Over her long career, Barbara Brend has focused on the study of Persian manuscript painting, a tradition that apparently developed in the late thirteenth and fourteenth centuries in circumstances that are still a matter of some scholarly debate. While the later history of painting in Iran is increasingly understood, thanks to the indefatigable work of Barbara Brend and other colleagues, many questions remain about its origins. Among the issues still to be resolved are: when the tradition started; why Iranian artists began inserting images into books; the relationship, if any, between Persian painting and earlier painting traditions in other media, such as wall-painting or ceramics; and the relationship between this tradition and painting traditions elsewhere in the Islamic lands. The evidence for earlier painting traditions elsewhere, whether in Iraq or in Egypt, is particularly fragmentary, yet scholars have often imagined a continuous tradition of book illustration in the Islamic lands, from Abbasid Iraq to Fatimid Egypt to Seljuq and Mongol Iran, despite the paucity of evidence for illustrated books or their existence anywhere in the Islamic lands before the thirteenth century. The following remarks are an attempt to re-examine the evidence for painting in the Fatimid period as presented by two of the most important scholars of the field, Richard Ettinghausen and Ernst Grube.

The origins of Islamic book production go back to the years following the death of the Prophet Muhammad in 632 ce, when the text of the Qur'an was copied on parchment sheets, which were gathered and eventually assembled into codices. But Arabic book production really flourished in the eighth century after Muslims encountered paper, following their conquests in Central Asia, where it had been known for centuries. The availability of paper, which was cheaper than parchment and—unlike papyrus—able to be produced virtually anywhere, encouraged an explosion of books and book learning, particularly in Baghdad, the Abbasid capital.Footnote ¹ Authors produced books not only on the religious sciences, but on every other conceivable subject, from astronomy to zoology, including translations from Persian and Greek, and all levels of literature. Large numbers of books were collected in libraries, of which the most famous was the Bayt al-Hikma, or “House of Wisdom”, the caliphal library in the capital, but there were also great libraries in cities from Marv in Transoxiana to Córdoba in Spain.Footnote ² Of the hundreds of thousands of books known to have been produced in this period, remarkably few survive, although many texts have been preserved through repeated copying and recopying. Although texts occasionally refer to books that had illustrations, no illustrated books have actually survived from the period, apart from an undated fragment of a popular romance which ends with a schematic illustration.Footnote ³

Of these many libraries, those of the Fatimid caliphs in Egypt (969–1171) are known to have been extraordinarily large, housing apparently hundreds of thousands of volumes on all branches of knowledge.Footnote ⁴ We know, for example, that in 993–994 the librarian of the caliph al-ʿAziz was able to produce for his patron 20 copies of al-Ṭabari's History, 30 copies of al-Khalil b. Ahmad's Kitāb al-ʿayn, and a hundred copies of Ibn Durayd's Jamhara. Twenty years later, in 1012–1013, the caliph al-Hakim transferred 1,298 manuscripts of the Qur'an from the palace library to the Mosque of ʿAmr in Fustat. Of these fabled manuscripts, only a handful are known to survive.Footnote ⁵ It has long been assumed that a substantial number of the secular manuscripts would have been illustrated, as some later Islamic manuscripts are, but no illustrated Fatimid manuscripts are known to have survived, although somewhat more than 150 paper fragments bearing images have been attributed to the Fatimid period in Egypt (969–1171). In the absence of surviving manuscripts, scholars have deployed several strategies to attempt to reconstruct the lost art of painting in the Fatimid period, most notably Ettinghausen in his seminal article of 1942, and Ernst Grube in a subsequent series of articles and chapters.

That the art of painting flourished in the Fatimid period in Egypt (969–1171) is known through a famous description preserved in the writings of the Mamluk historian al-Maqrizi (1364–1442). Maqrizi wrote about a contest between painters that had been organised by the Fatimid vizier Yazuri sometime before his execution in 1058. This incident was first noted in the European scholarly literature by H. Lavoix in 1875 and further cited by T. W. Arnold in his influential book, Painting in Islam (1928).Footnote ⁶ The actual text is not by al-Maqrizi, as Arnold and others had believed, but from a lost work by al-Qudaʿi, a Shafiʿi jurist who had served as deputy qadi under the Fatimids before his own death in 1062.Footnote ⁷ As Gaston Wiet noted in 1932, the larger context for this story is a description of the mosque of the Qarafa, which had extraordinary carved and painted decoration on its interior, the work of “painters from Basra and of the Banu'l-Muʿallim”.Footnote ⁸ The existence of this story does not allow us to determine how common figural mural painting was in the Fatimid period: it may have been mentioned only because it was so unusual.

In his 1942 article on Fatimid painting, Ettinghausen (1906–1979) attempted to reconstruct painting in the Fatimid period on the basis of what survived in various media.Footnote ⁹ To round out the textual evidence, he brought together A. A. Pavlovskij's 1890 reproductions of the paintings on the muqarnas ceiling of the Cappella Palatina in Palermo, some fragments of lustre-painted bowls attributed to Fatimid Egypt, a painted Sicilian ivory box, two fragments of painted muqarnas from a ruined bathhouse in Fustat (Old Cairo), a textile fragment, and a drawing on paper in the British Museum. He concluded that Fatimid painting presented several styles, including one that derived from the style associated with the ninth-century Abbasid capital at Samarra and another that seemed closer to Hellenistic models. He suggested that as more evidence was discovered, scholars would be able to delineate more styles. Ettinghausen returned to the subject in a 1956 article on early realism in Islamic art as well as in his 1962 book Arab Painting.Footnote ¹⁰

The subject of Fatimid painting was also enthusiastically pursued by Ernst Grube (1932–2011) as part of his greater interest in the history of Islamic manuscript painting. In the 1970s Grube published groups of what he believed to be Fatimid-period drawings in the Kraus and the Keir collections.Footnote ¹¹ He also published a series of articles on individual drawings on ceramics and paper, ultimately compiling a list of 158 “Fatimid” drawings on paper, even suggesting that some might predate the Fatimid period.Footnote ¹² In order to date these drawings he brought forth a range of comparative materials, adding to Ettinghausen's list the carved ivory plaques that Ernst Kühnel had attributed to the Fatimid period in Egypt,Footnote ¹³ as well as additional Siculo-Arab painted ivory boxes and more Egyptian lustre ceramics. To take one example, he compared a drawing of wrestlers in the Cairo Museum of Islamic Art (inv. 13004) first to a lustre-painted bowl depicting a wrestling match in the same museum and then to the two painted muqarnas fragments that Ettinghausen had previously mentioned as one of the few examples of Fatimid-period painting to actually survive.

Grube stated that these muqarnas, which had been recovered from the ruins of a small bath near the sanctuary of Abu al-Suʿud in Fustat, must date from before the burning of the city in 1163. He wrote that as they were executed in a manner almost identical to that used by the ninth-century painters of Samarra they might even predate what he called the “developed Fatimid ‘Abbasid’ style”. Noting that the bathhouse had stood in the al-ʿAskar quarter of Fustat, which had been constructed by the Abbasids in the eighth century, he rejected a Fatimid attribution for the muqarnas fragments and for the drawing, and suggested that both must date from the ninth century, if not earlier. He concluded that the painting is the “first tangible proof of a continuous tradition of figurative paintings in Egypt from the Abbasid to the Fatimid period”.Footnote ¹⁴

Two important facts, however, entirely escaped Grube's attention. First the muqarnas elements from the Bath of Abu al-Suʿud cannot possibly date to the eighth century or even the ninth, since the earliest surviving muqarnas are found in the Arab Ata mausoleum at Tim, located between Bukhara and Samarqand in Uzbekistan, which is dated 977–978. Scholars generally accept that the muqarnas developed in the eastern and central Islamic lands and appeared in Egypt only towards the end of the eleventh century.Footnote ¹⁵ Even if the bath had been an Abbasid foundation of the ninth century, it is highly unlikely that any interior decoration would have survived from the time of its foundation, as the heat and humidity associated with such a function would have necessitated repeated redecoration. It is therefore simply impossible for the original decoration to have included muqarnas elements, as they had not yet been invented! So, in the absence of any evidence other than the destruction of the city in 1163, one might plausibly date the bath paintings to the first half of the twelfth century, at the earliest.

The second fact that seems to have escaped Grube's attention is that the paper on which the wrestler drawing was done would simply not have been available to Egyptian draftsmen in the eighth or ninth centuries, nor probably even in the tenth century nor through much of the eleventh. Paper, which had been invented in China in the centuries before Christ, was introduced to the Muslim world through Central Asia in the eighth century. As far as we know, Muslims first used paper in Central Asia and Iran, whence it spread to Iraq, Syria, and eventually Egypt by the tenth century. The first papermills in Baghdad, the Abbasid capital, were established in the late eighth century to supply the needs of the burgeoning bureaucracy, and subsequently writers of all types adopted the relatively new medium, spurring a burst of writing and translation that has been characterised as the Abbasid Golden Age. Paper seems to have been available to writers in ninth-century Syria and tenth-century Egypt, where it quickly replaced papyrus, which Egyptians had used as a writing material for the past 4,000 years. But there is no evidence that paper would have been available to artists anywhere at this early date, for it would have remained relatively expensive.Footnote ¹⁶

Furthermore, artists in this period did not make preliminary sketches on a medium such as paper for eventual transfer to another medium, such as pottery or an ivory plaque. Artists would have normally done their preliminary sketches directly in the given medium (for example, on the back of a ceramic tile), and these drawings would disappear as the work was completed. At the same time, no doubt, some books, such as those about geography or astronomy, might have included illustrations as maps or charts, but no illustrated books survive from before the eleventh century, and the early references to illustrated books are so infrequent as to underscore their rarity.Footnote ¹⁷ Even the well-known copy of al-Sufi's Treatise on the Fixed Stars in the Bodleian Library, Oxford, said to date from 1009 and frequently cited as the earliest surviving example of an illustrated Arabic text, turns out to be a twelfth- or thirteenth-century copy, produced at a time when the increased availability of paper encouraged the development and proliferation of the illustrated book in the Islamic lands.Footnote ¹⁸ Before the twelfth century, most artists, as they had in the past, continued to work directly in their chosen media, and it seems quite improbable that drawings on paper would have been used to transfer motifs from one medium, such as ceramics, to another, such as wall-painting.

Thus, the attractive intellectual edifice that Ettinghausen and Grube, among others, constructed to describe painting in the Fatimid period seems open to revision, particularly since recent scholarship has called into question much of the evidence on which it is based. In the following pages I would like to address four media that have been used to delineate Fatimid painting: works on paper, ceramics, ivories, and the paintings of the Cappella Palatina in Palermo.

Comparative evidence

Grube identified nearly 160 “Fatimid” drawings or paintings on paper, but none has a secure provenance. Unlike the hundred thousand paper fragments excavated in the nineteenth century at the sites of Arsinoë, Ashmunayn, and Ikhmim, and now in Vienna, or the 300,000 documents found in the Cairo Geniza, the Fatimid drawings simply appeared on the art market. A leaf with a painting of a human figure now in the Keir Collection (see Figure 1), for example, was identified many years ago by Gaston Wiet as containing a poem of the Umayyad poet Kuthayyir (d. 723) of the tribe of Khuzaʾa, possibly collected by Abu Ayyub Sulayman ibn Muhammad ibn Abu Ayyub al-Harrani.Footnote ¹⁹

Figure 1. Painting of a human figure on paper, 26.35 × 15.88 cm. Published by Gaston Wiet in 1944. Source: The Keir Collection of Islamic Art on loan to the Dallas Museum of Art, K.1.2014.1111.

Wiet dated the piece to the Fatimid period based on details of its costume. Ernst Grube further hypothesised that this leaf was the frontispiece to a manuscript that would have had two such images facing each other.Footnote ²⁰ Nevertheless, the offset image of an upside-down face under the feet of the figure shows that the painting had been folded in half while the paint was still wet, an impossibility had the page been bound in a codex, so the image had to have been painted after the page was removed from a book. In sum, either the text or the painting may be genuine.

Similarly, a well-known paper fragment in the Metropolitan Museum in New York (see Figure 2) bears an almost calligraphic drawing of a lion on one side. The drawing is accompanied by a few lines of text, identified as a discourse on wild animals by the early Jewish convert to Islam, Kaʿb al-Ahbar (d. 652–653). The presence of this text would suggest that the drawing came from an early Islamic book about animals. The drawing's style, however, has suggested a twelfth-century date, which has led some to conclude that this drawing came from an early copy of a text illustrated at a later date. As the reverse of the same sheet, however, bears a drawing of a hare accompanied by a text irrelevant to that on the front of the sheet, it is difficult—indeed impossible—to imagine how this page could once have formed part of a book by Kaʿb al-Ahbar or by any other author.Footnote ²¹

Figure 2. Painting of a lion on paper, 15.72 × 12.06 cm. Bought from Michel Abemayor, 1954. Source: The Metropolitan Museum of Art, Rogers Fund, 54.108.3.

A meticulously executed drawing of two warriors in Cairo, which has been reproduced countless times and cited as evidence for Fatimid costume and weaponry, bears a truncated inscription, ʿizz wa-iqbāl li'l-qā’id Abī Manṣ[ūr] (Power and good fortune to the commander Abu Mansur), presumably referring to a specific individual of the early eleventh century (see Figure 3).Footnote ²² The inscription oddly truncates the most important part of the inscription, namely the dedicatee's name, which contrasts disconcertingly with the meticulous care with which the drawing has been planned and executed. Furthermore, the slight damage affects only unimportant portions of the design. If the drawing is genuine, what exactly would its purpose have been? It is certainly not an illustration for a book, nor was it a preliminary design for something to be executed in another medium, because the level of finish—despite the truncated inscription—is too high.

Figure 3. Drawing of two soldiers, ink on paper, 14 × 14 cm. Bought from Maurice Nahman, 1937. Source: Cairo, Museum of Islamic Art, 13703.

Perhaps the most impressive drawing to survive is a somewhat larger (280 x 180 mm) sheet in the Israel Museum bearing a representation of a nude and tattooed woman carrying a lute (see Figure 4).Footnote ²³ The drawing, on yellowish paper, was first done in red ink and then gone over in black, with touches of white and crimson, a technique that can be traced back to Classical times. Some scholars have suggested that this image of a dancing girl is a specific representation of a famous courtesan in Fatimid Egypt, while others have proposed that it is a generic representation of Venus playing the lute. As there is no text around or on the back of the image, it seems unlikely that the page was taken from a book, nor does it look like a preparatory study for something else, so its purpose remains a matter of speculation.Footnote ²⁴ And the drawing in the British Museum (see Figure 5) that Ettinghausen cited in the 1940s as Fatimid has now been attributed to the Ayyubid period (that is, after 1171) on the basis of the shields depicted.Footnote ²⁵ While there is no possibility of establishing the provenance of any of these drawings, particularly those in Cairo, technical analysis of the paper—especially the drawings in the Keir collection, now on loan to the Dallas Museum of Art in Texas—might be a place to start.

Figure 4. Drawing of a courtesan, ink and colour on paper, 28 × 18 cm. Gift of Mr and Mrs Ralph Harari, London. Source: Jerusalem, Israel Museum, M 165-4-65.

Figure 5. A battle beneath the walls of a town. Painting on paper, 21 × 31.4 cm. Purchased from Major H. Howard, 1938. Source: London, British Museum, 1938, 0312,0.1. © The Trustees of the British Museum.

If there is no evidence that Fatimid potters did preliminary drawings on paper before painting their ceramics, it would seem to be a pointless exercise to reconstruct Fatimid painting by comparing painted ceramics with unprovenanced and undated drawings, especially since the dating of Fatimid ceramics is such a problem in itself. Nearly 50 years ago, the late Oleg Grabar (1929–2011) ingeniously proposed that Fatimid ceramics had initially been non-figural and abstract until the dispersal of the Fatimid treasuries in second half of eleventh century, when representational styles would have percolated from the imperial to the urban realm.Footnote ²⁶ Unfortunately, the few signed ceramics show that figural representation was an integral feature of Fatimid lustre pottery from the beginning, and the two datable pieces—one abstract, one representational—are associated with courtly patrons.Footnote ²⁷ Current research suggests that the representational pieces are actually earlier than the abstract ones! As Ettinghausen and Grube already noted, some representations seem to develop from those on earlier Abbasid lustrewares, while others seem to derive from Hellenistic and eastern Christian traditions of representation and draftsmanship which apparently continued to be practised in Egypt in early Islamic times.Footnote ²⁸ The variety of representation is extraordinary—traditional subjects like animals, hunters, musicians, dancers, entertainers, and rulers, as well as “new” subjects like wrestling, cock fights, rope dancing, or duels—although the question of whether these representations are “realism” or “formalism” remains open.Footnote ²⁹ In any event, the sensitivity and elegance of the drawing on some of the ceramic fragments is sometimes quite amazing and often surpasses the quality of drawing seen on the paper fragments, suggesting that, as would be the case later in Seljuq Iran, the most talented draftsmen in Fatimid society worked not on paper but on pottery. Of course, in Iran all that would change in the Ilkhanid period, and in Egypt ceramics with representations went out of fashion.Footnote ³⁰

Just as it is unlikely that potters practised their drawing skills on paper before decorating the ceramic surface, it is equally unlikely that ivory carvers—if they existed at all in Egypt—drew their designs on paper before carving. Although scholarly consensus has attributed many exquisite carved ivory plaques and oliphants in various European and American museums to Fatimid Egypt on the basis of purported similarities between the ivories and the carved wooden beams thought to come from the Fatimid palace in Cairo, there is, as Silvia Armando has recently reminded us, absolutely no evidence that any of the ivory objects were produced there, and many, if not all, of them may actually have been produced in southern Italy.Footnote ³¹ Nevertheless, Eva Hoffman has proposed that four carved ivory plaques in Berlin, which have been assembled as a frame, once decorated the cover of a Fatimid book, although no such book cover is known from anywhere else in the Islamic world and the contents of such a book remain unstated.Footnote ³²

The most frequently cited evidence for reconstructing the lost art of Fatimid painting is the painted wooden muqarnas ceiling of the Cappella Palatina (1131–1140) in Palermo, erected under the patronage of Roger II.Footnote ³³ The painted figures, which from the ground are nearly invisible to the naked eye, have always been thought to represent the metropolitan style of Fatimid painting, although it must be said there is little evidence to make this assertion. Despite the presence of the two painted plaster muqarnas fragments found in Fustat, no complete muqarnas ceilings or vaults comparable to the Palermo ceiling, let alone painted decorations, have been found in or are known from Fatimid Egypt.Footnote ³⁴ David Knipp has suggested that the style of representation on the Palermo ceiling was less like Fatimid painting—whatever that might have been—than that of northern Mesopotamia, which he was able to illustrate with several examples.Footnote ³⁵

The closest parallel to the Palermo muqarnas ceiling in form and date, if not in technique or decoration, are the exactly contemporary muqarnas vaults of moulded plaster erected over the nave of the Qarawiyyin mosque in Fez, which were added to the mosque in 1136–1137 on the orders of the Almoravid sultan, ʿAli b. Yusuf (r. 1106–1143).Footnote ³⁶ The five muqarnas vaults—four resting on square bases and one on a double square (rectangle)—are far more elaborate than the two other Almoravid examples known: those in the corners of the Almoravid qubba in Marrakesh (1117) and those in the corners and centre of the dome in front of the mihrab at the Great Mosque of Tlemcen (1136) in Algeria. All of these high-prestige Almoravid vaults can be understood to represent not the indigenous traditions of northwest Africa but the architecture of metropolitan Córdoba, although few, if any, examples of Almoravid buildings survive in Spain.Footnote ³⁷ The recent discovery in Murcia of some muqarnas fragments with painted figural representations raises the possibility, largely ignored until now, that the Palermo ceiling was more a product of a western Islamic tradition of painted muqarnas ceilings that is now completely lost. In short, it is simply circular reasoning to argue that the Palermo ceiling represents Fatimid painting because Fatimid painting resembles the Palermo ceiling.

Nearly 75 years after Ettinghausen published his attempt to reconstruct Fatimid painting, it seems reasonable to ask whether the topic of Fatimid painting is still worth pursuing? Is painting a valid category? Is there anything that connects the extant depictions on ceramics with the lost depictions on walls and the imagined depictions in manuscripts? The art of the Fatimids has exercised a particular fascination because of the frequency of representations of humans and animals in its art, otherwise thought to be unusual in Islamic, particularly Arab, art, although it would become, as Barbara Brend's work has amply shown, a major feature of later Persian art. If not painting as a category, should we be thinking instead about representation?

An investigation into the nature of representation in the Fatimid period will not answer the question of whether there were illustrated manuscripts in the Fatimid period or whether this was a later phenomenon due to the increased availability and affordability of paper. While I have no doubt that some manuscripts produced in the Fatimid period may have been illustrated, particularly works of geography or astronomy, it seems to me that despite the existence of the so-called “Fatimid drawings”, copiously illustrated books of the type that became common in later centuries were not produced in the Fatimid period, principally because the medium on which they would have been made was not yet so cheap as to encourage people with drawing skills to use it.