The Trinitarias carpet: early masterpiece or modern reproduction?


The Trinitarias carpet (fig. 1) in the collection of the National Gallery of Victoria presents a series of challenges to the carpet scholar: it has an intriguing provenance story, but like those of many important early carpets, one based to an extent on legend rather than on documentation of its early history; it poses an interesting question of place of manufacture, an issue that is currently receiving considerable attention by scholars; and finally, it was recently exhibited in Australia with an assigned date that in effect branded it as either a later reproduction or even a forgery.1 See Pathways through Paradise: Oriental Rugs from Australian Collections (exh. cat.), Powerhouse Museum/Oriental Rug Society of NSW, Sydney, August 2004–January 2005, pp. 9–10, 64–5. All of these factors have focused attention on the Trinitarias, which is certainly the most important early carpet in an Australian museum and which was, until the recent publication of an article by Roger Leong, Bronwyn Cosgrove and Solitaire Osei in Hali magazine, relatively unknown to overseas scholars and the general public.2 See Roger Leong, Bronwyn Cosgrove & Solitaire Osei, ‘Unrolling a hidden treasure: the Trinitarias carpet at the National Gallery of Victoria’, in Hali, no. 168, 2011, pp. 86–7. I am much indebted to Roger Leong for the opportunity to examine this carpet, and to Solitaire Osei and Yvan Maes De Wit and his staff for facilitating my examination of the carpet in Mechelen. The present discussion of the Trinitarias carpet will focus on these three issues in both a specific and a broader context. First, how does our information on the provenance of the Trinitarias carpet reflect on the provenances assigned to great early carpets in general? Second, in light of the most recent scholarship, why do we presently assign the Trinitarias carpet to India rather than to Iran? And third, what should be our view of some recent ‘exposures’ of carpets long thought to be products of the classical age of production in the sixteenth and seventeenth centuries as later products?

Examination of the carpet

The author examined the Trinitarias carpet on 7 August 2010 at the De Wit laboratories in Mechelen, Belgium.3 The Royal Manufacturers De Wit (est. 1889) is a private tapestry workshop in Mechelen, Belgium, which today specialises in the cleaning, conservation and restoration of tapestries and large carpets. Two members of the De Wit staff were present, along with textiles conservator Solitaire Osei from the National Gallery of Victoria. Mr Yvan Maes De Wit, director of De Wit, briefly joined us. Preparatory to the actual examination of the carpet, I had received a number of extremely helpful documents from the NGV. These included a set of high-resolution digital photographs of the carpet taken in Australia, a report on dye analysis conducted by Dr Recep Karadag at the Marmara University dye analysis laboratories in Istanbul, and the account of a radiocarbon analysis undertaken on 31 March 2010. In addition, I received references or complete copies of a number of publications on, and published references to, the carpet, as well as a preliminary and very useful structural analysis of the carpet made by Susan Scollay4 Susan Scollay, ‘Appraisal of room-sized Oriental carpet, known as the Trinitarias carpet, in the collection of the National Gallery of Victoria, Melbourne’, 22 February 2010. and, from independent scholar Michael Franses in London, some more documentation of its twentieth-century history.

The bottom half of the carpet (that including the end from which the weaving process began) was unrolled in the washing area when I arrived. Most of the analysis work was done on an accessible area half the length of the carpet on its lower left side, about a metre into its width (fig. 2). While this was not an ideal situation for carpet examination, because of the huge size of the carpet and the available space in the De Wit laboratories, it sufficed for a careful examination of almost a quarter of the carpet’s entire fabric. A search in this area for jufti knotting (knotting over four rather than two warps) and irregular knot patterns (such as offset rows of knots), both of which would have contradicted an Indian provenance, produced nothing. A few spots of candle wax were noticed that had escaped the massive de-waxing operation undertaken in Melbourne before shipping the carpet to Belgium. The carpet was examined with a medium-power optical loupe, and 12MP RAW and JPG digital photographs were taken with a Nikon D3 camera equipped first with a 24–70mm zoom lens (for general views), and second with a 60mm micro lens (for close-ups). Additional 12MP JPG photographs were taken through the loupe itself with a Canon G9 digital camera. Measurements and observations for every aspect of the analysis were made in several different places along the half-length of the left side of the carpet; minor differences were averaged.

Description

The carpet is a very large (approximately 10.44 x 3.36 metres) oblong rectangle in medallion layout, with a ratio of length to width of about three to one. The field of the carpet consists of horizontal rows of palmette, rosette and cloud-band motifs on a dark purple-red ground. The sixteen-lobed dark blue ground medallion consists of a central white ground, eight-pointed starlike form with a wide border on a dark blue field filled with split-leaf and floral palmette motifs. The medallion is surrounded by a sixteen-lobed red ground border and a white ‘cloud collar’ (fig. 1 detail). The medallion is not quite centred in the carpet; rather, its centre is slightly toward the top end. Two pendant elements, each surrounded by the same cloud collar, project from the medallion toward the top and the bottom of the carpet. The design of the field is in theory infinitely repetitive, cut arbitrarily by the border of the carpet at the top, bottom and both sides; its lateral symmetry coupled with variant knot counts within identical forms suggests that a paper cartoon rather than a knot plan was used by the weavers.

The main border of the carpet has a white ground ornamented with small floral palmettes and cloud bands; it is decorated with large oblong red-ground cartouches alternating with smaller eight-pointed ochre-ground medallions, these motifs linked in turn by even smaller eight-pointed light-blue-ground stars (fig. 3). The corner resolutions of the major border are arbitrary; the bottom border was planned from the centre outward beginning with a cartouche, resulting in ‘Siamese twin’ corner motifs based on the small light-blue-ground stars; at the top of the carpet, the weavers apparently ran out of loom before being able to finish the central field symmetrically; the corners consist of small improvised forms based on the large cartouches. This corner articulation strongly suggests again that the entire design was based not on a knot plan but on a paper cartoon, a fairly common occurrence in early carpets.

The inner guard border is a yellow-ground stripe ornamented with small lotus palmettes. The outer guard border consists of yellow (pointing in) and red (pointing out) reciprocal trefoil forms. The corner articulations are arbitrary rather than planned in advance.

Structure

Warp (4 tested): All slightly brittle Und white cotton, eight Z-spun strands of highly variable thickness, plied S alternate warps moderately depressed.

Weft: All visible weft was blue cotton, one Z-spun yarn. The majority of the weft patterns were 2// + 2// +2//; that is, three shoots of two parallel wefts each between every row of knots. A few instances of 2// + 2// were observed (fig. 4).

Pile: Wool; dark yellow ochre, medium yellow, dark red, medium red, light red (pink), dark green, light green, light blue, medium blue, dark blue, medium brown, corrosive black/brown, undyed white (13 colours). Pile colours exhibit overall slight variation in horizontal banding (abrash). The dark brown exhibits the normal corrosion and consequent wear associated with older carpets; there is no indication that there has been any clipping or wheel-grinding of the brittle brown pile to simulate ageing. All pile yarns observed were composed of 3 Z-spun yarns plied S.

Knot: Asymmetrical open to the left. 47–49 V x 49–50 H; 230–2450 per dm2. Square-inch equivalent: approximately 120–130 kpsi.

Edges: No original edges observed on the part of the carpet I was able to examine; in old carpets the edges are frequently replaced or rewrapped.

Ends: The bottom end I examined was very slightly stripped, but no original end finish was visible.

Handle: The tactile experience of the carpet is closely similar to that of other large carpets of similar scale, age, and design; the handle is fairly dry and the carpet is stiff and substantial.

Dye analysis undertaken by Recep Karadag in Istanbul confirmed the use of traditional dyes from the seventeenth century in the carpet.5 The Marmara University laboratory for the study of carpets, founded by Dr Harald Boehmer and today directed by Dr Karadag, relies on paper chromatography for the bulk of its tests of dyestuffs. No tested colours provided any data that might cast doubt on the traditionally assigned date of around 1600 for this carpet. Carbon-14 analysis undertaken at the order of the NGV, while of limited reliability, also confirmed an early date for the carpet. The structural analysis undertaken in Mechelen effectively agreed with that undertaken by Susan Scollay in Melbourne. The implications of all of this data will be discussed below.

The issue of provenance

What’s in a name? It appears that a good deal of importance resides in the names given to carpets. Because neither the name of the designer nor, in many cases, the place of manufacture of some of the world’s most splendid and famous Islamic carpets are known, we associate them instead with great royal patrons, palaces or shrines or, barring these, with important subsequent possessors. The Emperors’ carpets, the Anhalt carpet, the Schwartzenberg carpet, the Ardabil carpets – all of these names reflect to some extent the combination of wishful thinking and shrewd salesmanship long present in the carpet marketplace. In determining provenance, one might posit a formula of reliability that parallels that of the efficacy of a camera’s flash device: as light diminishes in proportion to the square of the distance from the light source, so the reliability of provenance information in many cases diminishes in proportion to the square of the number of years of history involved.

In the case of the Trinitarias carpet, working backward, we know that it came to the NGV in 1959, due to the generous support of the Felton Bequest. Before that it was in the hands of a Scottish dealer and carpet manufacturer, James Templeton & Company of Glasgow. It came to Templeton’s after having been shipped to Canada during the Second World War; before that it was in the hands of London dealers Vitali Benguiat and Perez & Company. Perez exhibited the carpet as an ‘Isfahan Carpet’ of the seventeenth century. Perez’s typed house fact sheet on the carpet, generously supplied by Michael Franses, states that the earliest mention of the carpet in the convent records was in 1699, and that the carpet ‘is reputed to have been presented by Philip IV at the beginning of the 17th (sic) century’. The extremely good condition of the carpet was explained as follows: ‘During its long stay in the Spanish convent, the carpet was only brought out on special feast days and, because of that, it is in a remarkable state of preservation’.

It was supposedly exhibited in the Ibero-American Exposition in Seville in 1929; at this time it is not certain whether the carpet was still owned by the Trinitarias convent in Madrid, but by 1930, with the Spanish Civil War looming and the world in deep economic depression, it was apparently already on the market in London. In 1931 the official gazette of the Spanish government published a list from the Ministry of Public Instruction of dispersed works of Spanish art in foreign museums; the list included: ‘Madrid. Las monjas Trinitarias: Una alfombra persa importante’ (an important Persian carpet).6 Published on the eve of the Spanish Civil War, this article chronicles the dispersal of items of Spanish cultural heritage, a phenomenon which of course afflicted a number of European countries affected by inflation in the decade after the First World War, and then in the aftermath of the stock market crash of 1929 (‘El ministro de instruccion publica habla de la desaparicion de obras artisticas, entregando una nota de las desaparecidas’, ABC, Edicion de Andalucia, 4 July 1931, p. 22).

Since the carpet apparently resided in the Trinitarias convent for several centuries, the date of its arrival there is somewhat murky. A romantic view is put forward in the Perez document: although stating that the earliest mention of the carpet in Madrid was in 1699, that document presents the theory that the carpet was given to the convent by the King of Spain, Philip IV (r. 1621–65), who is perhaps best known today for the portraits of himself and his family painted by Velásquez, and his purported association with the carpet undoubtedly carried some cachet. A document discovered in the 1920s in the convent, however, mentions a more prosaic origin for the gift. The document was discussed in 1926 in an article  by a Spanish scholar, Elías Tormo, written at a time when the sale of the carpet was already being contemplated and publicly debated in Spain; the article by Tormo was cited in the recent Hali article on the Trinitarias carpet mentioned above.7 Leong et al., p. 89.

Tormo’s article, entitled ‘Trinitarias Descalzas, Madrid, alfombra en venta’ (‘Convent of the Barefoot Trinitarians, Madrid, carpet up for sale’)8 Elías Tormo, ‘Trinitarias Descalzas, Madrid, alfombra en venta’, in Boletin de la Real Academia de la Historia, vol. 88, April–June 1926, pp. 477–83. With thanks to Pierre Arpin, Director, Museum and Art Gallery of the Northern Territory, for discovering this article., is a presentation to the Royal Academy of History in Madrid arguing against the sale of the carpet, for which an offer of 300,000 pesetas had been made. Tormo’s article includes an exhaustive description of the carpet, the recent history of offers for it, and an impassioned argument against its sale. Tormo had examined the convent’s own records and discovered a late seventeenth-century document mentioning that the carpet indeed came to the convent in 1699 as the gift of a certain Don Juan de Guzman on 5 February 1699 to honour Sister Josefa de la Encarnacion, evidently a relative of his who had entered the convent at this time. While we have not yet been able to confirm this assertion by an examination of the convent’s records, its publication in a respected journal, together with the overall context of the article’s research, argues for its authenticity. Many convent records in Spain perished during the Civil War (1936–39), so it is not certain that this document can be recovered.

The issue of place of manufacture

India or Iran? This question is increasingly important as we begin to understand the full implications of new discoveries in carpet history and to gain new insights as we examine the technique of carpets. Long a major question revolving around the so-called Indo–Isfahan carpets that appear in vast numbers in the seventeenth century and are documented in numerous European paintings, the question has recently taken on an even more complicated meaning as we begin to realise that a goodly number of early carpets in collections around the world, some of which had previously been attributed to northern India in the seventeenth century, and others to a mysterious unbekannte Gruppe (unknown group) known largely only through representations in European oil paintings, were probably woven in the Deccan in south-central India.9 Brigitte Scheunemann, ‘Eine Unbekannte Teppichgattung’, in Kunst des Orients III, Franz Steiner Verlag GmbH, Wiesbaden, 1959, pp. 79–89, first isolated a group of carpets now thought to be from the Deccan. In his Flowers Underfoot: Indian Carpets of the Mughal Era, Thames & Hudson, New York, 1997, now something of a standard work in the field, Daniel Walker discussed the same group under the rubric ‘Kyoto group’ of carpets (pp. 136–46). In her groundbreaking 2011 PhD dissertation for the Institute of Fine Arts, in process of revision for publication, Dr Yumiko Kamada of Waseda University added carpets to this group from other Japanese cities, and uncovered extensive documentation about their origins in trade from the Deccan to Japan carried out by the VOC or Dutch East India Company.

The popular conception that serious carpet weaving did not begin in India until the late sixteenth century after the re-establishment of the Mughal dynasty with the return of the Mughal Padshah (emperor) Humayun from exile in Persia in 1555, is now open to question. Recent scholarship has focused on the Deccan as a place of carpet production, beginning at least as early as the early sixteenth century. Early carpets attributed to the Deccan include the famous silk Octagon Holbein carpet now in Qatar; another fragmentary Holbein-design carpet formerly in the possession of the German dealer Bertram Frauenknecht; as well as all carpets of the seventeenth-century unbekannte Gruppe; a significant number of carpets of that design group with early provenance documentation in the collections of Japanese guild organisations; and a number of carpets with north Indian designs but with structural and design characteristics that may argue for a south Indian provenance.10 Michael Franses, in a recent article (‘Ashtapada’, Hali, no. 167, Spring 2011, pp. 80–9) proposed a Deccani origin for the silk carpet with a central ‘large-pattern Holbein’ octagonal medallion, now in Qatar. The unpublished Frauenknecht fragment, also using a Turkish design, corresponds closely in technique to the bulk of the carpets in Japanese collections now attributed to the Deccan. On technical grounds Dr Kamada has also suggested that the well-known Morgan animal carpet in the Metropolitan (17.190.858), whose designs are indisputably of Mughal court origin, as well as others of its group, may well have been woven in south-central India rather than in Lahore. In the seemingly endless and inconclusive arguments about one or another provenance, a consensus seems to be emerging that focuses on a few variables. The first involves warp structure; a number of scholars have proposed that as a rule of thumb the inclusion of four plies or fewer in a cotton warp militates toward a Persian provenance (with certain possible exceptions, such as the ‘garden carpet’ group), while the use of more plies, especially the use of eight or more, characterises Indian production. Clearly this rule of thumb places the Trinitarias carpet, with its warp consisting of eight Z-spun cotton yarns plied S, in India, just as it places the so-called Indo–Isfahan carpets in Persia.11 See Steven Cohen, ‘Indian or Persian?’, Hali, no. 134, May–June 2004, pp. 91–3; and Daniel Walker, ‘Technical characteristics’, in Flowers Underfoot, Metropolitan Museum of Art, New York, 1998, pp. 21–8.

A second variable involves the use of what is referred to as ton-sur-ton in carpet design – that is, the juxtaposition in parts of the design of areas of knotting dyed in two values of the same hue. Indian carpet designers are thought to have been far more likely to have used areas of pile restricted to two closely allied colours – for example, a dark blue and a light blue, or a dark red and a light red – created by varying either the length of time the wool was placed in a dye-bath (a longer immersion producing a darker lac red, and a shorter immersion producing a lighter, pinker value) or, in the case of blue, the number of times the wool was repeatedly dipped in an indigo bath. There is an extensive use of ton-sur-ton in the Trinitarias carpet, which constitutes one of the strongest arguments for an Indian provenance (fig. 5).12 See Daniel Walker’s discussion of ton-sur-ton, in Walker, 1998, especially pp. 26 & 67.

A third variable, with limited application, and far more subjective in nature, is the alleged tendency of some carpets thought to be of north Indian proven-ance to employ very expressive, even flashy, designs, such as the use of highly developed ‘cloud-collar’ forms around central medallions. A fine example of this type of design is the large medallion carpet in the Gulbenkian Collection in Lisbon. Some see the Trinitarias carpet as including some of these Indian elements in its overall Persianate design character, especially in the central medallion itself.13 The Gulbenkian carpet is illustrated in Walker, 1998, pp. 62–3.

European provenance might provide another indication of place of manufacture. While in the later seventeenth and eighteenth centuries there is a substantial documentation of the import of Indian carpets into Europe (and also into Japan) through the Dutch East India Company, in the sixteenth and early seventeenth centuries the Portuguese were major players in the Indian trade, having set up communities on the west coast of India in the sixteenth century shortly after Vasco da Gama’s discovery of the sea route to India around the Cape of Good Hope. Portuguese–Spanish trade in luxury goods can be easily documented within the Iberian peninsula at this time. Thus it could be argued that the Spanish element of the carpet’s provenance militates in favour of an Indian place of manufacture.14 On the carpet trade, see Walker, 1998, ‘International commerce’, pp. 15–20.

Of course, all of these ‘litmus tests’ are still debatable, still subject to significant exceptions and thus still far from attaining widespread acceptance by scholars. This does not mean that we do not have early carpets that can be firmly assigned to India; there are many carpets that are unquestionably assigned to India on the basis of materials or design, just as there are many old carpets whose Persian provenance is beyond doubt. But this still leaves a large group of definitely Persianate carpets such as the Trinitarias in the realm of uncertainty.

Is this situation unusual? As stated above, proven-ance takes on a special significance for art historians in the study of the so-called decorative arts, in which the absence of artists’ names, lack of specific patron documentation and widespread commercial dissemination of works of art have always been common. In carpet history, the place of design and actual place of weaving often vary widely. Today India produces commercial carpets, carpet-like textiles and textiles with designs taken from carpets whose patterns and colours come from every corner of the world. We should therefore not be shocked, disturbed or puzzled if this phenomenon were to have existed in the distant past.

The issue of authenticity

In one respect the Trinitarias carpet is unfortunately in very distinguished artistic company. In the past century a number of important early carpets in major institutions, among them the Ardabil carpet, one of a pair, in the Los Angeles County Museum of Art; the Anhalt carpet in the Metropolitan Museum of Art; and the Salting carpet in the Victoria and Albert Museum, and its technically and design-allied group, have come under a cloud of suspicion as either outright forgeries or simply as very late artistic products in an early style. As early as the 1950s the German art historian Kurt Erdmann proposed that a number of prayer carpets in the Topkapi Palace in Istanbul, along with larger carpets of the so-called Salting group, were ‘Persischer Teppiche Turkischer Provenanz’ (Persian carpets of Turkish provenance) that Erdmann believed had been woven somewhere in Turkey in the late nineteenth century.15 Kurt Erdmann’s views on the so-called Salting carpets are sum-marised in his article ‘Persian carpets of Turkish provenance’, in Kurt Erdmann, Seven Hundred Years of Oriental Carpets, University of California Press, Berkeley & Los Angeles, 1970, pp. 76–80. Examination of these attributions may have useful implications for our study of the Trinitarias carpet.

What were the bases for these evaluations of well-known carpets assigned to the nineteenth century in the collections of major institutions? In the case of Mr Getty’s Ardabil carpet now in the Los Angeles County Museum of Art, the questioning came in a (thankfully never published) paper presented at the meeting of a scholarly organisation in California in the 1980s. The presenter, relying exclusively on the publication on the carpet by Rexford Stead,16 Rexford Stead, The Ardabil Carpets, Getty Trust Publications, Los Angeles, 1987. Sheila Blair, relying perhaps too much on Stead, discussed the carpets from a technical point of view in ‘The Ardabil carpets in context’, in Andrew J. Newman (ed.), Society and Culture in the Early Modern Middle East: Studies on Iran in the Safavid Period, Brill, Leiden, 2003, pp. 125–43. but never having examined the carpet itself, used an (unreliable) published structural analysis and even more unreliable black and white photographs of the carpet’s inscription in Mr Stead’s little volume for her arguments. She was unaware of a recent accomplished and exhaustive conservation survey of the carpet conducted by LACMA conservators that, among other things, recorded structural damage and repair to the area of the carpet bearing the inscription. These matters were brought up in a vigorous discussion following the presentation of the paper, and since that time the reputation of the Los Angeles carpet has not been further impugned. A more recent examination of the apparent dramatic difference in knot count between the Los Angeles and London Ardabil carpets as recorded in Stead’s analysis versus that of the Victoria and Albert Museum convincingly explains the differences as a result of a single measurement taken in too small an area of the Los Angeles carpet; and points out that in both Ardabil carpets the knot density of individual square inches may vary enormously within a small area – the weave is in fact not uniform at all in either example.17 See Walter B. Denny, ‘Anatolia, Tabriz, and the carpet design revolution’, in Jon Thompson, Daniel Shaffer & Pirjetta Mildh (eds), Carpets and Textiles of the Iranian World 1400–1700, The Ashmolean Museum, Oxford University, in association with the Bruschettini Foundation for Islamic and Asian Art, Genoa, Oxford 2010, especially p. 66.

In the case of the Anhalt carpet, given to the Metropolitan Museum of Art in 1946 by the Kress Foundation, the criticisms came from three sources. The late Charles Grant Ellis, an American scholar, communicated verbally to colleagues his belief that the carpet was modern, based largely on its unusual employment of a golden yellow ground. The late May Hamilton Beattie, a British scholar, communicated verbally to her colleagues her belief that the carpet was modern based on its very good condition, in which it differs markedly from almost all other known examples of sixteenth-century Persian carpet weaving. However, when she finally published her opinion of the carpet in Encyclopedia Iranica, shortly before her death, she accepted it as genuine, noting that it had probably been rolled up in storage for much of its European history and thus had received little wear.18 May Hamilton Beattie, ‘The Anhalt carpet’, in Ehsan Yarshater (ed.), Encyclopedia Iranica, Center for Iranian Studies, Columbia University, New York, online at . See also Walter B. Denny, ‘Islamic art: forgeries’, in The Dictionary of Art, Grove, New York, 1996, vol. 16, p. 545; and the same author’s review of Donald King & David Sylvester (eds), ‘The Eastern carpet in the Western world’, in Middle East Studies Association Bulletin, vol. 18, no. 2, 1984, pp. 244–5. Friedrich Spuhler, a distinguished and highly regarded German scholar, communicated verbally to his colleagues, and in writing in a detailed memorandum to the Metropolitan Museum of Art, his belief that the carpet was modern, produced in an unknown centre in Persia in the nineteenth century. His argument was based largely on small details of its design that he believed would never have been included in a sixteenth-century carpet. He asserted that these details constituted weavers’ mistakes that would never have been tolerated in the golden age of Persian carpet weaving. On the other hand, there is no documentation whatsoever that supports the notion that carpets of this size and classical technique were woven anywhere in Iran during the nineteenth century. Dye analysis of the carpet, a highly detailed examination of its structure, condition and of the almost complete corrosion and disappearance of its dark brown pile and, most importantly, significant changes in our perceptions of the way early Persian carpets were designed, woven and artistically evaluated, have all prompted the Metropolitan Museum of Art to confirm with confidence the original attribution of the carpet to sixteenth-century Safavid Persia.19 In the last several years the carpet underwent intensive examina-tion and, at last, analyses of all of the dyestuffs were undertaken. They are all traditional. Under the modern replacement of all of the original dark brown wool, the original highly corroded pile is easily visible under microscopic examination. Questions of provenance once obscured by the political division of Germany may provide new evidence concerning the carpet’s age. A study of the Anhalt carpet, which will include a great deal of recently uncovered evidence, is currently being prepared for publication in order that the question of the authenticity of the carpet can be publicly examined and evaluated by scholars worldwide.20 Because the Anhalt carpet’s case is such an interesting one when looking at the question of authenticity in carpets, it is the author’s intention to publish a detailed study of the carpet and its criticism within the next several years.

The case of the so-called Salting carpets is the oldest and, in some respects, the most complex of the three cases under discussion. The brilliant colouration of the Istanbul examples, their extraordinarily good condition and the lack of early documentation of such carpets in European collections led Erdmann to the extraordinary conclusion that the carpets were woven in a sixteenth-century Persian style in Turkey in the nineteenth century. Subsequently, most of Erdmann’s questions have been explained; the Istanbul carpets, mentioned in documents as gifts of the Shah of Persia to the Ottoman sultans in the sixteenth century, were almost certainly not exposed to light or used by the sultans for reasons of differences of taste and religious doctrine; dye analysis and the usual isotope-decay tests, whatever the latter’s usefulness, all confirm an early date for the majority of the Salting group.21 The Erdmann attribution of the Salting carpets to nineteenth-century Turkey (see note 12) led to an unfortunate ‘stampede effect’ in the literature. Charles Grant Ellis, Friedrich Spuhler and May Hamilton Beattie concurred, as did this author and many others; all of the Topkapi prayer rugs fell under the same pall. The matter was finally addressed more clearly in a series of articles published in Murray Eiland Jr & Robert Pinner (eds), Oriental Carpet and Textile Studies, vol. v, part 2, ‘The Salting carpets’, ICOC, Danville CA, 1999. Most importantly, there is no nineteenth-century documentation whatsoever for a production of carpets in this style and technique on the Ottoman Empire, in the Kumkapi district of Istanbul, in Hereke or in any other location.

The attitude of carpet scholars toward the phenomenon of good condition – of a very old carpet surviving in a very good state of preservation – so important in the discussions of the Salting, Anhalt and Trinitarias carpets, has undergone a major change in recent years. The discovery in the Pitti Palace guardaroba in Florence of two early carpets in almost pristine, newly woven condition has affirmed the fact that very old carpets do occasionally survive in excellent condition;22 See Alberto Boralevi, ‘Three Egyptian carpets in Italy’, in Robert Pinner & Walter B. Denny (eds), Oriental Carpet & Textile Studies II: Carpets of the Mediterranean Countries 1400–1600, ICOC, London, 1986, pp. 205–20. important textiles and carpets were sometimes stored away carefully and either not used or rarely used, sometimes for reasons of change or differences in taste, sometimes (as in the case of the Topkapi prayer rugs with Shi’ite inscriptions) due to the theological implications of their inscriptions, or occasionally because their large size kept them from practical use.

The Trinitarias carpet authenticated

Reattribution of an important work of art to a later period, or any suggestion that the work may be a forgery, is obviously a touchy matter. Scholars are understandably reluctant to offend a museum or influential owner with an attribution that might lessen the artistic significance or monetary worth of a work of art. As a rule of thumb, such matters should receive full scholarly scrutiny; ordinarily the re-attribution of a carpet such as the Trinitarias to a much later period should be accompanied by a citation of whatever evidence caused this re-attribution to occur. All the evidence for the carpet’s authenticity, together with the complete lack of documentation for any production of such a carpet in late eighteenth- or nineteenth-century India, force us to regard the attribution made in Sydney a few years ago as without substance.

The Trinitarias carpet defined

The Persianate design of the Trinitarias carpet places it in a distinguished group of medallion carpets now thought to have been woven in India in the seventeenth  century; the Persian taste remained strong throughout the Mughal period, despite the production in India of carpets in more typically Mughal designs. The use of blue weft, which gives the back of the Trinitarias a ‘Persian look’, is probably a tribute to the enduring popularity of the Persian taste in India, but the technical use of three pairs of wefts is a strong indication of Indian provenance. The excellent condition of the Trinitarias carpet is most likely the result of its careful preservation by the sisters of the convent in Madrid and the custom of using it only on very special occasions.

The evidence surrounding the Trinitarias carpet – its technique, colouration, design, as well as its Spanish history – all tend to militate toward a seventeenth-century Indian provenance and, for the present, the best attribution for the carpet remains India, probably northern India, in the first half of the seventeenth  century, during the reigns of the Mughal emperors Jahangir (1605–1627) and his son Shah Jahan (1628–1658). While the use of the words ‘probably’ and ‘approximately’ in the ‘tombstone’ descriptions of works of art is often frowned upon by museums, we should accept that a degree of uncertainty is often the norm rather than the exception in the history of art, and adjust our prose accordingly, despite the pressure from dealers, auctioneers and lawyers to err on the side of certainty.

Because of its good condition, complex and fluent design and powerful visual impact, the Trinitarias carpet may be justly regarded as one of the great artistic treasures of the National Gallery of Victoria, a major cultural asset for Australia and an important part of the world’s carpet heritage.

Walter B. Denny, Professor of Art History, Adjunct Professor of Middle Eastern Studies, University of Massachusetts (in 2012).

Notes

 

1   See Pathways through Paradise: Oriental Rugs from Australian Collections (exh. cat.), Powerhouse Museum/Oriental Rug Society of NSW, Sydney, August 2004–January 2005, pp. 9–10, 64–5.

2   See Roger Leong, Bronwyn Cosgrove & Solitaire Osei, ‘Unrolling a hidden treasure: the Trinitarias carpet at the National Gallery of Victoria’, in Hali, no. 168, 2011, pp. 86–7. I am much indebted to Roger Leong for the opportunity to examine this carpet, and to Solitaire Osei and Yvan Maes De Wit and his staff for facilitating my examination of the carpet in Mechelen.

3   The Royal Manufacturers De Wit (est. 1889) is a private tapestry workshop in Mechelen, Belgium, which today specialises in the cleaning, conservation and restoration of tapestries and large carpets.

4   Susan Scollay, ‘Appraisal of room-sized Oriental carpet, known as the Trinitarias carpet, in the collection of the National Gallery of Victoria, Melbourne’, 22 February 2010.

5   The Marmara University laboratory for the study of carpets, founded by Dr Harald Boehmer and today directed by Dr Karadag, relies on paper chromatography for the bulk of its tests of dyestuffs.

6   Published on the eve of the Spanish Civil War, this article chronicles the dispersal of items of Spanish cultural heritage, a phenomenon which of course afflicted a number of European countries affected by inflation in the decade after the First World War, and then in the aftermath of the stock market crash of 1929 (‘El ministro de instruccion publica habla de la desaparicion de obras artisticas, entregando una nota de las desaparecidas’, ABC, Edicion de Andalucia, 4 July 1931, p. 22).

7   Leong et al., p. 89.

8   Elías Tormo, ‘Trinitarias Descalzas, Madrid, alfombra en venta’, in Boletin de la Real Academia de la Historia, vol. 88, April–June 1926, pp. 477–83. With thanks to Pierre Arpin, Director, Museum and Art Gallery of the Northern Territory, for discovering this article.

9   Brigitte Scheunemann, ‘Eine Unbekannte Teppichgattung’, in Kunst des Orients III, Franz Steiner Verlag GmbH, Wiesbaden, 1959, pp. 79–89, first isolated a group of carpets now thought to be from the Deccan. In his Flowers Underfoot: Indian Carpets of the Mughal Era, Thames & Hudson, New York, 1997, now something of a standard work in the field, Daniel Walker discussed the same group under the rubric ‘Kyoto group’ of carpets (pp. 136–46). In her groundbreaking 2011 PhD dissertation for the Institute of Fine Arts, in process of revision for publication, Dr Yumiko Kamada of Waseda University added carpets to this group from other Japanese cities, and uncovered extensive documentation about their origins in trade from the Deccan to Japan carried out by the VOC or Dutch East India Company.

10 Michael Franses, in a recent article (‘Ashtapada’, Hali, no. 167, Spring 2011, pp. 80–9) proposed a Deccani origin for the silk carpet with a central ‘large-pattern Holbein’ octagonal medallion, now in Qatar. The unpublished Frauenknecht fragment, also using a Turkish design, corresponds closely in technique to the bulk of the carpets in Japanese collections now attributed to the Deccan. On technical grounds Dr Kamada has also suggested that the well-known Morgan animal carpet in the Metropolitan (17.190.858), whose designs are indisputably of Mughal court origin, as well as others of its group, may well have been woven in south-central India rather than in Lahore.

11 See Steven Cohen, ‘Indian or Persian?’, Hali, no. 134, May–June 2004, pp. 91–3; and Daniel Walker, ‘Technical characteristics’, in Flowers Underfoot, Metropolitan Museum of Art, New York, 1998, pp. 21–8.

12 See Daniel Walker’s discussion of ton-sur-ton, in Walker, 1998, especially pp. 26 & 67.

13 The Gulbenkian carpet is illustrated in Walker, 1998, pp. 62–3.

14 On the carpet trade, see Walker, 1998, ‘International commerce’, pp. 15–20.

15 Kurt Erdmann’s views on the so-called Salting carpets are sum-marised in his article ‘Persian carpets of Turkish provenance’, in Kurt Erdmann, Seven Hundred Years of Oriental Carpets, University of California Press, Berkeley & Los Angeles, 1970, pp. 76–80.

16 Rexford Stead, The Ardabil Carpets, Getty Trust Publications, Los Angeles, 1987. Sheila Blair, relying perhaps too much on Stead, discussed the carpets from a technical point of view in ‘The Ardabil carpets in context’, in Andrew J. Newman (ed.), Society and Culture in the Early Modern Middle East: Studies on Iran in the Safavid Period, Brill, Leiden, 2003, pp. 125–43.

17 See Walter B. Denny, ‘Anatolia, Tabriz, and the carpet design revolution’, in Jon Thompson, Daniel Shaffer & Pirjetta Mildh (eds), Carpets and Textiles of the Iranian World 1400–1700, The Ashmolean Museum, Oxford University, in association with the Bruschettini Foundation for Islamic and Asian Art, Genoa, Oxford 2010, especially p. 66.

18 May Hamilton Beattie, ‘The Anhalt carpet’, in Ehsan Yarshater (ed.), Encyclopedia Iranica, Center for Iranian Studies, Columbia University, New York, online at . See also Walter B. Denny, ‘Islamic art: forgeries’, in The Dictionary of Art, Grove, New York, 1996, vol. 16, p. 545; and the same author’s review of Donald King & David Sylvester (eds), ‘The Eastern carpet in the Western world’, in Middle East Studies Association Bulletin, vol. 18, no. 2, 1984, pp. 244–5.

19 In the last several years the carpet underwent intensive examina-tion and, at last, analyses of all of the dyestuffs were undertaken. They are all traditional. Under the modern replacement of all of the original dark brown wool, the original highly corroded pile is easily visible under microscopic examination.

20 Because the Anhalt carpet’s case is such an interesting one when looking at the question of authenticity in carpets, it is the author’s intention to publish a detailed study of the carpet and its criticism within the next several years.

21 The Erdmann attribution of the Salting carpets to nineteenth-century Turkey (see note 12) led to an unfortunate ‘stampede effect’ in the literature. Charles Grant Ellis, Friedrich Spuhler and May Hamilton Beattie concurred, as did this author and many others; all of the Topkapi prayer rugs fell under the same pall. The matter was finally addressed more clearly in a series of articles published in Murray Eiland Jr & Robert Pinner (eds), Oriental Carpet and Textile Studies, vol. v, part 2, ‘The Salting carpets’, ICOC, Danville CA, 1999.

22 See Alberto Boralevi, ‘Three Egyptian carpets in Italy’, in Robert Pinner & Walter B. Denny (eds), Oriental Carpet & Textile Studies II: Carpets of the Mediterranean Countries 1400–1600, ICOC, London, 1986, pp. 205–20.