sign in sign up
<<
Home , Amman Citadel, Umayyad architecture

PART 2

PALACES OF THE EARLY ISLAMIC

(SEVENTH-TENTH CENTURIES)

UMAYYAD PALACES RECONSIDERED BY OLEG GRABAR

5 THE LATEST, MOST COMPLETE, AND MOST AUTHORITATIVE Islamic foundations or ornamental inscriptions. survey of early -Allan's re- For Qusayr Amra, as we shall see shortly, there is working of Creswell's standard history-contains a valid presumption for an Umayyad dynastic seventeen partly documented buildings datable patronage on the basis of the paintings decorat- between 690 and 750, the main decades of Umay- ing the monument. As to Mshatta, it is primarily yad rule, and fitting into the general category of its inordinate size and its peculiar decoration that palaces. It is a loose category including nearly all make it reasonable to assume that only princes, foundations with living accommodations for and in all likelihood ruling princes, had access to which a reasonable assumption can be made of the funds and personnel necessary for its plan- sponsorship or use by the state (a daral-imara for ning and construction, and for a completion instance), by the ruling dynasty, or by members of which never took place.6 the Arabian aristocracy associated with the Umay- In short, we have no direct knowledge of these yads.1 buildings as palaces built for ruling Umayyad Out of these seventeen buildings, five-al-Mu- princes, and I shall return in my conclusion to the waqqar, Rusafa, Qastal, Qasr al-Tuba, and Tulul hypotheses which can be derived from this appar- al-Sha'iba-are too poorly known or too poorly ent absence of clearly cut, written or archaeolog- preserved to allow for significant conclusions,2 ical, external labels. Nor can we really argue that two--'Anjar and -pose avery complicat- their shape, the functions they imply, or their ed set of problems in the interpretation of ar- decoration make it necessary for all of them to be chaeological evidence, which, in my opinion, Umayyad creations. In fact, as I reread some of does not necessarily lead to the Umayyad dates the papers, articles, and books I wrote on them proposed by the investigators of these sites, and over the years, I realize that there may have been four-, Qasr Kharana, Jabal Says, and a major methodological flaw in parts of my rea- Qasr al-Hayr East-are not palaces or even royal soning. Following very respectable masters like or dynastic residences, at least not in my judg- Sauvaget, Herzfeld, Stern, and in his own way ment, although this judgment is not necessarily Creswell, I consistently tried to fit all these build- shared by all scholars.3 ings into a single pattern by assuming that they all This process of elimination leaves six relatively reflected the same architectural and functional well-known buildings-, Khirbat Minya, type: a villa with living and formal or official Qusayr 'Amra, Qasr al-Hayr West, Khirbat al- quarters arranged around a porticoed court in Mafjar, and Mshatta-which can be reasonably standard square buildings with a single, usually assumed to be dynastic foundations of the Umay- massive, entrance; a bath, most of the time an yad dynasty. For Kufa the evidence is essentially independent building; a ; a variety of historical in the sense that written accounts of service areas; and possibly formal gardens or events from the last decades of the seventh centu- orchards. The type would have developed within ry and the first ones of the eighth justify the the rich tradition of secular architecture in the interpretation of an excavated building as the dar late-antique Mediterranean area. And, following al-imara of the Umayyad governors.4 At Khirbat primarily Sauvaget but frequently using docu- Minya, Khirbat al-Mafjar, and Qasr al-Hayr West mentation gathered much earlier by Lammens (all of which were first thought by archaeologists and Herzfeld, 7 I further assumed that this partic- to be pre-Islamic Christian foundations), inscrip- ular type could be fitted into a reconstruction or tions were discovered which suggest that some a model of Umayyad aristocratic behavior. But aspect of the construction of these buildings was that reconstruction of behavior was based on very sponsored or ordered by a ruling Umayyad caliph different sources: written texts in which an orig- or else that a caliph was alive while the building inally orally transmitted poetry predominates; was being built. But not one of these inscriptions archaeological sources involving the interpreta- can be considered as a foundation statement or tion of excavated spaces and other remains; and as decorative writing in the manner of later finally a more complicated source which may be 94 OLEG GRABAR

called "ecological logic" and which consists in they found in spaces which had not previously apparent changes within the lands ruled by the been used for these purposes. Umayyads which require significant investments There are several other, more specific issues of 8 of money and labor. architectural history, like the technology of build- The resulting hypothesis was a simple one. ing for instance, which come up whenever there Umayyad patrons, a nouveau riche class of aristo- is a combination of a set of buildings from a crats from central Arabia, invested in the land restricted period of time, in a relatively limited they inherited or conquered and in a life of space, and with more or less the same patron- varying but usually considerable luxury. The pal- age.1 2 Beyond elementary measurements and the aces illustrate the setting for that life. Variations establishment of the most basic modular princi- and peculiarities in the model were then seen as ple of designing and with the exception of the aberrations reflecting individual needs or prefer- question of brick in and Palestine about ences in taste or behavior for which no explana- which a few things have been written, these tech- tion was likely to be found. A case in point is that nical issues have hardly been touched, however of the twin quarters of Qasr al-Tuba for which important they are for establishing the techno- some romantic or family explanation could al- logical potential of a time. ways be invented and proposed but no proof can A more specific approach to the establishments 9 be found. attributed to the can be derived To the historian of architecture, that is to say, of from a number of puzzles in the architecture and techniques which require considerable invest- decoration of remaining buildings. These puz- ments of funds and differentiated competencies zles led historians of art, of visual forms, to seek an from specifically trained practitioners, the as- explanation in the history provided by texts. Such sumption of types adapting rapidly or slowly to features as the elaborate reception rooms of new patrons was then and remains now a reason- Mshatta and the rich decoration of Khirbat al- able one, for the hitherto unknown sponsorship Mafjar did not quite coincide with the simple- of architecture by rich aristocrats from north- minded image created by earlier historians such western Arabian cities did not necessarily de- as Lammens and Creswell of an easy-going, fun- mand new functions or seek to satisfy new needs. loving, but not very serious life attributed to This patronage could easily accommodate itself freely moving and semi-nomadic Umayyad princ- to the prevailing habits of building and of design- es and assorted aristocrats. Nor is fancy decora- ing within the Mediterranean world, and one or tion necessary for landlords in their manorial two technical innovations from elsewhere-as, and private estates. As a result, several studies, for instance, the covering of stone walls with whether or not they acknowledge the original stucco-easily found a place in an essentially inspiration of Sauvaget for the method involved, traditional system. A typical problem of design sought to describe an Umayyad ceremonial life arose from the new requirement for an oriented and to fit that life within the ruins or reconstruc- place of prayer. The mosque became a composi- tions of whatever remained from alleged palac- tional problem with several early and more or less es.'3 The best among these studies, notably Hamil- successful solutions until buildings apparently ton's meditation on al-Walid and Khirbat began to be oriented in order to meet the neces- al-Mafjar, have provided a very tempting and sity of fitting a mosque into them. Oratories coherent picture, based to a large degree on appear, most commonly to the right of the en- poetry, of an unstable equilibrium between for- trance (Mshatta, Ukhaydir)."° mality and fun, between rigidity and freedom, all The more original aspect of these Umayyad of it bathed in virile sensuality. It is a coherent secular buildings is how they transformed, or at picture in the sense that the poetry of the time least modified, the architectural profile of an and especially the ways in which this poetry was area by introducing the amenities of a wealthier recited in male gatherings seem to fit with the and restricted private life of "masters" into a spaces provided by the ruins. simple landscape of farms and villages with their But there are problems. For instance, too many churches and occasional civic buildings such as of the examples of almost anything in these stud- baths or into old Roman military cities." The ies deal with al-Walid II, who was in many ways an Umayyads, this hypothesis implied, did not really eccentric hardly typical of his rank and of the invent much that is new; they simply putwhatever office he occupied. And there are always dangers UMAYYAD PALACES RECONSIDERED 95 in excerpting examples from literary contexts standards with local variations. The assumption, whose own rules have not been ascertained. In valid for religious arts with an ecclesiastical pa- general, there are many theoretical and concrete tronage, was that models (actual ones on monu- pitfalls in explaining architecture through poet- ments and objects or written ones in manuals) ry, unless, as for instance with much later Persian must exist for any visual program to be executed lyrical poetry, the poetry itself deals with architec- and to be understood. In reality, of course, the ture or, as in the Alhambra, ad hoc poetry adorns secular arts operate quite differently from reli- architecture.' 4 The danger exists because the gious art. It is therefore more fruitful and, in fact, sphere of life and of behavior in which poetry necessary to attempt to understand the individu- appears is the immediate and the evanescent; al combinations of themes and forms located in much ofitwas not written down at the moment of one monument.'8 The establishment of vocabu- recitation, but at the moment when the event in laries and grammars of forms and the definition which it participated was recorded. of typologies are proper endeavors, but because Architecture, on the other hand, is more per- in secular art almost every statement through a manent, unless, of course, we consider decora- given object or monument is a different combina- tion separately from building on the grounds that tion of terms and means something different, some of its techniques-especially painting and typologies must follow rather than precede indi- carved or molded stucco-were themselves vidual analyses. In a nutshell, system follows syn- changeable or could be covered with other mate- tagm.' 9 In the absence of formal manuals and of rials such as textiles.'5 The skin of architecture a doctrine about the arts, the only way one can like that of snakes can be changed, at least in part. eventually reach a sense of the language used in A remarkable feature of the art of Umayyad palaces is first to provide is the quantity as well as the quality of the decora- reasonable hypotheses about every monument tion it exhibits, still in situ, as in the of the by itself. Such hypotheses are all the more impor- Rock and the mosque of , or through tant since every one of the monuments involved masses of fragments, some in place, others recon- provides a range of information which is pecu- structed into their original compositions, many liarly its own. I shall illustrate the point with two still hiding in unopened crates in the basements examples. of museums or exhibited as bulk remains in their The first one is that of Qasr al-Hayr West, for drawers and galleries.' 6 The curious point is that which the last written and drawn statements by the decoration of the main Umayyad palatial the excavators, all of whom are now deceased, establishments has been less well studied than have recently been published. 20 The second one architecture or planning in that period, probably is Qusayr'Amra, the first of the Umayyad dynastic because Creswell spent relatively little time on establishments to have come to light, and the the decoration, much of which was discovered only one which is not in ruins. It underwent a after the first edition of his work. The only partial major overhaul some twenty years ago, but its exception is that of classical art historical issues complete actual appearance today is, rather dealing with genetic sources which have occu- shamefully, not yet available in print.2' pied a few scholars, especially in the early decades I shall be brief about Qasr al-Hayr. In addition of the century.' 7 to the sculptures and paintings which are well This absence of studies is all the more peculiar known, have been visible for over a generation in since the originality and the variety of decoration the Damascus Museum, have been mentioned in in Umayyad palaces are quite astounding. Yet, general books and manuals, and have even been precisely because of its variety and originality, the subjected to elementary scholarly analyses, we study of this decoration leads to a number of are presented with painted fragments which were methodological problems. Among approaches exhibited occasionally but never published and which have been tried is one I used for the first with stone and stucco sculptures of personages time some forty years ago, and which was proba- known only to visitors to the storerooms of the bly even more dubious than the procedure I used Damascus Museum.22 It is nearly impossible to for architectural planning and design. Following reconstruct how these paintings and sculptures the model of Christian and Buddhist art, I sought fitted anywhere, except for a few fragments on to find in the decoration of walls, floors, and the facade of the palace rebuilt in the Damascus ceilings a set of types, that is to say, of common Museum and for two large paintings found on the 96 OLEG GRABAR

floor of the staircases. 23 It is, therefore, nearly not lead one to interpret the decoration on the impossible to propose a program for these imag- walls within a social, official, or personal setting, es, or even sequences of visual impressions with but it does explain how certain combinations of some iconographic or expressive meaning. This forms ended up in a striking location of the is so even when the sources of a motif are clear. Syrian steppe. We become informed on the move- For instance, two sculptures found at Qasr al- ment of works of art and of artistic motives across Hayr resemble representations of rulers in Byzan- western Asia rather than on their meaning in the tine and Sasanian art.2 4 But if we argue, as I and setting they eventually adorned. This is so be- others have done, that, because these had been cause the heterogeneity of the vocabulary of the forms for the representation of pre-Islamic rul- palace seems to me to be its only characteristic ers, they must now be representations of Umay- which lends itself to a positive conclusion. yad princes, then we also have to argue that a Let me turn now to the bath house of Qusayr sculpture with obvious Palmyrene funerary asso- 'Amra which is only one of a larger group of ciations (fig. 1) should be connected to some constructions including now-ruined living quar- funeral theme in Umayyad times.25 What then is ters, a watch-tower, and a hydraulic complex. Its it doing on the facade of the building? discovery a century ago affected nearly all histori- A similar position of iconographic skepticism ans of Islamic culture because of the paintings can be argued for almost every fragmentfound at which adorn the bath and whose subjects seem to Qasr al-Hayr West. But the same skepticism need reflect the life of princely leisure. Now we have not apply if one tries to identify, even tentatively, the advantage of possessing all the information not so much the subject matter as the sources of we are ever going to get and of having it in situ.2 9 inspiration for the images. Within the limits and It is possible to reconstruct the probable mod- restrictions of existing information and investiga- els for many of the paintings. Just as at Qasr al- tions, the genetic pool from which these images Hayr, objects, mostly from the Mediterranean, came may be the only kind of meaning or seman- but also from the Sasanian world, had a clear tic range available for the decorative motifs from impact on the paintings. Moreover, as I have Qasr al-Hayr West which can be traced with some already pointed out elsewhere, there were also certainty. The most interesting and most unex- many images inspired by a direct perception of pected point about this genetic pool seems to be the actual setting of the building and of the the close relationship between the paintings and events which took place in it. This was not done sculptures from Qasr al-Hayr and the art of Soghd in order to present the illusion of a physical in Central Asia, and even the Buddhist and Man- world, but in order to commemorate events which ichean arts found farther east in the Tarim ba- had either taken place or been recalled there.3 0 sin.26 The question is how did themes and styles In fact, those images which are new, without a from the northeastern frontier of the Muslim clear model, seem often to be more caricatured empire (and even beyond) reach Syria, when and more awkward than the conventional repre- none of the patrons of Syrian and Palestinian sentations in the bath. It is as though someone, a buildings had ever set foot there and the move- patron no doubt, had requested topics that may ment of artisans, while not impossible, seems have been a bit outside the competence of the technically unlikely at this time. The answer lies painters, for the whole place has the appearance almost certainly in transmission through objects- of a lively amateur exhibition. The more impor- silver objects sent as tribute or as gifts, textiles, tant point, however, is that even here, where perhaps rugs. Books are not likely to have been about five-sixths of what had once been there is important at this time, a few decades before the preserved, no clear sense of what is shown emerg- appearance of paper through the same route, but es. This may well be an example of the art histor- painted bark and bone could have been among ical principle that it takes two to tango, or that a the marvels brought from the east.27 It is also unique building is almost impossible to under- through textiles and works in metal that Mediter- stand, especially when additional contextual evi- ranean motifs can be imagined to have reached a dence is missing. remote site like Qasr al-Hayr and through silver It is tempting to give up,just as it is tempting to objects that Iranian ones would have become excerpt something immediately understood like available.2 8 the Six Kings and to elaborate from there on Thus the evidence from Qasr al-Hayr West does some grand iconographic message." So far, to my UMAYYAD PALACES RECONSIDERED 97 knowledge, none of these approaches hasworked, 'Amra is the extraordinary number and variety of although admittedly so little has been written women represented on its walls in a truly minus- about this fascinating little place that even meth- cule space: half-dressed ladies around the prince odological judgments are dubious. What can be (figs. 5 and 6), Lady Nik6 (fig. 4), the tall lady by proposed is something akin to chemical tests the pool (fig. 3), nude dancers (fig. 7), a heavily used to determine the composition of some prod- dressed dancer (fig. 8), another standing naked uct, that is, try possibilities until something works. woman (fig. 9; now in the Berlin Museum), a Here are two examples. woman apparently alighting in front of a man The first one is difficult to illustrate because, (fig. 10), meditating personifications identified when the photographs I used were made,3 2 I was by Greek inscriptions (fig. 11), busts of well- concerned with recording two-dimensional paint- dressed women in the central nave of the main ed panels on walls. I did not try to record photo- hall as though looking at the goings-on below graphically or even to determine on the spot the (fig. 12), a cogitating nude woman in one of the impressions made on users and visitors from side rooms for bathing (fig. 13), and partly dam- various points of view logical and unique to this aged paintings of totally naked women and chil- particular place: the entrance, doorways, the floors dren also found in one of the small bathing on which one sat, the apse or throne niche in the rooms that seem to depict domestic activities that center of the building, and so on. would have taken place in precisely such a room Pending verification at Qusayr 'Amra itself, I (figs. 14 and 15). propose two visual impressions as correct assess- Almost none of these representations of wom- ments of the building's impact. One concerns the en finds adequate parallels in comparable or wealth of pictures hanging on the wall: there are reasonably comparable artistic traditions, except so many of them that one can hardly make sense occasionally in details like the stubby and awk- of them as a whole, and one moves almost auto- ward legs so typical of Coptic textiles, poses and matically to consider them individually. Just as in gestures clearly belonging to the ways of repre- a museum, one is overwhelmed by the quantity of sentation in late-antique art, and occasionally the pictures, and one can only handle them one at a memory of a classical personification or of a time. I am unclear about the psychological impli- Sasanian symbol. cations of this cascade of ever-present images. The point I would like to make, however, is a The second impression is that the first and richest different one. While the presence of so many hall of the bath, usually interpreted as a throne women predominates at Qusayr 'Amra, it is strik- hall, is dominated by the image of an enthroned ing that they are so different from each other. ruler in the apse and by a series of large panels in Sensuous performances cohabitwith highly prop- the rightvault ending at the vault's back wall (fig. er ones, possible narratives coexistwith apparent 2). The comparable area in the left vault is cut up scenes of leisure and domestic life, and trite into smaller sections and visually less effective personifications are found alongside concrete from a distance. references. Should one try to argue that what is The question is whether this visualjudgment, if shown or reflected in Qusayr 'Amra is the world accurate,justifies arguing that the woman by the of the women whose power, both political and pool (fig. 3), the Six Kings, and the "Lady Nike" cultural, was, in early Islamic times, consider- (fig. 4)33 are, with the prince in the back of the able? Or should one interpret all of this as a men's apse, more significant than other images. One locker-room view of the world? A preliminary could propose several other similar readings of argument can be made for either possibility, and the walls from a variety of points of view corre- maybe others exist as well. The full appreciation sponding to the likely uses of the building. The of this strange and fascinating museum of rather result of this kind of investigation would be to mediocre paintings most probably requires the establish hierarchies according to the way the elaboration of a series of categories of under- paintings are perceived, but I am not sure that standing, one of which may even be the tradition- such hierarchies would not themselves become al art historical one of specific iconographic arbitrary constructs rather than useful explana- meanings and of an equilibrium of sources from tions. many origins. My second approach is even more subjective, at What conclusions can one draw from these re- least in its premise. A striking feature of Qusayr marks? The most important one is that Umayyad 98 OLEG GRABAR palaces, or at least a small number among them, A second discourse concerns the Umayyad contain a considerable amount of data which phenomenon. It can explain how a new aristocra- should at this time be considered syntagmatical- cy from the outside established itself in ancient ly, that is, in terms of the unit in which they are lands, but the more tantalizing question is wheth- found, rather than systemically, in terms of the er the specific phenomenon of Umayyad palaces separate histories of each example. Before re- played any significant role in the development of turning to types and to broad vistas, it may be a new art of princes eventually associated with worthwhile to plunge into the depth of each Muslim rulers. One would also be able to return palatial establishment, to study all of its details, to with better information and fuller perception to imagine how it was built and what went on in it. standard problems of early like the This sort of analysis should not stop at informa- formation of a princely cycle and the myths that tion provided by the monument or by written developed around the Umayyads. At this time, I sources about its time, nor even by the sources am doubtful that the Umayyads were as impor- which can be identified for any part of it. It is tant and innovative in secular art as they were in perfectly appropriate to bring to bear everything the art of the mosque, but the matter will be from gender studies to technological analyses discussed over the years to come, I hope. which can develop a grip on the monument. Such Finally, there is a thematic discourse. The a monographic approach, like a proper trench in Umayyad establishments belong somewhere in an archaeological investigation, would at least one or more schemes of royal and aristocratic begin to show that each one of some six buildings living and behaving. Most of them are private can bristle with specific references and interpre- palaces which constantly use public artistic trig- tations, acquiring a full personality of its own. gers-like a decorated facade, for example-to The acquisition of such precise information audiences who are impossible to identify. But together with a range of possible interpretations although their most brilliant features are inside, would allow for the first time three additional to be treasured and enjoyed and not to be shared, discourses. One is the art historical one of wit- they lack clear spaces for private and family living. nessing, in the first half of the eighth century, a They seem different from anything else and yet series of transfers of forms and of meanings, as can be related to French Renaissance palaces or new sponsors and new resources contribute to to nineteenth-century villas in the West perhaps the recomposition of older forms, on the one more easily than to their Roman ancestors. In hand, and to the development of new meanings, short they have a significant contribution to make on the other. The most obvious example of the to the broader understanding of what rulers latter is the ceremonial pastime, which acquires sought to do with their architecture and its deco- new dimensions with the luxurious baths of Khir- ration: to proclaim and display their glory or bat al-Mafjar or the grandiose composition of perhaps to create places in order to hide and Mshatta. enjoy their wealth. UMAYYAD PALACES RECONSIDERED 99

Notes summarized in her thesis for a Doctorat de troisieme cycle from the University of Aix-en- This paper is neither a coherentreevaluation of Umayy- Provence, Qastal, chiteau umayyade (Aix, 1984). ad palaces nor quite a repetition of remarks begun For Rusafa, the pertinent results of the renewed some forty years ago for my doctoral dissertation at German excavations have not yet, to my knowl- Princeton University. It is rather a moment in a con- edge, been published. I am sure that departments tinuing meditation on the first monuments of Islamic of antiquities and private notes of travelers and art, perhaps with a different set of questions and inter- archaeologists contain large quantities of useful pretations than I and others developed in the fifties of information. this century. 3. For Qasr Kharana, see the rather complete study 1. I A. C. Creswell, A Short Account of Early Islamic and original interpretation by Stephen K. Urice, Architecture, revised byJames W. Allan (Aldershot, Qasr Kharana in the Transjordan (Durham, N.C.: 1989). The buildings are in the order in which they American Schools of Oriental Research, 1987). appear in Creswell, and in what is also supposed to For the other three sites, see the comments provid- be their chronological order: Kufa, Khirbat al- ed in Oleg Grabar et al., City in the Desert (Cam- Minya, Jerusalem, Qasr Kharana, Qusayr Amra, bridge, 1978), 151-52. On Jerusalem the only Jabal Says, cAnj ar, al-Muwaqqar, Qasr al-Hayr West, easily accessible source consists in the very impres- Rusafa, Qasr al-Hayr East, Amman, Qastal, Khirbat sionistic and personal accounts of Meir Ben-Dov, al-Mafjar, Mshatta, Qasr al-Tuba, Tulul al-Shaiba. In the Shadow of the Temple (New York, 1982), esp. I have not included in this list places like Qasr 273-321. More detailed views on these buildings Burquc or Bayir, where there were Umayyad foun- will appear in my forthcoming book on the early dations but which are too damaged to allow for medieval city of Jerusalem. In my view, the date significant architectural rather than archaeologi- and functions of the have not yet cal or geographical conclusions. Nor does this list been resolved, in spite of the assiduous and metic- include buildings like the Khadra in Damascus ulous works by A. Almagro, El palacio omeyyade de which are known to have existed but about which Amman (Madrid, 1984), and by Alistair Northedge too little can be reconstructed, even though some in his unpublished dissertation on the topic com- of these constructions, like the Damascus palace of pleted in 1982 at the University of London. 'An- Mu'awiya or Wasit's palace of al-Hajjaj, in their jar's problem is quite different. Too little has been time or eventually acquired a mythical value. Both published of whatwas discovered there and its size of these buildings are mentioned in Creswell-Al- as well as apparently commercial functions do not lan, but there are many places like , Basra, fit with what is known of Umayyad economic pol- Merv, and probably other cities for which written icies in this area as opposed, for instance, to north- sources yield some information. In the absence of ern Mesopotamia where new trade routes can be a reasonably complete survey of these sources, assumed. Furthermore, its alleged palace does not occasional accidental references would have been look like known Umayyad ones. misleading. I have also not counted in the list of seventeen "palaces," buildings like Qasr Hallabat 4. Most of the immediate evidence on Kufa is found in or Raqqa in Syria, where recent excava- in Creswell's larger volumes, pt. 1,46 ff. But, since tions have found significant Umayyad uses and Kufa remained as a major and very lively city for modifications of older buildings or Umayyad sec- several centuries, it should be possible to analyze ular buildings like the ones being excavated in and interpret the continuing use of this building, Aqaba or Humayma in Jordan which may have a possibility which cannot be pursued with the been palaces, but whose investigation is still in- evidence which has been published. complete. Finally, I have left aside the full investi- gation of the long list of possible Umayyad prince- 5. At Qasr al-Hayr West and at Khirbat al-Mafjar ly establishments made, partly as a heuristic fragments of statements written casually on mar- exercise, by Jean Sauvaget, "Observations sur les ble plaques mention the name of the caliph monuments omeyyades," Journal Asiatique 231 Hisham, once as the sender of a message, the other (1939). This list contains many places which are time as the recipient of one. Both statements merely settlements without palatial functions or contain the names of otherwise unknown and whose exact dates are still under discussion. unidentified individuals, and it is extremely diffi- cult to imagine the circumstances which would 2. Actually quite a bit is known about Rusafa and have led to the writing of these messages on mar- about Qastal, but much of the available infor- ble plaques presumably destined to be used in mation is not easily accessible. For Qastal, for the buildings. They are correctly interpreted instance, the activities of Patricia Carlier are as graffiti, not as inscriptions. See Daniel Schlum- 100 OLEG GRABAR

berger, "Kasr el-Heir," Syria 20 (1939): 373, and have small outside the main residential R. W. Hamilton, KhirbatalMafar (Oxford, 1957), complex. 43. The written fragment found at Khirbat al- Minya is possibly a genuine inscription since it 11. The landscape of Christian times can best be seen begins with the traditional invocation of God's in books like G. Tchalenko, Villages antiques de la name and with the words mimma amarabihi, "what Syrie du Nord (Paris, 1959). Much has been accom- has ordered," the typical beginning of thousands plished since that time, and numerous excavations of medieval inscriptions on objects or buildings. and surveys, especially in Jordan, have modified The problem is that the stone with the inscrip- some of the views developed a generation ago, tion is reused in the palace and there is no way especially around the extent and character of the of knowing to what it referred. See O. Put- cultural continuity of the area. Consult the reports trich-Reignard, "Die Palastanlage von Chirbet found in various archaeological journals. For a el-Minje," Palastina-Hefte des Deutschen Vereins recent publication which touches on many of the vom Heilige Lande 17-20 (1939). The three pertinent problems, see P. Canivet and J.-P. Rey- documents are quite different from clear-cut Coquais, La Syrie de Byzance d l' (Damascus, foundation inscriptions like the one from Qasr 1992). al-Hayr East. 12. See, as an example, the work ofJ. Ward-Perkins on 6. Elsewhere I have argued for an early Abbasid date vaults in University of St. Andrews, The GreatPalace of for Mshatta (Oleg Grabar, 'The Date and Meaning the Byzantine Emperors, 2nd report (Edinburgh, of Mshatta," Dumbarton OaksPapers41[ 1987] ), but 1958). On a much smaller scale similar analyses it is only slightly later than the one usually given to have been carried out on Palestinian churches and the palace and does not affect the reasoning of this synagogues. paper. See now the excellent summary byVolkmar Enderlein and Michael Meinecke, "Mshatta-Fas- 13. J. Sauvaget, La mosquie omeyyade de Midine (Paris, sade,"Jahrbuchder BerlinerMuseen 24 (1992), with 1947), is the first book to have used this approach a discussion of epigraphic information I did not of blending written and visual documentation. For know. later studies, see Oleg Grabar, "Notes sur les cer- monies omeyyades," in Studies in Memory of Gaston 7. There is no one statement of Sauvaget's definitive Wiet, ed. Myriam Rosen-Ayalon (Jerusalem, 1977); opinions on the subject of Umayyad palaces. The Robert Hillenbrand, "La Dolce Vita in Early Islam- most important ones are "Remarques sur les mon- ic Syria," Art History 5 (1982); and especially Rob- uments omeyyades," JournalAsiatique231 (1939); ert Hamilton's admirable Al-Walid and HisFriends and "Chateaux omeyyades de Syrie," Revue des (Oxford, 1988). In reality the historiography of itudesislamiques39 (1967), published posthumous- this particular approach is older than Sauvaget, as ly. But he made remarks on these "palaces" on Lammens in particular and Herzfeld were inspired many other occasions. H. Lammens, "'La Badia' et by buildings to search for texts, but they did not la 'Hira' sous les Omayyades," Melanges de laFaculti look for ceremonies and for behavior, something Onrientale, Univ. St.Joseph, Beirut, 4 (1910); E. Herz- which has been compelled on a scholarship affect- feld, "Mshatta, Hira und Badiya,"Jahrbuchderpreus- ed by the decoration which came out of excava- sischen Kuntsammlungen 42 (1921). tions for the most part.

8. One example of ecological studies with implica- 14. For the Alhambra, see Oleg Grabar, The Alhambra tions for Umayyad palaces isA. M. Watson, Agricul- (London, 1978). To my knowledge no study has tural Innovation in the Early Islamic World (Cam- been devoted specifically to the references to ar- bridge, 1983). chitecture in Persian poetry or in Ottoman pane- gyrics, whether descriptions of buildings or meta- 9. Such is essentially my position; see Oleg Grabar, phors and images using architecture. I owe to The Formation of Islamic Art, 2nd ed. (New Haven, Professor Necipoglu the reference to A. S. Levend, 1989), and Richard Ettinghausen and Oleg Gra- TiirkEdebiyatznda ehrengizlerveSehrengizlerdestanbul bar, Islamic Art 650-1250 (London, 1985). (Istanbul, 1957).

10. A thorough study of early solutions to the problem 15. A typical early Islamic example of changing tex- of finding a place for mosques in a language of tiles is that of the Kacba in which was cov- planning that did not have them lies beyond my ered anew every year with fancy embroideries; al- purposes here, but I do want to note that many Azraqi, AkhbarMakka, in Geschichte der Stadt Mekka, early Islamic buildings (Ziza, Qastal, Umm al- ed. G. Wuistenfeld (Leipzig, 1858), 1, 104 ff. Walid, Qasr Hallabat, all in Transjordan) Another kind of change using textiles were the UMAYYAD PALACES RECONSIDERED 101

temporary modifications made for special occa- the one held at the Muse du Monde Arabe in sions, as when the Baghdad palaces were festively 1990-91 which focused on Umayyad palaces. Their transformed through textiles for the arrival of a catalogues provide a few additional pictures and Byzantine embassy; see, among many places, Gra- minimal comments. bar, Formation,168 ff. For a more general theory of textile aesthetics, see Lisa Golombek, "The Draped 23. DanielSchlumberger, "Deux fresques omeyyades," Universe of Islam," in Content and Context of Visual Syria 25 (1946-48); Richard Ettinghausen, Arab Arts in the Islamic World, ed. P. Soucek (University Painting(Geneva, 1962), 35, 37. Park, Penna, 1988). 24. Oleg Grabar, "Islamic Art and Byzantium," Dum- 16. The principal structures with decoration are Khir- barton Oaks Papers 18 (1964). bat Minya (practically only mosaics), Qasr al-Hayr West (paintings and sculpted stucco), Khirbat al- 25. It is, of course, true that as an exercise in pure Mafar (mosaics, stucco, and sculpture in stone connoisseurship the Qasr al-Hayr sculpture differs and in stucco), Qusayr cAmra (paintings), Qastal in many details from actual Palmyrene works. Does (mosaics), and Mshatta (stone sculpture, a few it matter? How different must an imitation be in stucco fragments). The basic bibliography on all order not to be meaningful as an imitation? Con- of these except Qastal can be found in Ettinghau- versely, are meanings carried together with forms? sen and Grabar, IslamicArt andArchitecture, 26-70. For Qastal, see above. 26. There is no convenient access to the immensely rich treasures of early medieval art from Central 17. Genetic discussion of the origin of motifs was and Inner Asia. A very general survey is Tamara central to much of what has been written on the Talbot Rice, Ancient Arts of CentralAsia (New York, facade of Mshatta and on the mosaics of the Dome 1965). Rich and accurate, but geographically more of the Rock. Most of it derived from the complex restricted information is found in the published ways of analyzing the decorative arts developed catalogue of an exhibition that never took place, out of grammars of ornament on the one hand Culture and Art of Ancient Uzbekistan, 2 vols. in and the theoretical studies of A. Riegl on the Russian and English (Moscow, 1991) .A little-known other. and unexpected source of major importance for Qasr al-Hayr as well as Khirbat al-Mafjar is Vara- 18. I first approached the subject of a general system khsha in Uzbekistan, which has friezes of animals in Umayyad art in my dissertation, "Umayyad Art strikingly similar to the Umayyad ones; see V. A. and Ceremonies," Princeton University, 1955. It is Shishkin, "Arhitekturnaia Dekoratsia Dvortsa v outdated in manyways and in error in some, but its Varahshe," Hermitage Museum, Trudy Otdela Istorii impact can be found in most of the works quoted Kultury i skusstva Vostoka,vol. 4 (Leningrad, 1947). in n. 13. For further examples see my comments on the paintings of Mafjar in Hamilton's publication of 19. The major exception to the rule lies in coinage, the palace and A. von Le Coq, Die Buddhistische where both written sources and existing series of Spdtantike (Berlin, 1922-33), :pl. 20 and ff.; 5:pls. coins emphasize the necessity for a system of forms 12 and 33, comparable to fragments from Mafjar in order for any one item to be understood. and Qasr al-Hayr.

20. DanielSchlumberger, Qasral-Heiral-Gharbi(Paris, 27. It is well ascertained that considerable booty came 1986). from the conquest of Central Asia and that taxes were at times paid in objects rather than cash. A 21. See M. Almagro and others, QusayrAmra (Madrid, complete gathering of all appropriate texts would 1984), which unfortunately does not give a com- be a very worthy enterprise. plete survey of the restored frescoes. It is also unfortunate that there is no account of the degree 28. The arguments justifying this position are much of restoration which has been provided in some beyond the range and purposes of this paper. instances, nor is anything really known about the From a methodological point of view I am trying to technical details of the paintings. distinguish artistic transfers which took place through objects from the ones made by traveling 22. Bibliographic references will be found in Creswell- artisans or artists. Allan, Early Muslim Architecture and in Ettinghau- sen and Grabar, IslamicArt. Some pieces have been 29. The restoration work was accomplished by a team shown in various traveling exhibitions organized sent by the Spanish government. The results of its by the Syrian Department of Antiquities, especially work are rather impressive, as can be seen in the 102 OLEG GRABAR

book by M. Almagro and others quoted earlier. To the right of the tent, in the upper part of the There are, however, several placeswhere the paint- panel, there is a Greek inscription reading quite ings visible now differ from the fragments seen by clearly NIKE, "Victory." In an article published a Musil a century ago and photographed by the few years ago, I connected NIKE known for a Reverend Fathers Jaussen and Savignac, Mission hundred years with a number of Greek letters archiologiqueen Arabie, vol. 3 (Paris, 1922). While discovered by the Spanish restorers to have been important for the overall interpretation of the painted on the left of the tent at the same level as bath's paintings, these queries are not particularly NIKE. In "Une inscription grecque a QusayrAmra," pertinent to the point of this essay. Revue des itudes islamiques (MilangesDominique Sour- del) 54 (1986), I argued that there was a single 30. Oleg Grabar, "La place de QusayrAmrah dans l'art name and proposed thatitbe readARIS [TO] NIKE. profane," Cahiers archiologiques36 (1988). Unfortunately, I do not believe any longer that this reading is possible for a whole series of reasons. A 31. Oleg Grabar, 'The Painting of the Six Kings at suggestive alternative will be proposed by Dr. Garth Qusayr Amrah," Ars Orientalis1 (1954). Fowden in a forthcoming book which, if accept- able, will be of extraordinary importance for the 32. This is an appropriate occasion to recall the help understanding of Qusayr 'Amra. I am most grate- of Fred Anderegg, for some time Head of Photo- ful to Dr. Fowden for having shared his interpreta- graphic Services at the University of Michigan, tion with me and for many wonderful discussions who came with me on so many Syrian andJordani- of Qusayr Amra. See Garth Fowden, Empire to an trips and who took and often developed his Commonwealth (Princeton, 1993), 143-49. pictures in most unusual places.

33. The representation on this wall is of a well-dressed woman in a tent with various figures around her.