, the first recorded owner of the portrait; but Diego could very well have been acquainted with Giovanni’s cousins, Giovanni di Arrigo and Michele, both of whom died in 1472,²⁵⁵ or with Reale Reali, another Lucchese merchant who lived in Bruges, who was still there in 1498²⁵⁶ and who had dealings with the great nobles of the Netherlands and France. Occasionally, at least, he was concerned in transactions involving works of art.²⁵⁷ In 1452, Reali and Giovanni di Nicolao had been arbiters in the same dispute in Bruges.²⁵⁸ Diego might also have come across another Lucchese, Francesco Micheli, whose sister had married Giovanni di Nicolao’s brother Battista. Francesco was still in Bruges in 1476.²⁵⁹ Petronella Arnolfini, perhaps one of Giovanni di Arrigo’s illegitimate daughters, was a nun in the convent of the Rich Clares in Bruges and died on 15 October 1496.²⁶⁰ She would almost certainly have remembered Giovanni di Nicolao. Diego de Guevara, who may well have acquired the painting from the executors of the couple portrayed, had every opportunity to find out about them and must have known exactly who they were.

The Marriage Theory

NG 186 is often called ‘The Arnolfini Marriage’ but in the earliest references, in the inventories of 1516, 1523–4 and 1556–8, it is described merely as a double portrait. The idea that the picture is more than a portrait, that it tells a story, seems first to have been recorded in 1568, twelve years after it had been taken from the Netherlands to Spain, by the Ghent chronicler and rhetorician Marcus van Vaernewijck (1518–69). In Den Spieghel der Nederlandscher audheyt ( ( The Mirror of Netherlandish Antiquity), he wrote that Mary of Hungary had owned ‘a very small panel’ (een cleen tafereelkin) by Jan van Eyck in which was painted ‘a marriage of a man and a woman who are married by Faith’ (een trauwinghe van eenen man ende vrauwe/ die van Fides ghetrauwt worden) and that Mary had acquired the picture from a barber whom she rewarded with an office worth 100 florins a year.²⁶¹ Den Spieghel has been described as an ‘unbelievable jumble of nonsense and rubbish’, a ‘disordered compilation of fables’. Van Vaernewijck, though an accurate reporter of contemporary events, was a totally unreliable historian.²⁶² In the passage quoted, he was almost certainly referring to NG 186, which did indeed belong to Mary of Hungary, but clearly he had never seen it: it is not a ‘very small panel’ and Mary of Hungary did not acquire it from a barber but inherited it from Margaret of Austria. As two of van Vaernewijck’s statements about the portrait are false, there is no reason to believe the third, that the subject is a couple ‘married by Faith’. The entire passage can be dismissed, with most of Den Spieghel, as nonsense and fable. Van Vaernewijck appears to have thought that the picture included an allegorical figure of Faith. In 1604, van Mander, inaccurately paraphrasing van Vaernewijck and having no further information on the painting, certainly believed that it included a figure representing Faith.²⁶³

In 1599 Jakob Quelviz saw the painting in the palace at Madrid, identified the subject as ‘a young man and young woman joining hands as if they are promising future marriage’ and recorded that, among many other inscriptions, there were two lines of Latin poetry – identifiable as lines from Ovid’s Ars amatoria. According to the 1700 inventory, these lines were inscribed on the frame.²⁶⁴ They come from a section of Ovid’s poem where he offers the cynical advice that a lover should make promises, but not gifts, to the woman who is the object of his desire.

Quelviz saw the painting in the Salle chiqua, next to the Room of the Furies and adjoining the room where Philip II had lived when he was ‘very old and ill’ (he had died in 1598). In the same room hung Cranach’s picture of Charles V and John Frederick of Saxony hunting at Torgau (dated 1544, now in the Prado), portraits of Philip II’s Savoyard grandchildren and a double portrait of his daughters Isabella and Catherine – probably the painting by Sánchez Coello now in the Prado (fig. 9).²⁶⁵ Another German, Hieremias Gundlach, visiting the palace in 1598–9, did not mention NG 186 but saw the Cranach and stated that it hung in the room where Philip II had slept.²⁶⁶

The collection of pictures in this small room was a very personal one: portraits of Philip’s beloved daughters and his grandchildren; a portrait of his father whom, even on his deathbed, he strove to emulate; the hunting scene, which would have recalled an activity of which he had been passionately fond. He clearly admired van Eyck’s portrait very much and may well have been able to read and understand the other inscriptions which Quelviz neglected to copy but which may have explained who the couple were. He may have interpreted Ovid’s lines in a moralising sense.

The Ars amatoria was widely known and several French translations had been made during the twelfth and thirteenth centuries. Four manuscripts of French translations were in the library of the dukes of Burgundy, though only one survives.²⁶⁷ The Latin text, first printed in 1471, cannot have been difficult to obtain. In principle, van Eyck could himself have written the inscription on the frame of his portrait. It might perhaps be argued that ‘Hernoul le Fin’ should be translated as ‘shrewd Arnold’ and that the painting illustrates a folk‐tale to which the lines from Ovid were appropriately applied. This is totally implausible. NG 186 is clearly a portrait and was recognised as such in 1516, 1523–4 and 1558; the signature makes little sense if the picture is a genre scene. As it appears impossible to see the painting as a depiction of a deceitful man making empty promises and as, until 1558, the picture is consistently described as a double portrait, there is no need to take seriously the idea that the lines from Ovid were original. They appear to have been added after [page 199]1558 and before 1599, probably by someone who, unused to fifteenth‐century conventions, imagined that the woman was pregnant, that the man was swearing an oath and that the room was a bedroom. The inscription would then have been written when the portrait belonged to Philip II, who perhaps considered it a moralising, rather than a lewd or suggestive, image.

The compiler of the 1700 inventory of the Spanish royal collection identified the inscription on the frame as lines from Ovid and described the couple as a youth and a pregnant German woman: ‘and it appears that they are getting married by night and the verses declare how they are deceiving each other …’²⁶⁸ No doubt he saw the painting in a dirty state and in a dark room. Mistakenly believing that the woman was pregnant, misled by the candle into thinking that the picture was a night scene and putting his trust in an inscription that was unlikely to have been original, he interpreted the portrait as if it had been a seventeenth‐century genre piece and made up a story (‘it appears that they are getting married’) to suit his interpretation. His story had been forgotten by 1794, when the picture was described merely as ‘a man and a woman holding hands’.²⁶⁹ It was exhibited in 1841 as ‘Portraits of a Gentleman and Lady’. The compiler of the National Gallery catalogue of 1843 was cautious: ‘The subject of this Picture has not been clearly ascertained.’²⁷⁰

From the time of the 1841 exhibition, many explanations have been put forward; most of the arguments advanced rest on mistaken premisses. An anonymous reviewer in 1841 thought that ‘this strange pair, hand‐in‐hand, resemble nothing better than Simon Pure about to atone for a faux‐pas by making Sarah Prim an honest woman’.²⁷¹ A second anonymous writer in 1843 recorded that ‘Others again are “almost sure” the portraits represent either a mother elect consulting her medical attendant, or even honest Giovanni, the painter, and his wife, engaged in a similarly delicate business’. He himself imagined that the picture was the wing of a triptych, that the centre panel ‘may have represented the Virgin and child, or some patron saint, and the wings, both on their inner and outer sides, the great epochs of modern domestic life’. The subject of NG 186 was ‘an astrologic medicine man in the act of examining the lines on his patient’s hand, with a view to determining, by palmistry, the chances – male or female – of her majority of months …’.²⁷² The idea that the lady was a pregnant ‘fallen woman’ had great appeal for the Victorian public. De Laborde in 1855 took up once again the theory that the painting depicted a marriage and the legitimation of the child of the supposedly pregnant lady; he identified the figures in the mirror as witnesses.²⁷³

In 1934 Panofsky published an erudite and influential article in which he attempted to prove that the painting represented the private marriage of Giovanni di Arrigo Arnolfini and Jeanne Cenami.²⁷⁴ Since Giovanni di Arrigo married Jeanne Cenami in or after 1447, and since the portrait apparently represents Giovanni di Nicolao Arnolfini, Panofsky’s theory has little to recommend it. Because his article has been so very influential, however, it is necessary to consider his arguments in some detail.

His starting point was the passage from van Vaernewijck which has already been mentioned (he did not know about the lines from the Ars amatoria inscribed on the frame). Though Panofsky did not allude to the notorious unreliability of van Vaernewijck, he realised that van Vaernewijck had not seen the painting and proposed that he had heard of it from a Spanish correspondent. There is no proof that van Vaernewijck had contacts in Spain and it is much more likely that he concocted his story from reports by people who had seen the picture before it left the Netherlands. Panofsky, however, thought that van Vaernewijck’s putative Spanish correspondent wrote to him in Latin and that van Vaernewijck mistranslated the Latin. Van Vaernewijck in fact had little formal education, knew no French²⁷⁵ and had no competence in Latin, though presumably he must have realised that Fides meant Faith.²⁷⁶ According to Panofsky, when van Vaernewijck wrote that the couple were being ‘married by Faith’ (van Fides ghetrauwt), he had misunderstood his hypothetical Latin source. By using the Latin word Fides and by giving it a capital letter, he implied that an allegorical figure of Faith was represented. Still according to Panofsky, van Vaernewijck should have written that they were married per fidem, which, Panofsky claimed, was a legal term implying a private marriage. Panofsky believed that the couple were in a bedroom (it is better described as a reception room), in the presence of two witnesses (they are more plausibly visitors), one of whom was van Eyck himself; and that the man was raising his right hand to take the matrimonial oath. Panofsky stated that his gesture was called fides levata and was a usual part of marriage ritual; but in fact he had invented the term.²⁷⁷ He also claimed that the single candle was a ‘marriage candle’ but was unable to produce any evidence to support his idea that candles were burned during marriage ceremonies. Concluding that the portrait was a ‘pictorial marriage certificate’ and that the signature was the attestation that ‘Jan van Eyck was here’ as one of the witnesses, he then proceeded to interpret various objects as disguised symbols. He claimed, for instance, that Arnolfini had discarded his shoes because he was standing on holy ground;²⁷⁸ but, as Arnolfini has merely taken off his pattens, which are for outdoor use, and is still wearing boots, Panofsky’s argument is unconvincing.²⁷⁹

Panofsky’s ingenious interpretation rests on a very slight basis in van Vaernewijck’s muddled and unreliable text. Schabacker and Dhanens were puzzled that the couple did not join their right hands to take the marital vows, for the union of right hands, the dextrarum iunctio, was indeed part of most marriage rituals; they therefore suggested that the picture represented a morganatic marriage.²⁸⁰ Smith rightly pointed out that, if a specific event was recorded, then the exact date should have been given. On three other occasions, van Eyck dated to the day portraits which make no allusion to any events or actions; yet in NG 186 he gave only the year.²⁸¹ Hall, successfully demolishing Panofsky’s and Schabacker’s arguments, preferred to see the picture as the representation of a betrothal.²⁸²

Again referring to van Vaernewijck, Hall sought to buttress his case by comparing NG 186 with miniatures from [page 200]manuscripts of Boccaccio’s Decameron. According to Hall, these miniatures are near‐contemporary representations of betrothals and show couples making the same gestures as the couple in NG 186. It is indeed true that in the French Decameron of c. 1415, now in the Vatican Library (Pal. lat. 1989), and in the derivative Netherlandish Decameron of the 1430s, now in the Arsenal Library in Paris (MS 5070), the illustrations for the seventh novella of the tenth day show Perdicone raising his right hand and taking Lisa’s right hand in his left hand.²⁸³ According to Hall, these miniatures represent the betrothal of Perdicone and Lisa and the gestures are ritual gestures. There is, in fact, no proof that the illuminators intended to depict the moment of betrothal. The miniatures may represent the introduction of Perdicone to Lisa and his right hand may be raised in greeting. There are other miniatures in both manuscripts which could equally depict betrothals but where the couples are making different gestures.²⁸⁴

In some manuscripts of the period, written instructions were included which guided the illuminators and which can be taken as accurate descriptions of their miniatures. Though directions, in Dutch, exist on some of the folios of the Arsenal Decameron, none refers to a miniature which may represent a betrothal. In a manuscript of the Grandes Chroniques de France (Bibliothèque municipale, Valenciennes, MS 637), which was illuminated in France in about 1400 and which afterwards belonged to Philip the Good, the instructions to the illuminators survive and may clarify the gestures of the couple in NG 186.²⁸⁵ The miniature on fol. 134 shows Charlemagne presenting his young son, Louis the Pious, to a bishop (fig. 25). Charlemagne raises his right hand in greeting and takes his son’s right hand in his left; Louis lays his left hand on his heart. The direction to the illuminator reads: ‘How the king, accompanied by two nobles, stands and holds his little son by the hand and presents him to the two bishops and the first blesses him and behind him there is a chaplain.’²⁸⁶ The illuminator has decided to omit the second bishop but has otherwise followed the instruction. By analogy, the man in NG 186 is raising his hand in greeting – the same gesture occurs in many Annunciations, where Gabriel greets the Virgin by raising his right hand – and is presenting the woman to the two visitors whose reflections are visible in the mirror.²⁸⁷ The man in blue is moving towards the couple and is a visitor entering the room, not a witness standing still to watch a ritual.

Since van Vaernewijck’s misleading reference to the painting may be disregarded, since the woman is not obviously pregnant, since the room is a reception room, since the gestures of the couple do not imply that they are being betrothed or married, there seems little reason to believe that the portrait has any significant narrative content. Only the unnecessary lighted candle and the strange signature provoke speculation.

The lighted candle in a room filled with natural light is not without parallels. Another candle burns above the fire in the Virgin and Child in an Interior (NG 6514) by a follower of Campin; and in van Eyck’s lost Bathing Women, where there must have been an open window, through which an extensive landscape was visible, there was also ‘a lantern in the bath chamber, just like one lit’.²⁸⁸ Van Eyck may have wanted to represent and contrast natural and artificial light both in the Bathing Women and in NG 186. One of Marmion’s greatest accomplishments was to have painted, in a picture which was clearly not a night scene, ‘a candle which seemed truly to be alight’.²⁸⁹ Alternatively, the candle has been lit in honour of the visitors. In Les Honneurs de la Cour, Aliénor de Poitiers described the conventions observed when great ladies of the Burgundian court were lying in. She stated that two silver candlesticks were kept, with other plate, on the dresser; in the candlesticks ‘there must be two large wax candles, to be lighted when someone comes to the chamber’.²⁹⁰ Similar conventions were no doubt observed on other occasions when visitors were received. Wax candles were expensive.²⁹¹ Burning a candle may have been a way of honouring a visitor, not only when a countess was lying in, but also on less exalted occasions.

The signature Johannes de eyck fuit hic/.1434. is unusual in its placing and in its form. According to Panofsky,²⁹² ‘the artist has set down his signature – lettered in the flourished script normally used for legal documents – as a witness rather than as a painter. In fact, we see him in the mirror entering the room in the company of another gentleman who may be interpreted as a second witness.’ The ‘flourished script’, however, was used for any elaborate or decorative inscription and does not necessarily have any legal connotations. The formula fuit hic or hic fuit was used in graffiti²⁹³ but never in legal documents, where the witnesses’ names were introduced by the words coram his testibus or presentibus N et N testibus.²⁹⁴ Panofsky’s interpretation of the signature is further weakened by the facts that the exact date is not given and that no reference is made to the ‘second witness’.

The meaning of the signature may not be altogether clear but, if van Eyck had wished to convey his meaning more directly, he would have chosen other words. He used the perfect tense, fuit, rather than the imperfect, erat; he must [page 201]have done so advisedly. If the inscriptions on his pictures are difficult to understand, they may be deliberately obscure or ambiguous. ‘Jan van Eyck has been here’ may be, at one level, an assertion of the truth of his representation and in that sense it would be in line with the ‘Loyal Remembrance’ on his Portrait of a Man, NG 290, and the ‘My husband painted me …’ on the ‘speaking likeness’ of his wife (Bruges). At another level, ‘Jan van Eyck has been here’ may assert that the image is not, after all, an accurate one, that the room is an invention and that ‘Jan van Eyck has been here’ – and only Jan van Eyck – because, as his subjects must very well have known, the interior is an illusion, the product of Jan’s imagination. Whatever interpretation is put on the inscription, it is an assertion of Jan’s skill in counterfeiting reality; and it seems to be a clear statement that he is the foremost of the two men reflected in the mirror. The signature is not, however, a statement that van Eyck witnessed a legally significant event. It is probably an indication that he was on friendly terms with the couple, who tolerated his rather blatant intrusion into their portrait and who understood the subtleties of his illusions and allusions.

The fact that Jan painted a second portrait of the same man tends also to suggest that the two were friends. If the man is indeed Giovanni di Nicolao Arnolfini, they would have been in contact both in Bruges and at the Burgundian court. The number of changes in NG 186 is probably further evidence of constant contact between the artist and his subjects, who must surely have approved every departure from the first design and underdrawing.

Most of the objects which, according to Panofsky and his followers, have symbolic significance, turn out not to have been underdrawn: the candle, the carving of Saint Margaret, the oranges, the discarded pattens, the dog. The gestures were altered during the course of painting – which suggests that they were not well‐rehearsed ritual gestures which had to be formally executed. The signature itself cannot have been envisaged in its present form until the chandelier had been added and the mirror had been reduced to its present size and shape. There are no drawn guidelines for the inscription, which is not quite horizontal but rises from left to right. A signature is often the last thing to be painted and this one may not have been planned very far in advance. The portrait was certainly not conceived according to any carefully worked out programme.

The versions are of no great use in elucidating the subject‐matter, since the composition was adapted for many different scenes. Loyset Liédet, who often borrowed from the painting, seems never to have used it for marriage or betrothal scenes. When he had to depict a private or clandestine marriage, he did not think to copy the double portrait.²⁹⁵ Only one among all the possible versions seems to refer to marriage: the double portrait at Godesberg (fig. 3). This does not represent a marriage, though it must allude to the marriage of the couple. The painter had some knowledge of the Arnolfini composition but he has not included the men reflected in the mirror, a candle or an inscription. He has, however, altered the gestures so that the couple join their right hands in the ritualised dextrarum iunctio and the man extends the index finger of his raised left hand in a movement which suggests that he, unlike Arnolfini, may be swearing an oath. Because the painter of the Godesberg picture had to change the composition radically in order to introduce allusions to marriage, his version, far from supporting the idea that NG 186 represents a marriage or a betrothal, provides strong evidence against such an interpretation.

It has been argued that there is a connection between NG 186 and a lost painting by Jan van Eyck known from two copies and depicting a nude woman with a clothed female attendant.²⁹⁶ The copies of the lost painting do indeed show an interior similar to that in NG 186 and also including a bed, a high‐backed chair, a window with six shutters, a convex mirror, a chest, a folding chair, discarded pattens and a small dog. Assuming that NG 186 represents a marriage, Held and others have argued that the lost picture depicted the ritual purification of the bride and that it was originally the cover of the portrait. If NG 186 does not represent a marriage, this argument loses its validity; it was in any case dubious, as NG 186 was probably designed to be protected by wings rather than a cover. Covers were suitable only for small paintings. In the lost picture, the bed‐curtains were not red and the mirror frame was undecorated. The resemblances between it and NG 186 were probably coincidental.

The hypothesis that NG 186 depicts a marriage or a betrothal may therefore be rejected. When the painting was first mentioned, in the 1516 inventory of Margaret of Austria’s collection, it was described as ‘Hernoul le Fin with his wife in a chamber’. Diego de Guevara, who almost certainly transmitted this description to Margaret of Austria, probably knew exactly what the portrait represented. It is only sensible to return to the 1516 description and to treat the painting as a portrait, without any significant narrative content, of Giovanni di Nicolao Arnolfini and his second wife.

The Interpretation of the Portrait

As a full‐length double portrait, NG 186 is not at all unusual.²⁹⁷ Such portraits occurred frequently in dynastic series, for example those of the Counts and Countesses of Flanders in the Aldermen’s Hall at Ypres,²⁹⁸ in the Chapel of the Count in the church of Our Lady at Courtrai (where only countesses in their own right were represented, with their husbands standing beside them)²⁹⁹ and in the Hall of the Aldermen of the Keure at Ghent (where, behind at least some of the counts, scenes from their lives were depicted).³⁰⁰ The Ypres series was in progress by 1323, when a count and countess were added; the Courtrai series was apparently begun in the 1370s by Jan van der Asselt and in 1406 Melchior Broederlam was paid for having added Philip the Bold and Margaret of Flanders; the Ghent series was commissioned in 1419 from Willem van Axpoele and Jan Martins. Copies survive of some of the figures from the Courtrai series but unfortunately Broederlam’s double portrait has vanished without trace.

NG 186 may be the earliest surviving portrait on panel of two persons who were not rulers and the earliest known portrait on panel where the subjects are seen in a domestic [page 202]setting.³⁰¹ Because vast numbers of works of art have been lost, it is impossible to say whether, in either respect, NG 186 was very unusual. It was not without precedent but it seems unlikely that many such portraits were produced before 1434. Only in the sixteenth century did this kind of portrait become relatively common.³⁰²

By 1516, NG 186 had shutters, which were still attached in 1700. According to the inventory of 1700, they were marbled; according to the inventories of 1516, 1523–4 and 1556–8, they were ornamented with the coat of arms and device of Diego de Guevara.³⁰³ He may have had his arms and device painted on wing panels that already existed; or he may have commissioned the shutters and had them attached to the portrait. The marbling could have been painted in 1434, in van Eyck’s workshop, or in about 1500, for Diego de Guevara. It was not unusual for a portrait to be protected by shutters. The Master of Frankfurt’s Self Portrait with his Wife (Antwerp), dated 1496, was the centre panel of a triptych. It retains its original frame, to which is still attached part of one of the four original hinges.³⁰⁴ Dürer’s portrait of Oswolt Krel (Munich), dated 1499, has wings representing wild men holding coats of arms;³⁰⁵ while Jakob Elsner’s small portrait triptych of Conrat Imhof (Munich, Bayerisches Nationalmuseum), dated 1486 but apparently painted in about 1500, shows on the left wing a coat of arms and on the right wing an allegorical figure.³⁰⁶ Most of the portraits listed in Margaret of Austria’s inventory of 1516 seem to have had shutters or covers – in this inventory the two terms were used interchangeably – and it is noted when a portrait lacked a cover or shutter: for example, ‘a medium‐sized picture of the face of a Portuguese lady which Madame had from Don Diego, done by the hand of Johannes, and it is done without oil on linen, without a cover or shutter’.³⁰⁷ The Master of Frankfurt’s Self Portrait with his Wife appears to be described in the same inventory: it certainly had shutters, but they were not mentioned, perhaps because they were not decorated. If they had been missing, their absence would have been noted.

It seems unlikely that a fashion for portrait triptychs arose only at the end of the fifteenth century and very much more probable that portraits, especially fairly large portraits, had always, as a matter of course, been equipped with shutters. In that case, the marbled wing panels once attached to NG 186 could have been original and Diego de Guevara merely had his arms and device added to existing shutters. The fact that van Eyck seems not to have decorated the reverse of NG 186 may support the theory that it was designed as the centre of a triptych. It was apparently usual to decorate the reverses of single panels but unusual to decorate the reverses of the centre panels of triptychs and polyptychs.³⁰⁸ Although the triptych form was also used for devotional paintings, it did not necessarily have religious connotations. The shutters were there to protect the portraits.

Opening the doors of the triptych‐portrait, the spectator would have confronted the signature, which is virtually an invitation to identify with van Eyck, reflected in the mirror and walking through the reflected doorway. It is perhaps the invitation into the picture which, together with the credibility of every detail, makes it so compelling an image. The candle flame, the reflection of the orange in the polished wooden casement, the mud on Arnolfini’s pattens are so accurately observed and rendered that it seems that the whole painting can be accepted as a faithful imitation of reality.

It is surprising, perhaps, to discover that the attention to detail is rather less sharp here than in van der Weyden’s Magdalen Reading (NG 654), where Saint Joseph’s beads may be compared with the beads hanging by the mirror, where the small figures in the landscape may be contrasted with the figures reflected in the mirror. Van Eyck is less careful, less literal, less interested in linear definition, more economical of his time and effort. Van der Weyden may have been able to lavish greater attention on detail because he included less incident. One of van Eyck’s supreme achievements was to hold in control, by means of his unrivalled mastery of composition, tone and colour, all the wealth of incident.

Some critics, seduced by van Eyck into taking his image too literally, have noticed that the perspective of his interior is not altogether accurate.³⁰⁹ Van Eyck may conceivably have known about the theories of perspective recently developed in Florence but, if he did, he chose to ignore them. He was of course perfectly able to represent correctly anything that was put before him, anything that he chose to imagine.³¹⁰ Even if he knew nothing of the theory of perspective, he has made a deliberate choice to distort the interior and the distortions have allowed him to construct a symmetrical composition around the central axis marked by the chandelier, his signature, the mirror and the vertical line between the two middle floorboards. He has also contrived to avoid overlaps so that each of the principal figures, each of the principal objects, is presented with the greatest possible clarity. As we have seen, the white fur on the woman’s dress is apparently of a colour and texture that could not have existed in 1434 and could only have been imagined.³¹¹ It seems very probable that flaws in other materials have been disguised or omitted to enhance the beauty of the surfaces. The room is in any case an imagined space, where logic of scale is not observed, although, as we have seen, it must bear some relation to the main reception room of Arnolfini’s house and its furnishings are probably recognisable representations of objects owned by the couple.

They themselves seem credible likenesses yet, in order to concentrate attention on their heads and hands, in order to enhance likeness by presenting each feature from the most suitable angle and in order to idealise slightly both his subjects, van Eyck, according to his usual practice, has adjusted the proportions of their heads and bodies. Arnolfini has a very large head and rather large hands, which agree in scale with his face but not with his arms. His shoulders are extremely narrow and the whole upper part of his body is severely diminished. His head is turned to our right. His face is apparently enlarged; his nose seems to be elongated and is turned too little into profile but is presented frontally, perhaps to disguise as much as possible the flaring nostrils. The ear is [page [203]][page 204]also seen frontally, to restrict its visibility. His wife, too, has a small thorax and short arms, though her hands are long and elegant. Her thumbs are impossibly long. Her head is distorted according to van Eyck’s usual system for female heads: while he normally diminishes the cranium of a man, he enlarges the forehead of a woman.

In the initial underdrawing, Arnolfini’s face and features are even larger and his features are higher (fig. 12). The head was then redrawn, with the distortions less pronounced, and the painting follows the revised underdrawing. His garment is shorter in the underdrawing and his legs are further apart. The alterations here make him look very much less ungainly. In the underdrawing, his wife’s features are lower and at slightly different angles; the height of her forehead is even more remarkable than in the painting (fig. 13). Van Eyck’s processes of distortion were very carefully considered but they were often moderated. The underdrawn heads of the Arnolfini couple were perhaps too obviously distorted to have been accepted as truthful likenesses.³¹²

It is very easy to fall into the error of thinking that van Eyck recorded with impassive objectivity all that he saw, to take his image too literally, to treat it as if it were a photographic record of a single instant. Van Eyck has succeeded in convincing his public that he was indeed here and that, as the doors of the triptych‐portrait opened and as his reflection passes through the reflected doorway, we can accompany him through the unseen door into the room and into the presence of Arnolfini and his wife. The couple are distorted and idealised, the room is an imagined space, the objects are arranged with marvellous artifice. Only after much thought and many changes of mind did van Eyck arrive at the immutable perfection of his image, so compelling and so powerfully engaging that it has never ceased to inspire fantasy. Its artificiality and above all the art that conceals its artifice are among the greatest triumphs of painting. Jan van Eyck was here and has persuaded us that we may follow him; but his image is so contrived, is so much the creation of his imagination, that, in truth, only the inimitable Jan van Eyck has been here.

General References

Friedländer, vol. I, pp. 40–1; Panofsky 1934; Davies 1954, pp. 116–28; Davies 1968, pp. 49–52; Dhanens 1980, pp. 193–205; Bedaux 1986; Hall 1994; Billinge and Campbell 1995; Paviot 1997.

Notes

For help with this entry, thanks are due to Rocío Arnáez, François Avril, Angus Fairrie, Noël Geirnaert, G. Houwen, Christiane Klapisch‐Zuber, Anne Korteweg, John Mills, Lisa Monnas, M. Nuyttens, Christine Paquet, Reinhard Strohm, Giorgio Tori, Anne van Buren, Stefaan Vande Cappelle, Elspeth Veale and Alison Wright. I am particularly grateful to Michael Bratchel, Christine Meek and Jacques Paviot, who have been exceptionally generous in sending information and advice.

Curator’s update

Lorne Campbell, September 2021

Further information has come to light regarding the painting’s location between 1794 and its ownership by James Hay. The picture remained where it was recorded in 1794, in a Retrete in the Cuarto del Rey in the Palacio Nuevo at Madrid, and was listed in two inventories taken during the reign of Joseph Bonaparte: in 1808, ‘[Otro «cabinet», aunque no lo especifica] J. de Holbein, peintre célèbre de Henry VIII: Un chevalier & une dame enceinte; à bien soigner, magnifiques, curieux’;³¹³ and in 1811, ‘OTRO GAVINETE DEL QUARTO DE DORMIR, 575 [871], Un hombre y una mujer embarazada agarrados por la mano HOLBEIN’.³¹⁴

In addition, derivations from the composition have been found in several more manuscript illuminations. These include two scenes in a copy of Vincent de Beauvais’ Miroir historial, illuminated by Willem Vrelant ( BN , MS fr 309): on fol. 179r in volume 2, showing the martyrdom of Saints Cosmas and Damian, the supervising figure repeats the pose of the man in the London painting, and wears his hat and tabard. On fol. 233r of the same volume, showing the martyrdom of St Margaret, the London painting’s dog appears in the foreground.

Further works by Loyset Liédet and his workshop which are derived from the London painting include the miniature of Pilate and his wife on fol. 10r of the first volume of a copy of Eustache Marcadé’s Mystère de la Vengeance de Nostre Seigneur Ihesu Christ in London ( BL , Add MS 89066/1). Here, Pilate’s wife is posed in front of a high-backed chair and a bed in a manner which distinctly recalls, but does not literally copy, the London painting; the same is true of the reflection in the circular mirror in the illumination, and – to a lesser degree – of the three oranges which are rather randomly placed on the bench and the column supporting the canopy in the miniature. The setting in which Pilate and his wife appear is very similar to the rooms represented in the Louvre Annunciation, attributed to a follower of van der Weyden (fig. 3 on p. 27), and the miniature of Joscelin II, Count of Edessa, attributed to Dreux Jehan, in the Chroniques de Jérusalem abrégées ( ÖNB , Cod. 2533, fol. 15; fig. 2 on p. 26). As discussed on pp. 178–179, it is possible that Liédet, and Petrus Christus, had access to a more elaborate version of the Arnolfini composition where a more spacious and credible interior was depicted. This could have been a preliminary drawing. Like Vrelant, however, Liédet seems also to have known the finished painting or a faithful copy.

Notes