Five Essays on Islamic Art

Aniconism and Figural Representation in Islamic Art

by Terry Allen

excerpt, without illustrations. Copyright 1988 by Solipsist Press.

(back to: Palm Tree Books Home Page)

(back to: description of Five Essays on Islamic Art)

If Islamic art is an outgrowth of Late Antique art that does not mean that Islamic art is only late Late Antique art, or even that most of Late Antique art survives in Islamic art. The most obvious difference between the two is the lack of figural representation in Islamic religious art and a shift in the use of figural representation in secular art in the Islamic world. The difference between religious and secular usage is important to observe.

Religious and Secular

The Umayyad palaces of the inland Levant, such as Mushattâ and Khirbat al-Mafjar, show that the art of the early Muslims was not entirely devoid of figural imagery (Mushattâ, triangle A; Small apsed room, Bath Hall, Khirbat al-Mafjar; Caryatid in Bath Hall entry, Khirbat al-Mafjar).

It was early religious art and architecture, monuments such as the Dome of the Rock or the Great Mosque of Damascus (Dome of the Rock, interior of the octagonal arcade, pier 5; mosaic in W arcade, Great Mosque of Damascus), that were without figural representation.

The sharp difference between these two sets of monuments is not some peculiarity of Umayyad culture, peculiar though Umayyad taste may at times seem to us. Except for particularly repressive periods and the damage caused by the actions of zealots, figural representation has always been a part of secular art in the Islamic world.

The twentieth century is no exception. Even the most reactionary religious regimes display in public photographs of people, particularly religio-political leaders; and when a Persian mob sacked the Moroccan embassy in Tehran in 1986 in the name of religion (the Israeli Prime Minister having just visited Rabat) photographs of King Hasan were burned along with Moroccan flags while photographs of Persian leaders were brandished at the television news cameras. I found a somewhat less charged though not quite so apt example in Algiers, which experienced a heavy dose of Westernization during its time under French rule between 1830 and 1962. Since independence many colonial buildings, particularly churches, have been put to new uses. The former Church of St. Charles in Algiers has been so converted (to a mosque, I believe), and Figure 36 (Algiers, mutilated capital of porch of former Church of St. Charles) shows how one of the Roman-revival capitals of its porch, originally adorned with a bust, has been mutilated so as to eliminate the human likeness. But only a few minutes's walk away, on one of the main thoroughfares of the city, the second-floor balconies of a late nineteenth-century building sport Gibson Girl caryatids, who have been exposing themselves to view unmolested but no doubt admired for the better part of a century (two illustrations; compare Caryatid, Bath Hall entry, Khirbat al-Mafjar).

Neither representation was Algerian in origin, and probably neither was carved by an Algerian, but their present states of conservation and destruction still represent local attitudes. There is no contradiction in the coöexistence of the caryatids and the mutilated capitals, and never has been, as one can see in the facade of Mushattâ. Mushattâ's carved decoration includes animal and even human representations (Mushattâ facade, triangle A) except for the portion of the facade corresponding on the exterior of the building to the qiblah wall of the mosque inside, the wall on the side of Mecca, toward which prayer is directed (Mushattâ, triangle M).


The decoration of that portion, and only that portion, is entirely vegetal. At Mushattâ and in Algiers the division in decoration corresponds to the division between that which pertains directly to the practice of religion and everything else.

Still, these examples are sculpture, and sculpture was not a widespread form of art in the Islamic world after the Umayyads, who represent in many ways a last gasp of Late Antique art. Historical accounts indicate that the palaces of rulers usually had public figural representations, sometimes including sculpture of some form, but for most of the population figural representation was two-dimensional and usually applied to objects of some utility.

Religious Aniconism

The early history of Islam is recorded in oral tradition that was written down only later, when the Islamic historical tradition began in the mid-eighth century. This oral tradition and the written histories that were constructed from it are difficult source material, concealing many problems in interpretation. [Note: For a sharp statement of the problems see Patricia Crone, Slaves on Horses, pt. 1. ] To judge from these sources, in the early period of the development of the religion it was thought to be entirely appropriate for figural representations to occur in the context of ordinary life and even in religious settings, if those representations were not meant for religious purposes. Where there was no intent to create an icon (for which there was supposed to be no purpose in Islam), figural imagery seems to have been unexceptionable. Thomas Arnold, writing in 1928, gathered material on the early Islamic use of images from historical and literary sources which shows that a creeping iconoclasm replaced in pious doctrine this sensible if casual practice. For example, Arnold found hadîths, or purported oral traditions about the acts and words of Muhammad, clearly supporting the secular use of images, for example the report that Muhammad objected to figural curtains in his house-cum-mosque in Madînah but was satisfied when the curtains were cut up for cushion covers: their different orientation as cushions made them unlikely objects of prayer, and hence apparently acceptable. Whether these stories are accurate or not, they represent a straightforward and uncomplicated response to the issue. The traditional Muslim theological objection to images, which may have been observed more in the breach than in ordinary life, was eventually codified in a quite rigid form and extended to the depiction of all animate beings. It is captured in the prediction that "on the Day of Judgement the punishment of hell will be meted out to the painter, and he will be called upon to breathe life into the forms that he has fashioned; but he cannot breathe life into anything.... In fashioning the form of a being that has life, the painter is usurping the creative function" of God. [Note: Arnold, Painting in Islam, pp. 5--7. For secular examples of the use of images on objects and in architecture see Oleg Grabar, "Ceremonial and Art at the Umayyad Court," ch. 4, pp. 174ff. ] There is a clear contrast with official Byzantine Iconoclasm on this point: in Byzantium it was pictures of the saints, of Christ, and of God that were forbidden because they might be worshipped; these images could be replaced with scenes of the circus or hippodrome, which certainly do not lack animate subjects.

Nor was Western European iconoclasm opposed to all figural representation: to the anonymous author of the Libri Carolini, a collection of Latin iconoclastic texts compiled at the Carolingian court, "the artist was free to do what he liked... provided that he remained irremediably profane." [Note: Peter Brown, "Dark Age Crisis." ] Despite later theological doctrine, this rational division between religious and nonreligious subject matter was probably the original basis of Islamic iconoclasm too, and the reasons given by the theologians were probably secondary to the original motives for the nonuse of figural representations in religious contexts. That is to say, there can be an "Islamic" point of view on figural representation, articulated by theologians and considered as foremost by the pious, that differs from "Islamic" points of view held by most of the Muslim population. In view of this likely division of sentiment, which contrasts with Byzantine iconoclasm, I prefer to term the Islamic phenomenon not iconoclasm, the rejection of images, but aniconism, the nonuse of images. The important issue for the history of art is not the eventual theological justification for Islamic aniconism but the degree of continuity of practice between the pre-Islamic past and the early Islamic period---how people actually lived their lives in the seventh and eighth centuries---which is a matter not well handled by our written sources.

To explain the aniconism of Islamic religious art it is necessary to set early Islamic religious art in its historical context, to understand at least partly its manifestation in the mosque, and also to indicate why later Islamic religious art did not cast off aniconism and adopt figural representation. (Some religious representations were certainly produced from at least the fifteenth century onward, particularly by the Shî`ah, but they do not constitute an established and formal practice when seen against the background of continued institutional religious aniconism.)

Statues were destroyed as pagan idols in the fifth and sixth centuries in Byzantium and the Near East, just a few centuries after Constantine ransacked the Mediterranean for antique statues to decorate his newly refounded capital. As an example of this prominence of statuary in the Late Antique period, the Baths of Zeuxippus in Constantinople, a large structure located on the Hippodrome, had close to a hundred statues within it before its destruction in the Nike riot of 532. [Note: Cyril Mango, The Brazen House, p. 39. ] Nearby was the Basileios Stoa, which was evidently a public square surrounded by university buildings, bookstores, and the like, built ca. 410, burnt and rebuilt in the fifth century, and partially burnt again in 532. In the sixth century it contained "a gilded statue of Justinian II in a kneeling posture, a statue of his Khazar wife, a huge elephant together with his attendant, and a seated bronze figure of Theodosius I." [Note: Ibid. p. 50. ] The Byzantine emperors, and others, continued the antique practice of collecting statues as art at least as late as the reign of Theodosius II (406--50), though they did not generally display them publicly. [Note: Dericksen M. Brinkerhoff, A Collection of Sculpture in Classical and Early Christian Antioch. ] While few statues were still produced, there was no lack of relief sculpture, which must have been viewed as less real and therefore less threatening. It was the perceived reality of sculpture that was the problem, since at this date an image was identified with what it represented, in the street as well as in the church. For at the same time the population in general was coming to regard statues as imbued with immanent magical powers. The Emperor Maurice (assassinated 602) destroyed statues not for religious reasons but to counter their magic. This was thought to be a dangerous undertaking requiring considerable courage. Still, imperial statues were produced as late as the eighth century; sixth-century emperors were represented on the facade of the Chalke Gate, [Note: Cyril Mango, "Antique Statuary and the Byzantine Beholder," and The Brazen House, pp. 102--07. ] the main entrance to the imperial palace, where an image of Christ mounted over the gate was attacked in 726 and restored only in 843 after the defeat of the Iconoclastic party. [Note: John Beckwith, The Art of Constantinople, p. 61. ] It seems that until the Iconoclastic controvery the antagonism to public images applied only to figures that were strange or had no meaning anymore for the man in the street with a stone in his hand.

On the religious side there had been ferment over the use of images as objects of devotion for several centuries before the proclamation of Iconoclasm in 726. [Note: On this identification in prayer see Ernst Kitzinger, "The Cult of Images before Iconoclasm." For the possibility that it was Muslim iconoclasm that touched off the outbreak of Iconoclasm as official Byzantine doctrine, see Patricia Crone, "Islam, Judeo-Christianity and Byzantine Iconoclasm."] Clive Foss relates an anecdote about a late seventh-century depiction of an angel that had a label which apologized to the angel for representing it. [Note: Clive Foss, Ephesus After Antiquity, p. 91.] The Monophysites in particular (who in opposition to the church at Constantinople believed that Christ had but one nature, and who constituted a large part of the Syrian population) seem to have found representations of religious entities distasteful. When Iconoclasm was declared as a policy in the Orthodox church it set off decades of serious political clashes. The way the issue was used for political purposes indicates a similarity between the social context of Iconoclasm---that is, Byzantine society---and Islamic society, in which politics and religion are as intertwined as they often have been in countries that are formally Catholic. In all these cases government is identified with the enforcement of official religious morality.

But regardless of its political overtones Byzantine Iconoclasm cannot account for Islamic aniconism. Marlia Mundell Mango has shown that even in the sixth century the decoration of Monophysite churches in inland Syria and upper Mesopotamia---the area and sect of many Arab Christians---was entirely nonfigural. [Note: "Monophysite Church Decoration."] Thus a nonfigural Islamic religious art can be seen as continuing nonfigural Arab religious art before Islam. This is not iconoclasm but aniconism.

Early Islamic aniconism is specific to the mosque, a new architectural form, or group of forms, that resembles Christian and Jewish houses of worship in its general appearance but differs from them---certainly deliberately---in many important specifics. Whether one prefers Christian or Jewish prototypes for the mosque, the implications of the connection depend on one's answer to an almost unanswerable question: Was the mosque's likeness to the houses of worship of other monotheistic religions meant to indicate that the mosque was also a monotheistic house of prayer, in which case its similarity to synagogues and churches is primary, or was the user meant to notice the difference between the mosque and other places of worship? In any event, the rectangular plan of the mosque is suited to the arrangement of worshippers in rows facing the qiblah, but its liturgical furniture, the mihrâb or prayer niche, and minbar or pulpit, is not. And mosques, at least early ones, do not have programmatic decorative schemes, with the partial exception of inscriptions taken from the Qur'ân. If Arab Christians before Islam had no figural images in their churches, we need not see the mosque as a near-church stripped of the images that formerly accompanied prayer, but rather as a setting designed without images in mind. On this view the Muslims would have been acting out of conviction in avoiding the plethora of figures that characterized Byzantine art, not, as Oleg Grabar may seem to suggest, out of a lack of choices. [Note: In Formation; for a cultural analysis somewhat in contrast to mine see Grabar's "Islam and Iconoclasm."] After all, there is no proper Islamic equivalent to the Christian practice of prayer to a holy intercessor, which is the point of a cult image in Christianity. [Note: Brown, "Dark Age Crisis," p. 270. ]

So the mihrab, an empty niche with no explicit primary iconography, became the apparent focus of mosque decoration. In the more developed form of the mosque the antemihrab bay that includes the mihrab is treated as a theatrical version of the cosmos, with the structure of the heavens suggested by an intricate design. The minbar, the pulpit of Muhammad, is reproduced in the same bay for the prayer leader to speak from. The address given by the imam from the minbar (the khutbah) by convention invoked the name of the ruler, and the sovereign prayed nearby, perhaps in order to indicate his place in the divine scheme. The antemihrab bay is the most decorated part of the interior of the mosque by far, and the hypostyle structure of many early mosques ensures that it can be appreciated only when one is within it.

Thus Islam adopted a monumental setting for prayer composed of a mihrab, minbar, dome over the antemihrab bay, and, beyond, a relatively unhierarchical prayer hall. As symbols, these elements are extremely understated, particularly when compared with Gothic cathedrals or Hindu temples, and not only in their lack of figural imagery. But what was new in the mosque as a form of religious architecture was that the formal focus of the decoration, the mihrab, was not the visual focus of the prayer hall, not that the mosque had no figural decoration. Most importantly there is no reason to think that the motifs, techniques, and designs that later came to be used in decorating the mosque originated in a religious context. They were specific applications of arts that existed in secular architecture as well---and that could include figural components. Islamic religious aniconism applies to the mosque first and foremost.

But why did the mosque remain devoid of figural representations? This question is more easily answered by looking at secular art, whose figural content also changed significantly in the transition from Late Antiquity to the Islamic era.

Figural Representation in Secular Art

In secular Islamic art there is a shift in use of figural representation, not in the seventh century, but surely by the eleventh. The Umayyad princes certainly showed no distaste for figural art, but it seems clear that by the `Abbâsid period there had been a reduction, in both scale and physicality, in the use of figural representation in even the most lavish constructions: while the palaces of the caliphs at Samarra had frescos of dancing girls, there is nothing at Samarra like the carved facade of Mushattâ or the three-dimensional erotic sculpture of Khirbat al-Mafjar. This shift affects all of Islamic art, but can be seen most strongly in the Arabic-speaking parts of the Islamic world. In the Islamic art of the Persian-speaking world and, later, Islamic India, the most recognizable group of images contains representations of legendary heros, usually those of the Shâh nâmah; [Note: On narrative images of the Shâh nâmah see Marianna Shreve Simpson, "Narrative Allusion and Metaphor in the Decoration of Medieval Islamic Objects."] and in this region a tradition of fine manuscript illustration unequalled elsewhere in Islamic art was created following the Mongol invasion of the early thirteenth century. The Arabic-speaking world produced illustrated manuscripts too, but with only a few exceptions Arabic manuscripts are continuations of the Late Antique genre of scientific or luxury books and are not artistic products directly connected with literature or contemporary life. By comparison with Persian manuscript illustrations the pictures in Arabic books are not artistic successes, and they do not stand comparison as art with other forms of art in the Arabic-speaking part of the Islamic world. The internal cleavage that this comparison reveals points once again to the importance of language sharing and language barriers for the shape of artistic developments. There are other cleavages in figural art, both in time and between genres of representation.

In this discussion I shall discount early manifestations of figural representation such as Umayyad architectural decoration, along with the almost lost art of later wall painting. I shall also set aside manuscript illustration. All these things strongly reflect the high status of their pre-Islamic prototypes: the illustrated codex, the ruler's palace, the bath. Persian manuscript illustration after the Mongol invasion reflects, I think, the artistic quality of Chinese painting. In these things the prototype is strongest, and it is an essential part of the prototype that it includes figural representation.

I shall concentrate instead on more common media: pottery, metalwork, and wood. These portable arts were highly developed in Islamic material culture, and display a considerable degree of similarity of decoration from one medium to another. Likewise, they follow much the same pattern in both Arabic- and Persian-speaking spheres. Here too a differentiation of genres, or at least of types, of figural representation is neeeded.

I leave aside such quasisculptural survivals of pre-Islamic prototypes as figural coins and medals; animal protomes, finials, handles, thumbstops, and legs; zoomorphic vessels and aquamaniles; and dolls. None of these is an invention of the Islamic era, and all are quite restricted in their figural range, both in pre-Islamic times and later. Much more interesting than objects shaped as figures is the larger and more variable class of representations applied to objects.

The repertoire of representations on these portable objects features single figures, figures paired, and figures displayed in emblematic fashion rather than depicted in specific moments of a known story. These emblems are often grouped in cycles and contained within nonfigural decoration, which may even dominate the design. Whatever meaning they have, their context seldom suggests that they were the primary interest of the piece on which they appear. A few categories account for most examples: individual animals or figures, particularly birds or horsemen (2 Persian pots); friezes or fields of animals such as dogs, fish, or ducks(another); couples seated side by side (a fourth); heros, rather than stories, from the Shâh nâmah, although stories do on occasion appear; cycles of zodiacal symbols, of planetary symbols, or of vignettes standing for the labors of the months or seasons (Persian brass penbox, 1281, BM 42624; ivory plaque from a book cover, Cairo, 12th C, Florence, Museo Nazionale); and finally the so-called princely cycle. [Note: A term introduced by Grabar; see Formation, pp. 163ff. ] Fantastic creatures such as sphinxes, harpies, and unicorns occur much more rarely, though just for that reason they have been studied more intensively. [Note: Eva Baer, Sphinxes and Harpies in Medieval Islamic Art, and Richard Ettinghausen, The Unicorn. For animals of all kinds on metalwork see Baer, Metalwork, pp. 154--87. ]

These representations seldom occur in connection with explanatory inscriptions. Most, it seems, appear on objects that carry no inscriptions at all; other figured objects have standardized religious invocations and inscriptions expressing generalized and conventional good wishes; still others have personalized inscriptions, with the name and titles of the customer who ordered the piece. [Note: For invocations (du`âs) see Assadullah Souren Melikian-Chirvani, Islamic Metalwork from the Iranian World 8th--18th Centuries, pp. 29--35, and cf. my review in Ars Orientalis, v. 15, 1985 [1986]. ] Persian inscriptions have been studied more thoroughly than Arabic ones. Inscriptions on Persian luster pottery have been surveyed by Oliver Watson, who remarks a

lack of relationship in general between image and inscription on both vessels and tiles. While the common image of two figures seated together may possibly be thought suitable as an illustration to some of the "love" verses (though these usually deal with the tortures of separation rather than the pleasures of the loved one's company), we also find such verses together with images of enthroned figures, hunters, polo players, and animals. These images are often explained as part of a "princely cycle" and sit ill with the tenor of the verses.

Even Shâh nâmah images are coupled with what "appear to be inappropriate `love' quatrains." [Note: Watson, Persian Lustre Ware, pp. 152--53. ] Watson refers to Grabar's observation that they correspond not on "a narrative and illustrative level but on some other level,... both reflect[ing] sentiments accepted as being appropriate to the occasion." [Note: "The Visual Arts," p. 647. ]

"Correspond" may be too strong---perhaps the image and the inscriptions merely parallel each other. But otherwise this insight is surely valid, although it does not help immediately with understanding what the iconography of these pieces is. Watson's disappointment in finding no greater coherence unfortunately leads him to suggest that both images and quatrains were "imbued with religious significance" to the Shî`ah, but there is no justification for believing this and good reason to believe the contrary, that the images were not so important that they had to bear the kind of meaning Western art is accustomed to assigning them. The same may apply to the texts: it may be that sometimes a text was to be seen and not read; that would explain why the same lines occur repeatedly. Whatever the solution to this puzzle, it is fair to say that inscriptions do not generally explain the figural representations they accompany.

There are two separate issues to be resolved: the significance of the figural representations and their distribution, and the artistic practice of emblemization.


Much has been written on the meaning of figural representations in the portable arts of the Islamic world, though no general interpretation has gained favor. I do not intend to recapitulate this literature, but rather to point to what seem to me the most sensible interpretative categories and how they fit together, so as to indicate the general character of these images.

Some of the isolated images seem to be apotropaic, or what one might call loosely auspicious animals. The many birds, especially on ceramics, seem to fall into this group. Probably no provable meaning will ever be found for them, since the contemporary folklore that would explain them has been lost.

Certain objects decorated with images of Shâh nâmah heros might have been seen as continuing an ancient tradition, and the post-Sâsânian "Sâsânian silver" must belong to this group. But Shâh nâmah scenes, or at least figures, appeared on other objects as well, such as Persian luster ceramics. The meaning of these figures is clear, even if the specific reasons they were represented on these vessels and tiles have not been explained. An entirely different, and fairly large group of Fâtimid Egyptian luster bowls (some of them fakes?) shows diverse human figures, evidently referring to real life (Cairene luster bowl, Freer, 46.30). [Note: Ettinghausen, "Early Realism in Islamic Art." ] These, too, invite further explanation, as do a few Persian ceramics such as a plate in the Freer Gallery published by Ettinghausen as a Sûfî work. [Note: Grace D. Guest and Richard Ettinghausen, "The Iconography of a Kâshân Luster Plate." ] In this latter case the inscriptions still need to be read; unfortunately the Cairene luster bowls are seldom inscribed with anything but potters's names. Neither group seems to match up with works in other media.

Many of the cycles refer comprehensively to the cosmos: the planets are distributed throughout the heavens, the labors of the months and seasons cover the whole year, and the zodiac encompasses both the heavens and the year. The elements of these cycles seem to be fairly simple and routine, though there is considerable variation among sets of cycles, particularly in the seasonal images. [Note: D. S. Rice, "The Seasons and the Labors of the Months in Islamic Art." ] A notable exception deserving of further analysis is the "Blacas" ewer in the British Museum, whose emblems are quite specific in their imagery, approaching the specificity of narrative scenes ("Blacas" ewer, brass inlaid with silver, Mausil, 1232, BM 6612-2961; and detail). If there is an internal logic to the specific application of one or another of these cosmic cycles to an individual object, such as a zodiac indicating the time of year the piece was made, scholars have yet to find it. Less specifically, cosmic cycles may be centered on a seated figure usually interpreted as an enthroned mortal or (or, also as) the sun, possibly in reference to the owner or user. [Note: See Baer, Metalwork, pp. 258--74. ] So it may be appropriate to invoke the evidence of the inscriptions usually found on objects. These very often are composed of good wishes, invoking what may appropriately be called a cycle of kinds of well-being. Similarly, then, the cosmic cycles, or indeed, any category of figural representation, may have been intended to suggest, even to promote, cosmic well-being. Other associations, derived from themes of the passing of time, the seasons of the year, and astrology, would have been appropriate and may even have motivated the use of a cosmic cycle. James Allan has argued for a pervasive solar symbolism in inlaid metalwork, a symbolism more believable for candlesticks than for ewers. [Note: Allan, Islamic metalwork: the Nuhad es-Said Collection, pp. 46--53ff. Melikian-Chirvani sees esoteric imagery in many other places as well. ] The cycles could have been intended to refer to the contents of the vessels on which they occur, or they could have been meant to glorify the owner or user of the vessel by glorifying the vessel itself. In either case the owner or user obtained intellectual as well as aesthetic satisfaction by putting the object to its intended use.

The inscriptions on these objects are not often indicative of their use or, aside from the good wishes, of their iconography. I can see hardly any difference in the inscriptions on figural as opposed to nonfigural vessels, nor any difference between inscriptions on objects of one function and those of another. Likewise, the figural decoration seems not to be adapted to the type of vessel it decorates. In the thirteenth century inscriptions on vessels, particularly Persian ones, begin to speak in the voice of the object, but this direct relationship between the inscription and the use of the objects is rare earlier and much more common in the fourteenth and fifteenth centuries. This lack of specific connection between inscription and function or iconography leads me to conclude that good wishes really are the basic iconography and that figural representations were simply intended to further that theme visually. [Note: For early inscriptions in the voice of the object, see Baer, Metalwork, pp. 216--17. Allan, loc. cit., cites an inscription that clearly asks for cosmic protection for the artisan. ] Here the inscriptions do parallel the imagery. Otherwise the emphasis of the inscriptions is on the patrons who ordered the objects made and on their titulature. Are these personalized objects simply customized versions of ordinary luxury goods, or were they made for great occasions, to commemorate the assumption of an office or a title? Such a practice might explain the large number of extant inlaid metal penboxes, for example, but the distributional analysis needed to test this idea is yet to be done.

The "princely cycle" is clearly a separate case. In objects decorated with this cycle musicians, hunters, dancers, and cupbearers are juxtaposed with a seated figure who usually holds a cup filled, one presumes, with wine. The earliest of these cycles on luxury objects may be the well known series of ivory boxes from Umayyad Spain. [Note: John Beckwith, Caskets from Cordoba. ] There are many examples of such figures in Persian mînâ'î (enamelled) ceramics of the twelfth and thirteenth centuries, arranged in multifigure compositions that can be read as scenes but are actually composed of individual figures simply juxtaposed with each other as in any other cycle (Persian pot).

The princely cycle as it appears in these works must be distinguished from the relatively infrequent objects, such as the inlaid brass basin in the Freer Gallery made in Cairo for the Ayyûbid sultan al-Malik al-Sâlih Najm al-Dîn Ayyûb about 1240 (Freer 55.10), that show friezes of "princely" figures in loosely arranged groups, presented in a vaguely defined but apparently unitary space and not in a cyclical fashion. [Note: Most recently published by Esin Atil, Islamic Metalwork in the Freer Gallery of Art, pp. 137--46. They are not the source of the princely cycle: D. S. Rice thought the earliest such frieze showing a procession to be on the shoulder of a ewer made ca. 1200 in Mosul, "Studies in Islamic Metalwork---II," pp. 73--79; Eva Baer publishes a mid-fourteenth century bowl with a frieze illustrating a Shâh nâmah scene in Metalwork, pp. 277--79. ] An entire group of fourteenth-century south Persian inlaid metalwork has similar unitary scenes. [Note: Discussed in detail by Melikian-Chirvani, op. cit., ch. 3. Several such pieces are illustrated in color by Atil, Renaissance of Islam. ] Once again these groups find no parallel in other media. The iconography of these friezes may be the same as, or at least related to, that of the princely cycle, but their presentation is quite different. I shall set these friezes aside, like the Shâh nâmah representations, on the assumption that they are not relevant to the main artistic development of emblemization and not characteristic of the Islamic portable arts more generally.

Since the identifiable cosmic cycles are tightly organized with clearly coöordinate emblems, it is tempting to see the princely cycle as unanalyzed residue, which need not be understood as an entity. But elements of it occur often enough in works of high quality, and therefore high self-awareness, that some meaning probably existed for it. [Note: Melikian-Chirvani, op. cit., p. 141, suggests that in the princely cycle "the literary themes of bazm-o-razm, `feasting and fighting,' are illustrated systematically," but he provides no literary citations. ] Since the princely cycle shows a variety of human recreations, apparently fairly standard at any one time and place, as the plantary symbols were, it would be sensible to see it as somehow a humanized version of the cosmic cycles. Indeed, to judge from its use, the princely cycle could be another cosmic cycle whose symbolism has not been guessed by modern scholars. Since most of the princely cycle images appear on objects that would have been used in social situations, it may be that the intention was to lend the social occasions the quality of the occupations depicted in the cycle. The psychology of the princely cycle seems to go beyond (or differ from) that of the cosmic cycles, which suggest harmony with or invoke the protection of the world indicated by the emblematic figures; the princely cycle identifies the owner or user with whatever it is the figures themselves are doing.

Grabar, following Herzfeld, has argued that the princely cycle depicts the pastimes of monarchs, especially the royal feast, as a way of indicating their imperial status. Drinking, listening to singers and poets, watching female dancers, and hunting are all recreations belonging to this imperial cycle, which existed already in Sâsânian art, as seen in the cave reliefs at Tâq-i Bustân; it was developed further in the Umayyad period. [Note: Grabar, "Ceremonial and Art at the Umayyad Court," pp. 81--90 and ch. 4, esp. pp. 223--24, focussing on hunting; p. 94 for the royal feast. These views are adumbrated in Formation, pp. 163--64. ] But pastimes such as music and the hunt did not belong solely to monarchs: while good historical sources say that the Umayyad caliphs drank alone when in assembly, these texts may well reflect the etiquette of formal occasions, rather than everyday custom, and drinking may have been commonplace under the Umayyads, which is evidently when the princely cycle was formed (after all, the norm against which new Islamic values were promulgated was the free consumption of wine, not abstention). It was not the mere consumption of wine that made drinking a royal action for the Sâsânians and the Umayyads but rather its consumption in the context of the royal feast. More generally, the emblems of the princely cycle are the indications of success, wealth, and cultural or political security. The world referred to here is the world of secular self-celebration, and it must have had wider appeal as imagery than a strictly imperial explanation can give it. Grabar's explanation of the princely cycle as related to the Umayyad court is satisfying because it helps to account for the extensive iconographic assemblages of the Umayyad palaces, which range far beyond mere cycles, though they seldom seem to be narrative. But other wealthy individuals pursued the same recreations, and may have adopted the princely cycle as well.

In view of the consistency with which the Umayyads equipped their rural palaces with baths on the Roman model, it is interesting that no representation of bathing appears to have been made part of the princely cycle; this is perhaps proof that the cycle is inherited from the Sâsânians and not invented in the Islamic period, but may also reflect a lack of Late Antique bath imagery, or perhaps indicate that mere bathing was no status symbol. Bathers or no, the princely cycle never lost its popularity during the period considered here.

The emblems of the princely cycle, then, seem not to refer to abstract entities such as signs of the zodiac, but to a set of activities that are somehow appropriate to either the use of the object on which they appear or to its user. This iconography establishes harmony rather than invoking protection; perhaps one should conclude that the cosmic cycles should be read the same way, as indicating the harmony between the decorated object, its contents, or its user, and the cosmos as represented emblematically.

Whatever their true nature, it is clear that in most cases the images of the princely cycle are as emblemized as those of any other cycle: the hunt does not take place in the same time or space as the drinking party with singers and dancers. And the princely cycle need not be presented as a cycle: individual figures apparently drawn from the princely cycle (such as women with some object in hand) occur on Persian luster bowls, and may have been intended to refer to the iconography of the entire cycle. The same point recurs: on these Islamic objects representation is broken into emblems.

I have narrowed the range of my discussion of emblems by discarding single figures and representations of literary characters, and by ignoring unusual inscriptions and certain classes of objects. Thus the evidentiary base for my conclusions is narrow, and is probably skewed by the patterns of traditional scholarship on Islamic art---that is, by what is known already. Further investigation depends on finding groups of objects whose imagery expands or crosscuts my categories of understanding, or by showing that the categories I have set aside are connected in their meaning with categories I have discussed. Another caveat arises from the fact that the best works artistically are often more sophisticated iconographically: the "Blacas" ewer is a good example. Can one claim that the atypical but unusually good object is evidence for the hidden meaning of objects of ordinary quality? Where the objects concerned were created primarily as art (Western oil paintings, for example), indeed one can. But when the objects are pottery bowls, metal ewers, artfully designed objects of common use and standard shape, the argument from quality may be turned around, and it is just as reasonable to conclude that the uniqueness of high-quality pieces is the result of extra attention to the circumstances of their patrons, a view that is confirmed by the concentration of personalized inscriptions on high-quality work.


Whatever one is to make of my conclusions regarding the iconography of emblemized figures, it is plain to see that the representations considered above occur on objects of secular use, refer to nonreligious ideas or entities, and are accompanied by inscriptions that, whatever the religious tone of their phraseology, do not generally include the phrase b-ism allâh, "in the name of God." There is no reason to suppose that emblemization is a result of aniconism in the religious sphere. How does it relate to the history of figural representation in the art of the pre-Islamic world?

In Antiquity figural representation was turned to several uses, which may overlap: generalized representations of types; images of identifiable characters such as gods, personifications, or literary characters; portraits (which are also identifiable characters but have features based more closely on real life); scenes from literary texts or stories, presented separately or arranged in cycles; and scenes relating directly to real life, again arranged separately or in cycles. Identifiable characters in Islamic art do not include gods, of course, but may include personifications or literary characters. Portraiture, too, is hardly to be found in Islamic art (one might cite author portraits and early coins, but these are continuations of antique types). But then portraiture had diminished in importance already in Late Antiquity. Emblemization is a prime feature of Late Antique art, and it is convenient to cite as examples the emblemata of floor mosaics, mosaics that contain the seeds of the Islamic artistic interest in geometry as well as those of the arabesque (Floor mosaic from al-Jîm, Tunisia, fourth century).

A more interesting comparison, however, is between the emblemized Islamic cycles and antique narrative cycles.

Richard Brilliant, in dealing with visual narratives in ancient art, points out the importance of "scene making" for conveying the story being told. [Note: Visual Narratives: Storytelling in Etruscan and Roman Art, pp. 17--18. ] Narrative can occur within a scene or in groups of scenes. In any single scene visual details (props or gestures) identify the characters and their setting, and if these details do not specify the event being depicted the characters must exist in a composition that indicates or refers to a known story. A set of scenes permits---and demands---that the characters change their poses and settings from one scene to the next for the viewer to infer action. By contrast, emblems are not narrative. They employ visual detail only to identify the character, not to indicate action or setting. Furthermore, narrative cycles derive their full meaning as cycles from their context and arrangement: in the most complex cases, cycles of pictures may be related to each other in ways derived from or paralleling literary devices and the exposition of these literary devices in the classical literature on literature. [Note: Ibid., ch. 2. ] Thus pictures from different narratives may be juxtaposed in order to suggest a single theme that connects them.

There seems to be no Islamic parallel to this phenomenon: the images in the cosmic cycles are linked by what they denote, not by what they connote. The images of the princely cycle seem to exemplify rather than denote activities, but they refer to no narrative or text. The apparent lack of connection between cycles appearing together on one piece tends to show that each Islamic cycle has a single unified set of connotations.

The reason for this difference between antique and Islamic figural cycles is that the Islamic cycles are not narrative: even the princely cycle does not refer to events in a literary text or orally recited composition, although such works doubtless provided some of the connotations the princely cycle evoked. The emblems are just that: they are not scenes. Instead the figures in a cycle, taken together, refer broadly to all stories and associations connected with the cycle's subject. Individual figures from the princely cycle, or individual animals, appear also to have the same broad field of reference. The mechanism of reference is the same whether the image represents an actual human literary character or personifies a heavenly body or an abstraction. [Note: For an abstraction see Willy Hartner, "The Pseudoplanetary Nodes of the Moon's Orbit in Hindu and Islamic Iconographies." ] The relationship of the viewer to the image could have been guided by the context of the image, e.g., its appearance on a vessel that was used in a particular way, though this remains to be shown. This contextual influence must often have occurred with antique portraits or cult images, too (both of which are nonnarrative arts of Antiquity that nearly disappeared in Islamic civilization). But under such circumstances there is no real relationship between literature and art, since there is no strong relationship of the image with any specific text. The lack or avoidance of narrative in Islamic art made inaccessible a whole realm of humanistic artistic reference familiar to Western culture. But recognizing this lack allows us to understand that figural art did not vanish from the Islamic world but was turned to a different purpose than it had been in Antiquity. What disappeared was not figural representation but the use of figural representation to show human actions and states of mind---in the Islamic era these were the domain of poetry. Cycles of narrative were assembled in writing instead of in figures, which ceased to be either narrative or illustrative.

The illustration of the Shâh nâmah in manuscripts and the portable arts is again the great exception, to which can be added the smaller ones of the illustration of manuscripts of the Maqâmât of Abû Muhammad al-Harîrî, a text of quintessential Arabness, and the Arabic translation of the Kalîlah wa Dimnah. There are no illustrated manuscripts of either the Maqâmât or `Abd Allâh b. al-Muqaffa`'s Arabic translation of the Kalîlah wa Dimnah before the thirteenth century. [Note: Grabar, Illustrations of the Maqamat, ch. 2; EI 2, s.v. "Ibn al-Mukaffa`." ] In all these cases, the illustrations are at least at times narrative in intent. Scientific manuscripts may again be excepted, such as Ibn al-Razzâz al-Jazarî's Kitâb fî ma`rifat al-hiyal al-handasîyah (Book of Ingenious Mechanical Devices), written in the early thirteenth century in the Jazîrah (roughly, northern Mesopotamia), a region that produced many of these illustrated manuscripts, evidently as a result of an upsurge in scholarly activity. These manuscripts are related to interest in their texts, not interest in art. At this time the region produced a noted series of figural coins, however, and it seems to have had a distinctive and unusual bias toward figural representation.

These exceptions are the clue to why no narrative cycle ever extended itself across the whole Islamic world, as the arabesque did. Islamic civilization was a supraethnic one: its religion, law, and science transcended the ethnic and linguistic groups that became Muslim. On this level its art was aniconic and was lacking precisely in narrative. In the absence of a high-level figural religious art, within any one national or language group the story-telling urge was associated with that group's own literature: the Shâh nâmah was of no interest to Islamic Spain, nor the Maqâmât to the post-Mongol East. Figural art blossomed within language-defined cultural subdivisions of the Islamic world, and particularly in manuscript illustration, which has a naturally close relationship to texts. That is why the mosque remained aniconic: it belongs to the supraethnic level of the civilization, to another sector of personal experience. Of course the decoration of the mosque need not have been identified with one level of the culture or another, and mosques were built in all the regional architectural styles. More to the point, a large portion of the best Islamic art, made for secular as well as religious contexts, has no narrative figural representation, and much of it no figural representation at all. The exception to this rule is the princely cycle, which occurs in art from Spain to Afghanistan. The universality of the princely cycle was possible because it was unconnected to religious life but closely associated with the high culture of the ruling groups, whatever their nationality. To the extent it pictures directly the life of these groups it descends from its Late Antique prototypes.

As in Late Antique floor mosaics, with the elimination of narrative in favor of emblemization the focus of the artisan shifted from the figural representation to the framing decoration and its layout. These aspects of the Islamic object gained the intrinsic interest formerly held by the figures in antique narrative art.

An Art Without Narrative

Does the relative absence or unimportance of narrative figural art make Islamic art something radically different from its cousin traditions? Since there was no figural representation at the supraethnic, integrative level of Islamic culture, the focus of art was different for the Islamic world than for Antiquity or mediaeval Western Europe. But that is not the same as a radical difference in the pursuit of artistic form. Instead this difference in focus points out the different importance in Islamic art of the abstract elements of art that have come to be appreciated in this century for themselves, but which are also the stuff of art at all times: line, color, and texture, for example; proportions are just as important in Islamic art as in any other art. The question all this leads to is: what is one meant to attend to in Islamic art?

In an antique narrative cycle each scene can be the subject of individual contemplation. In the Islamic piece it is seldom the individual figure or emblem that is presented for extended contemplation but rather the arabesque or geometric decoration that enframes it, or upon which it is imposed. And like the arabesque or geometric design, in which a small segment implies the whole, each element of the Islamic cosmic cycle also implies the complete cycle; one need not turn the ewer around to see all the signs of the zodiac in the way one must turn a similar antique piece, or the beaker decorated with Shâh nâmah scenes in the Freer Gallery, to follow the narrative. [Note: Marianna Shreve Simpson, "The Narrative Structure of a Medieval Iranian Beaker." ] It seems to me that in Islamic art abstract elements such as geometry serve the integrative function of representations in classical art: they tie together the whole work of art by belonging to the same conceptual category. They are the visual though not the intellectual "subject matter" of Islamic art. Geometry and the arabesque are the intended focus, and the meaning, if any, to be assigned to them should be visual and artistic, rather than literary or covertly religious. What made Islamic art accessible to its contemporaries was its abstract values and its visual familiarity, its play with conventional forms (cf. classical architectural moldings, which remained standard for centuries to provide familiar forms that could be manipulated creatively). Why else does the arabesque remain vineous so long? Surely not because of any literary association, such as with wine, but because a successful arabesque is an inventive variation of familiar patterns with familiar elements, and those patterns and elements need to remain recognizable if the variation is to have any point. In the same way, in every regional style of Islamic architecture certain combinations of decoration, and combinations of decoration with other formal elements, became standard because they had become the limits of a certain visual genre.

This abstract aspect of Islamic art does not necessarily give it the intellectual scope and power of antique art, which was derived from its human interest as well as its own abstract qualities. It is very hard to believe that the cosmic and princely cycles supplied this kind of human interest instead. In the end the values we assign to art come down to what our culture is interested in. Western art has been fixated on the human figure since Antiquity, and on the relation of art to verbal narrative. Islamic art moved away from this classical heritage by leaving artistic narrative to speech rather than displaying it in pictures, which alluded to the world of ideas by way of emblematic reference or explicit words: cycles and inscriptions. The allusions were to a literary tradition that like so many others was rich in metaphors and learned by memory, so that it could be called to mind at any time. In contrast to antique visual allusions, which were so often to a particular passage in a text, these were evidently invitations to the user to specify his own associations within the general realm indicated by the emblem. Islamic culture did not lack human interest, but it did not look for it in what we, from a modern standpoint, are pleased to call art, a category not really applicable to the subject at hand.

As art, nonfigural Islamic art is concerned with control, balance, and regularity; it is severely self-referential. The focus is on pattern rather than organic forms. These are not Wölfflinesque terms meant as poles: there is plenty of freedom of execution, ambiguity, and harmonic imbalance in Islamic art. But even when figures are involved, the viewer's reflexive comparison of them to the real-life appearance of people (that is, the process of perceiving the representation as an artistic representation) is not what gives the figures artistic interest. It is the rhythm of their poses and their placement as a cycle within the larger work of art that make these figures worth regarding: there are few cycles without substantial decorative enframements.

The development of an Islamic art distinct from other postantique Western traditions was the result of a profound cultural redeployment, in which Arabic-speakers steeped in Late Antique culture reorganized the Near East's network of cultural contacts and slowly cut loose from its Late Antique past, a process to be explored in the next chapter. While their language---Arabic---defined the limits of their new empire, an aniconic religion left no central cultural story to be told by visual narrative. This does not mean that early Islamic civilization was essentially hostile or indifferent to images. If one considers the mediaeval West or Byzantium and ignores their religious art, in which stories are narrated constantly and even urgently, there is hardly more figural art to be found than in the Islamic world. It was not until the Renaissance that the humanistic focus of the art of Antiquity was regained, and the humanistic culture that had made antique art possible had died long before Muhammad was born. There is no dearth of secular Islamic figural art, only a dearth of religious figural art, a field in which there is much cultural variation around the world. The nonfigural art of the Islamic world is much more than mere decoration, it is a serious and visually intriguing exploration of pattern. To the extent it has figural representations, they are either artistically subordinate to the rest of the decoration, whatever their meaning, or tied to specific linguistic traditions, such as the illustrations of the Shâh nâmah.

(back to: Palm Tree Books Home Page)

(back to: description of Five Essays on Islamic Art)