In his History of Mehmed the Conqueror, Kritoboulos of Imbros (d. after 1467) sings the praises of the mosque of Ottoman grand vizier Mahmud Pasha Angelović (d. 1474), noting that it “was built with dressed stone and gleaming marbles, and with columns of outstanding beauty and size.”1 These words fit within the common practice of describing the luxurious materials and great financial expense that go into such buildings, a trope that readers encounter in a range of texts that mention architecture, beginning with medieval Islamic travel literature. Notions of wonder and astonishment are often expressed in such texts as well, in order to refer to the splendor of monuments whose means and materials of construction may not be apparent at first sight. This interest in process – or, rather, the fascination caused by the absence of knowledge about technical details – frequently appears as a cause of wonder (Arabic: ʿajab), not just in Ottoman contexts but also more broadly in medieval and early modern Islamic sources. In poetry, artifice is considered a crucial device intended to evoke wonder and aesthetic pleasure in the recipient, a notion that medieval Arabic works on poetics examine in great detail.2 In the fifteenth-century Ottoman context, where such poetry was well known, artifice as a device to evoke wonder appeared as well – including, I argue, in the creation of architecture. The Ottomans’ close engagement with transregional intellectual and artistic networks has so far been examined overwhelmingly in the context of Istanbul and treated as a phenomenon rooted in Mehmed II’s cosmopolitan patronage, but it had already begun in the first half of the century. Therefore, I begin this book with a chapter that reevaluates well-studied monuments in Istanbul, highlighting themes and connections that have not been previously explored. These themes and connections carry profound relevance for architecture in the first half of the fifteenth century, which I explore in subsequent chapters. I also turn to a late fifteenth-century structure in Skopje in order to explore how innovations were applied or created beyond the capital.
In this chapter, I investigate how the material politics of Ottoman architecture operated in Istanbul in the decades immediately following the Ottoman conquest of the city in 1453. Conscious choices about the materials to be used and the types of structures to be built were enacted, drawing on both earlier and contemporary ways of building within and beyond the Ottoman Empire, in a cultural universe that stretched from Venice to Central Asia. The monuments that Mehmed II (r. 1444–46 and 1451–81) and notables connected to his court built in this city offer a range of approaches to questions surrounding sensory perception, the meanings of different building materials, and the various layers of style and construction techniques.3 These questions are combined with an investigation of the larger issue pursued throughout this book – namely, how Ottoman architects experimented over the course of the fifteenth century with a wide range of forms and references to past and present coming from cities near and far. If a clearly recognizable Ottoman dynastic architecture eventually emerged, as I argue in Chapter 5, beginning in the reign of Bayezid II (r. 1481–1512), this development was not predestined at the time the Mahmud Pasha mosque and its contemporaries were built. Instead, architecture was in flux, as malleable as the empire whose elites sponsored it.
The Ottoman conquest of Constantinople in 1453 and the resulting end of the Byzantine Empire marked a dramatic shift in the political and cultural setting of the eastern Mediterranean, with far-reaching effects on Europe and the eastern Islamic world from Anatolia to Central Asia. Yet this event was not a singular occurrence but rather part of a time of upheaval within the Islamic world as the Timurid dominance of the first half of the fifteenth century began to unravel. The transition from Qaraqoyunlu to Aqqoyunlu rule in Iran and parts of eastern Anatolia in the 1460s, the Ottoman conquest of most of Anatolia, and the conflict between Ottomans and Mamluks are only a few of the political developments that took place within a fragmented world where the post-Mongol political landscape of the late fourteenth and early fifteenth centuries slowly gave way. For the Ottomans, the conquest of Constantinople came in tandem with a larger project of establishing control over Anatolia – something that Mehmed I and Murad II had not fully achieved in the first half of the fifteenth century – which was reflected in major conquests such as that of Trebizond in 1461 and Karaman in 1468. These developments would peak during the reign of Bayezid II, when the Ottomans expanded into southeastern Anatolia, the rising Safavid enterprise was increasingly present, and a decisive campaign against the Mamluks began to brew, which would be completed by Selim I (r. 1512–20) in 1516–17. By the 1520s, a new age of early modern empires was in the making. And yet the architectures that those empires would build to reflect the identities they were consolidating are rooted in the cultural and artistic developments of the fifteenth century.
In the current chapter, I examine how after the Ottoman conquest of Constantinople, and in the course of the establishment of the imperial Ottoman capital, architecture was reshaped within parameters that balanced the Byzantine heritage of the city with the aesthetics of the greater post-Mongol Islamic world. While there was a larger imperial project at stake, as Çiğdem Kafescioğlu has examined in detail, there was also a continued engagement with Saljuq, Timurid, and Aqqoyunlu architecture that had begun earlier in the fifteenth century (more on this in Chapter 2). Focusing on the period after the Ottoman conquest of Constantinople allows me to highlight some of the major questions at stake throughout the book, such as the use of wide-ranging stylistic references, eventual shifts in patronage and workshop practices that would come around 1500, and the question of when Ottoman architecture became integrated into the process of the empire’s political and cultural centralization. These questions are tied to the concept of material politics, in which the actions of makers and patrons intersect in the process of forming a distinct architectural style for the Ottoman Empire.
Timurid Style in Ottoman Lands?
As we investigate Ottoman architecture in Istanbul, it is helpful to address the notion of an international Timurid style, a concept that has dominated much of the scholarly narrative about fifteenth-century Ottoman culture. Overarching questions that arise immediately are: How was an international Timurid mode adopted in the Ottoman context? Which motifs were adopted in art, and what was their relationship to the overall convergence between Ottoman and Timurid scholarly and courtly cultures in the second half of the fifteenth century?
Discussing a poem by Tacizāde Cafer Çelebi (d. 1515) that was composed in 1493–94 in praise of the mosque of Mehmed II in Istanbul, Gülru Necipoğlu notes that the text refers to the interplay of “‘rūmī’ (i.e., foliate arabesque of Rūm, known as islīmī or islāmī in the Timurid world) and ḫiṭāyī (chinoiserie motifs of Cathay) patterns.”4 Rūmī motifs are composed of vegetal scrolls, often stylized into spirals, with pointed palmette leaves attached to them (see Figures 62 and 63), while the khiṭāyī repertoire consists of flowers such as peonies and chrysanthemums (see Figures 143 and 144), both prevalent in Chinese art, along with cloud bands. Bernard O’Kane has examined these terms in Timurid sources, and noted that the motifs they refer to appear already in the art of Ilkhanid Iran in the fourteenth century.5 Thus the scroll-and-leaf patterns that had frequently appeared in the Islamic architecture of medieval Anatolia under Saljuq and Ilkhanid rule since the twelfth century, both in stone carving and as part of tile mosaic, were integrated into a larger repertoire of Islamic ornament through the channels of the Mongol Empire that from the 1240s to the 1330s spanned from the Lands of Rūm to China. These types of ornaments were used in a wide range of materials, from tiles to stone to wood to metalwork, and would become firmly established within the repertoire of Islamic art – thus rūmī was not limited to Anatolia, and khiṭāyī spread across the Mongol Empire in the thirteenth century, persisting after this empire’s breakdown in the fourteenth.
Overall, the combination of these types of motifs – the rūmī emerging from Saljuq Anatolia, and the khiṭāyī derived from Chinese art and transferred into Islamic art by way of Mongol patronage in Iran – is what came to be regarded as typical of the international Timurid style beginning in the late fourteenth century, with local variations stretching from Samarqand to Edirne. From that period onward, the arts of the Timurid courts of Samarqand and, later, Herat came to be regarded as the cutting edge of Islamic art in the eastern lands of Islam, from Anatolia to Central Asia. In part, the prestige associated with the art developed at these courts had to do with the world domination that Timur (r. ca. 1370–1405) had aspired to with the extensive campaigns he undertook within the Islamic world. (As we see in Chapter 2, the Ottomans were at the receiving end of one of these campaigns, resulting in upheaval in the first decade of the fifteenth century.) A further source of prestige was, however, the accomplishment of artists who were assembled in Timurid court ateliers. The members of the Timurid kitābkhānas (court workshops whose name means “house of the book” but that worked with a much wider range of materials than the name suggests) reached high levels of skill and had access to all necessary materials and tools for their work, sponsored by Timurid patrons who included rulers, scholar-administrators, and poets.6 Many of these artists were prisoners taken during the campaigns of Timur and his successors, while others were workers who joined this artistic enterprise of their own free will. The high level of artistic skill in combination with extensive artistic patronage permitted the production of a wide range of sophisticated objects, including ceramics and metalwork, as well as works on paper. From artistic centers such as Samarqand, Herat, Tabriz, and Shiraz, these books were distributed widely and became important vehicles in the transfer of Timurid motifs across the eastern Islamic world. As we see in Chapter 2, contacts between the Timurid cultural sphere and the Ottomans developed in the early fifteenth century and had an impact on the material politics of Ottoman architecture at that time.
Within the Ottoman context, the taste for architectural decoration inspired by Timurid models persisted into the late fifteenth century, as reflected in the Çinili Köşk, built on the premises of Topkapı Palace in Istanbul in 1472 (Figure 1). As I discuss in what follows, craftsmen from Khurasan, possibly invited to the Ottoman capital to decorate this pavilion, were part of Mehmed II’s project to advertise his goal of universal rule in the architecture of his palace and capital. Some fifty years later, a group of tile makers who were taken to Istanbul when Sultan Selim I (r. 1512–20) defeated the Safavids at the battle of Çaldıran in 1514 renewed Timurid stylistic impact in the Ottoman context.7 This group “perpetuated a post-Timurid repertoire established earlier in the fifteenth century” in a time when Ottoman vocabularies were crystallizing – essentially, international Timurid style after Timurid rule had ended.8

1. Lateral view of east and south façades, Çinili Köşk, 1472, Istanbul, Turkey.
With the emergence of so-called saz-leaf decoration in the 1540s (Figure 2), named after the dominant element of the design, an elongated leaf with serrated edges (though this appeared in combination with other vegetal motifs such as pomegranates, tulips, and carnations), the new preference for a specifically Ottoman style that turned away from the Timurid and post-Timurid motifs of the fifteenth century would become obvious. On tiles, these new motifs – initially developed on paper – were more easily executed in underglaze technique. Underglaze permitted more freedom in drawing than the black-line technique (explained in detail in Chapter 2) generally associated with Timurid and post-Timurid style, which required a design to be created with clear boundaries between colors. This lack of technical flexibility in black-line tile was most likely the reason for the technique’s eventual disappearance from the Ottoman repertoire as expertise grew in creating complex underglaze designs requiring stencils and multiple colors (each with its own challenges in composition, application, and firing).9
As Necipoğlu notes, this turn from a largely Timurid to a fully developed Ottoman aesthetic in the sixteenth century was the result of a slow process of transformation rather than an abrupt shift. Earlier, cross-cultural interaction between Ottoman and Timurid courts was partly responsible for the pervasive Ottoman engagement with the culture of Timurid and post-Timurid Central Asia.10 This is true for the movements of both objects and people, the latter including makers as well as scholars who moved between courts, establishing an international community of knowledge that connected the Ottoman, Mamluk, and Timurid realms (see Chapter 3).11 This epistemological community included tile makers, painters, calligraphers, and architects who could move either along with their products or independently of them – or who could participate in this exchange without moving at all, simply by sending finished products and/or drawings to the place of destination. Although much of the architectural patronage in the fifteenth century is tied directly to the Ottoman sultan or to the inner circle of the Ottoman court, a clearly identifiable Ottoman style had not yet emerged, and building projects were sites of intense, focused experimentation.
Moving toward Constantinople
The conquest of Constantinople and its subsequent transformation into the new Ottoman capital of Istanbul have been studied in detail, and some key elements of this process bear repeating here.12 The demise of the Byzantine Empire in 1453 and the Ottoman takeover of its capital led to ideological changes for the Ottomans, who became more interested than ever in ideas of universal rule now that the long-standing goal of conquering Constantinople had been achieved. The Ottomans came into closer contact with the Adriatic world, already on their horizon through their Balkan possessions and through trade between Italy and the coastal regions of western Anatolia. With the presence of a Genoese colony in Pera, across the Golden Horn from the new Ottoman center, Renaissance Italy became an integral part of the Ottomans’ cultural orbit, although Bayezid I had already fostered connections to Europe.13 Venice was a major trade partner as well. As Julian Raby has shown, Mehmed II was an avid collector of art, interested in translations from Latin and Greek, and well read in classical literature. In the famous painting of the sultan by Gentile Bellini and in several bronze medals produced in Italy that depict him both in a detailed profile portrait and on horseback, Mehmed II is portrayed as an enlightened Renaissance patron and his empire is inserted into the world of early modern Europe.14 In a study of patronage at Mehmed II’s court, Gülru Necipoğlu argues that there was a conscious, complex set of negotiations in place that led to a “visual cosmopolitanism” attuned to both Renaissance Italian and Islamic art forms.15
This attitude of interest in other regions was not new to Mehmed II at the time of the conquest of Constantinople. When Mehmed II was again a prince following his brief, unsuccessful first reign as sultan in 1444–46, his marriage to the daughter of Sulayman b. Dulkadir, from the Dulkadir tribal confederation, had become a way to reach out to those powerful eastern Anatolian frontier lords. The Dulkadir tended to conclude alliances with the Mamluks, but Sulayman b. Dulkadir had been supporting the Ottomans in their protracted fight against the Karamanids, and the marriage consolidated these ties. As historians Neşri and ʿAşıkpaşazāde tell it, the unnamed wife of Hızır Ağa, governor of the province of Amasya, traveled from Amasya to Elbistan in order to select the bride, and again to fetch the young woman once Prince Mehmed’s father, Sultan Murad II, had approved the choice.16 The chosen princess, Sitti Shah Sultan, was taken to Edirne. The wedding took place in 1449 or 1450, before Murad II’s death in February 1451 and the prince’s second accession to the throne as Mehmed II.17 A double frontispiece in a Byzantine manuscript held in the Biblioteca Marciana in Venice has long been identified as a double-portrait of the bride, seated in a litter on top of an elephant, and of her brother, Melik Arslan; the manuscript was thought to have been a wedding gift to the future sultan.18 Recently, however, Merih Danalı has reexamined the manuscript and concluded that, first, it was never owned by the Ottoman court and, second, that the frontispiece is contemporaneous to the fourteenth-century manuscript produced in Thessaloniki. Thus Danalı lays to rest a long-standing art historical myth surrounding this double-portrait, which she identified as a representation of the philosopher and mathematician Hypatia of Alexandria (d. 415) and the geographer Ptolemy.19
Mehmed II also had close ties to his aristocratic Serbian stepmother, Mara Branković, who, along with Mahmud Pasha Angelović, was instrumental to connections to the Balkans.20 Rum Mehmed Pasha (d. 1474?), another one of the notables of Mehmed II’s court and grand vizier for a time after 1471, was married to a sister of Qilij Arslan, the Karamanid ruler of Alanya deposed by the Ottomans in 1471.21 In the decades before, even Ottoman princesses had been married to Karamanids: a daughter of Murad I was married to ʿAla al-Din (r. 1357–98?) and commissioned a madrasa in Karaman in 1381–82.22 In 1435, when Murad II campaigned against Taj al-Din Ibrahim (r. 1433–64), Murad’s own sister was the Karamanid’s wife, and she interceded on behalf of her husband, convincing her brother to cancel the deportation of the population of the city of Karaman (historical Larende) and to make peace.23
This interconnected courtly world thus abounded with family relationships between various dynasties in the Balkans and Anatolia; lengthy conflicts, movements of goods and artists, and complex modes of visual representation emerged. Necipoğlu points out that around 1400, Bayezid I was striving to include both Eastern and Western elements in his patronage, although these efforts were cut short by Timur’s invasion of Anatolia in 1402 and the subsequent events.24 As Raby notes, Mehmed II, who sponsored translation between Greek, Arabic, Persian, and Ottoman Turkish, showing interest in a wide range of works from Byzantine and Islamic cultures, also continued to engage in this kind of flexible and cosmopolitan patronage within the framework of his universal and imperial aspirations.25
The reshaping of Istanbul as an Ottoman capital was a central project in the fashioning of Mehmed II’s royal identity. Nevertheless, as I argue, the focus on imperial projects in Istanbul provides only one, if central, side of the picture. Various conflicts notwithstanding, there was close cultural interaction between the Ottoman, Aqqoyunlu, Mamluk, and, beginning in the early sixteenth century, Safavid empires. Mehmed II’s conquest of large parts of Anatolia and connections through scholarly and artistic networks to the Aqqoyunlu and Timurid realms were central to the political history of the time, together with the question of the Ottomans’ engagement with the Byzantine imperial heritage. Influence from the eastern parts of the expanding empire reached Istanbul, where on several monuments there appears tile decoration that can be tied in various ways to the so-called international Timurid style. More than a simple adoption of style as a sign of triumph, these monuments may be seen as part of an ongoing conversation with an Eastern, post-Mongol world.
The varied and diverse character of Ottoman architecture had deep roots in earlier building campaigns, particularly in Bursa but also in Edirne, as we see in Chapters 2, 3, and 4, and so this transformation cannot solely be attributed to the material politics of Mehmed II’s reign. These roots should be kept in mind as I move on to examine the architecture built in Istanbul. In this chapter, first I discuss the so-called Çinili Köşk, built in 1472 within the gardens of Topkapı Palace, and its combination of Persianate and Byzantine modes of construction. Then I analyze the mausoleum of Mahmud Pasha, built sometime before his execution in 1474, and investigate where it stands aesthetically between Persianate and Anatolian modes of construction. These two case studies will lead to a new understanding not only of these specific buildings but also of the aesthetic principles at hand as architecture engaged with Byzantine, Ottoman, and pre-Ottoman Islamic pasts, as well as Iranian and Central Asian presents, ideas that stemmed from the early fifteenth century.
The Çinili Köşk: Between Constantinople and Khurasan
The Çinili Köşk (Tiled Pavilion), built in 1472 on the premises of Topkapı Palace, is a central example of the negotiation between various architectural modes, some current at the time and others drawn from the past of newly conquered lands. The Çinili Köşk (Figure 3; see Figure 1) today stands separated from Topkapı Palace by the massive complex of the Istanbul Archaeological Museum, built in the late nineteenth century. The original garden setting no longer exists.

3. Frontal view from the steps of the Istanbul Archaeological Museum, Çinili Köşk, 1472, Istanbul, Turkey.
The fact that the building was originally part of the palace gardens, though, is important to consider, as such man-made landscapes were crucial parts of aesthetic experience within Islamic architecture. As D. Fairchild Ruggles notes in her discussion of gardens in Islamic al-Andalus, the sounds of water and wind, the singing of birds, and the smells of vegetation and flowers would have thoroughly transformed perception of a given site.26 Such was the case as well for the Çinili Köşk, known in Ottoman sources as sırça sarāyı: it was located in the palace’s outer gardens, juxtaposed with two other pavilions built in Ottoman and Byzantine styles, as Tursun Beğ notes, in order to present within close proximity the stylistic variations available in the Ottoman realm.27 In Tursun Beğ’s text, the Çinili Köşk is specifically referred to as built “in the mode of the Persian kings.”28 In a study of water features within Topkapı Palace, Necipoğlu notes the extent to which these outer gardens were wooded and studded with water features, and argues based on several sources that the larger garden landscape around the Çinili Köşk contained wide-ranging references to the extent of Mehmed II’s realm.29 Further, water features included in this landscape also evoked paradisiacal motifs. In a late sixteenth-century painting in the Hünernāme of Seyyid Loqman, the Çinili Köşk and its neighbor, the “Ottoman Pavilion,” are shown standing in front of a reflecting pool.30 As Necipoğlu notes, the pool and “paradise garden” were praised in a fifteenth-century eulogistic poem.31 Such elements are also evoked in the building’s inscriptions, as we see later in this chapter, and the notion of reflections in water is relevant with regard to the extensive use of marble and tile on the structure, particularly given the strong water symbolism that marble carries in Roman and Byzantine contexts. The juxtaposition of this reflecting pool in the original garden setting with the marble facing used on the building would have created a doubling of meaning: actual water to reflect the symbolic water of the marble.32 (I discuss this question further in Chapter 4.)
As Necipoğlu has argued, the Çinili Köşk may have specifically been commissioned in commemoration of the Ottomans’ triumph over the Karamanids in 1468.33 It features elements that evoked the architecture of the central Anatolian city of Karaman in the eyes of Giovanni Maria Angiolello (1451–ca. 1525) who spent time there in 1471–74, while he was living as an enslaved person at the Ottoman court.34 Scholarship on the monument so far has focused on its connections to Timurid, Karamanid, and Aqqoyunlu architecture under the larger umbrella of the international Timurid style, which in the Ottoman context is thought to have been first adopted in the early fifteenth century under the patronage of Mehmed I in Bursa. (This context is discussed in Chapter 2.) I argue, however, that the Çinili Köşk’s aesthetics cannot be entirely identified as either Persianate or Karamanid.
The Ottoman conquest of Karaman in 1468 to which the building is connected was the culmination of an extended conflict between the two dynasties that had begun in the early days of Ottoman emergence in the late thirteenth and early fourteenth centuries. Starting in the late thirteenth century, the Karamanids had made the city of Karaman their center; rivalry between them and the Ottomans began early on and was often entangled with the complex relationship between Ottomans and Mamluks. The latter had first tried to expand into Anatolia as part of their rivalry with the Ilkhanids in the second half of the thirteenth century, marked by Mamluk sultan Baybars I’s incursion into southeastern and central Anatolia in 1277, an event that would take center stage in later Ottoman history writing because it boosted Karamanid power.35 As noted before, in the late fourteenth century the Ottomans and Karamanids attempted to appease relations using dynastic marriage between a daughter of the Ottoman sultan Murad I and ʿAla al-Din I, the ruler of Karaman.36 This is one of the earliest documented Ottoman dynastic marriages, a practice that would end in the early sixteenth century.37 The attempt at reconciliation was hardly successful in the long term, and the Ottomans’ struggle with the Karamanids would continue until the mid-fifteenth century.
After Timur defeated Bayezid I in 1402 and restored many of the beyliks that Bayezid had subdued, the Karamanids too reemerged as powerful lords in central Anatolia.38 The Ottoman attack on the Karamanid realm in 1468 was triggered by the revolt of Pir Ahmed, who had ruled parts of the Karamanid lands under Ottoman suzerainty since 1466, the result of a war between brothers after his father, Taj al-Din Ibrahim, had died at the end of a long reign.39 Yet the revolt did not come out of the blue; Sara Nur Yıldız argues that it was triggered by the Ottomans’ increasing encroachment on the Karamanid hinterlands in previous years, and that after being touched off in this way, the rebellion may have served Mehmed II as an excuse to finally lead his armies toward Konya and Karaman.40 In the city of Karaman, with the exception of military construction, the conquest was not marked with Ottoman construction projects and rebuilding – as happened in Constantinople – but rather with the destruction of Karamanid palaces. As the historian Şikari reports, these buildings were razed, a new citadel was built in their place, and inhabitants were forced to move to Istanbul to repopulate that city for the Ottomans’ use.41
Between Persianate and Byzantine Architecture
The entrance façade of the Çinili Köşk features a tall marble porch that extends over two levels (see Figure 3). On the lower level, a central stone panel with a grilled window in the middle hides the stairs leading up to the entrance, which is located in a recess on the upper level. To the left and right of this central panel, five arches with the flattened profile of what is sometimes referred to as a Bursa arch (based on its frequent use in the early fifteenth-century architecture of that city) are supported by square pillars. On the upper level, a larger central arch is flanked by six narrower ones on each side. Fourteen octagonal columns with geometric capitals and square impost blocks support the roofline above.
One risk of analyzing the Çinili Köşk based on the presence of this monumental porch is the fact that it is not original to the building; in fact, a fire in 1737 destroyed what was likely a wooden porch and the current stone porch replaced it, although the exact date of the addition is unknown.42 Similar wooden porches can still be seen in Iran and Central Asia – for instance, on the citadel of Bukhara – although extant examples date to the seventeenth and eighteenth centuries (Figure 4). In merely technical terms, a possible comparison also lies in the wood-column and wood-ceiling mosques of late thirteenth- and early fourteenth-century Anatolia, such as the Aslanhane Mosque in Ankara or the Eşrefoğlu Mosque in Beyşehir, although there is no continuity between these monuments and the Çinili Köşk.43 Earlier pavilions from the Saljuq period survive only in fragments: the kiosk of Qilij Arslan II (r. 1156–92) in Konya, connected to the citadel walls, which collapsed in the early twentieth century, was built of brick with wooden beams and decorated with stucco and mināʾī tiles.44 Larger Saljuq garden landscapes within central and eastern Anatolia extended to hunting preserves such as those in the region of Alanya in which small structures, generally built of stone, have survived as traces in the archaeological record.45 As such, traces of gardens and pavilions predating the Çinili Köşk are rare.

4. Reception hall, 1605–06, citadel, Bukhara, Uzbekistan.
The architecture of the Çinili Köşk as such is that of a garden pavilion of the type that was frequent across the eastern Islamic world, although few have been preserved that predate the seventeenth century; two central examples are the Chehel Sotun (1647; current porch 1708) and Hesht Behesht (1669–70) in Safavid Isfahan.46 Overall, as Necipoğlu has noted, comparisons to garden pavilions in central Anatolia, Iran, and Central Asia are hampered by the fact that no fifteenth-century examples from those places have survived; still, her detailed analysis of the façade elevations and plan of the Çinili Köşk clearly supports such a comparison.47 Once the visitor arrives on the upper level of the Çinili Köşk, in front of the entrance, however, the Persianate view is immediately, if not entirely, upset. It is true that large parts of the façade are decorated with tile mosaic and a long foundation inscription runs across the deep recess around the entrance (Figure 5). Yet the lower zone of the upper entrance level is composed of a dado of white-and-grey book-matched marble, and the door frame – probably spolia from a Byzantine building – is carved from two slightly different shades of white marble with purple veins (Figures 6 and 7). These reused architectural elements add another layer of visual cues and historical references.

5. Upper section of entrance īwān with tile mosaic decoration including foundation inscription, Çinili Köşk, 1472, Istanbul, Turkey.

6. Partial view of the porch interior to the right of doorway, Çinili Köşk, 1472, Istanbul, Turkey.

7. Marble revetment to the right of doorway, Çinili Köşk, 1472, Istanbul, Turkey.
Looking at the marble facing of the dado around the entrance, we can see that the slabs are cut and arranged so that the veins in the stone match. Examples of such book-matched marble cladding are prominent in the Byzantine monuments of Istanbul, including the nearby sixth-century church of Hagia Sophia, which was converted into a mosque immediately following Mehmed II’s conquest of the city. The use of spolia for the door frame also introduces the past of the new Ottoman capital into the Çinili Köşk, breaking up an aesthetic that at first sight appears entirely Persianate. In fact, when the monument is seen from afar, this detail of the decoration remains obscured by the porch and does not emerge into view. It is only as one draws closer and finally enters the building that this other element, related to Istanbul’s Byzantine (hence Roman imperial) past and Mehmed II’s chosen role as qayṣer-i rūm, appears.48 Thus the visual cosmopolitanism that Necipoğlu and Raby attribute to Mehmed II and his court is integrated into this building, which is constructed – figuratively and literally – as a microcosm of the meanings built into the garden’s larger landscape. This observation points to the need to reassess the dynamics of past and present artistic forms and of engagements with Ottoman and non-Ottoman lieux de mémoire in this building.49 Yet with regard to the tile mosaic (Figure 8) on the façade, the question remains whether it is a reference to Iran or Central Asia, fifteenth-century Karaman or thirteenth-century Konya, or a combination of all of these. This question is examined after a discussion of the cosmological concepts and notions of artifice involved in the Çinili Köşk’s inscriptions, made of tile mosaic.

8. Detail of tile mosaic on entrance arch of Çinili Köşk, 1472, Istanbul, Turkey.
Epigraphic Artifice and Poetic Cosmology
The inscription, composed in Persian, above the entrance to the Çinili Köşk spans all three sides of the portal recess. It is a particularly fine example of tile mosaic, and nothing similar has been preserved in Istanbul (or anywhere else in the fifteenth-century Ottoman world). The calligraphy is complex in that two texts overlap: the main inscription (Figures 9 and 10), which contains the date of construction, is the larger text set in white that runs in two lines. Another inscription in verse, further praising the beauty of the monument, is set in smaller script overlapping the top of the main inscription, with some words in yellow and others in white. Both inscriptions contain textual references that are crucial to an understanding of the cosmological aspects of the monument; more than this, they also provide insights into ways in which materiality operated in the fifteenth-century Ottoman context. Artifice was involved at multiple levels: in making the tiles needed for the tile mosaic, in putting together the tile mosaic itself, and in the poetic work of composing the text itself before it was inscribed on the building. Translated here into English, the text of the larger inscription reads:
This pavilion, which is as lofty as the heavens, was so built that its great height would stretch up to Gemini. Its most worthless part would adorn the stars of Ursa Minor (firq ferqdān) and the highest sky of Saturn (saqf-i keyvān). Its emerald dome (qubba-i zumurrudīn) sparkles like the golden sky (āsumān-i zerrīn) and is adorned with inscription of stars (kitābe-yi kewākeb). Its courtyard floor of turquoise with multicolored flowers and designs (nuqūsh) of būqalamūn is a place of nature’s beauty.50 May it be everlasting – it was made to display the glory of the khāqān’s state and the good fortune of divine providence’s assistance. By God’s order – the building reflects the patron’s importance51 – it reached the honor of completion in the last days of the month of Rabīʿ ‘l-Ākhar in the time of the year eight hundred and seventy-seven.52
By mentioning the dome and the courtyard of the building, the larger inscription invites the reader to enter and admire the interior as well, with its splendid floor (which is either no longer extant or the subject of poetic hyperbole) and vault. In this inscription, as Necipoğlu has noted, the building is exalted with allusions to cosmic motifs: the emerald dome (qubbat-i zumurrudīn) and the heavenly mansion (qaṣr-i falak).53 These motifs correspond to ones used in panegyric and ekphrastic poetry composed by Ottoman authors; such poems describe the Çinili Köşk using comparisons to the firmament and the sun and moon, and include allusions to activities such as feasts and receptions taking place there.54 The motifs used certainly corresponded to ideas widely available among the poets and scholars of the Ottoman courts, and ones that were in line with the image of universal rule that was projected for Mehmed II. Open questions are who composed the inscription, who approved it for use on this royal building, and who read it. Would Mehmed II have been involved in the selection of these texts? Most likely so, considering that he was a keen patron, interested in art and architecture as much as literature.

9. Detail of portal inscription, Çinili Köşk, 1472, Istanbul, Turkey.

10. Detail of portal inscription, Çinili Köşk, 1472, Istanbul, Turkey.
The cosmological framework proposed in the inscription can be tied into notions of kingship such as the title ṣāḥib-qirān (lord of the auspicious conjunction, adapted from the Timurids), as well as the Ottomans’ interest in ʿilm al-nujūm (astrology and astronomy combined), documented since the early fifteenth century.55 The image of the emerald dome (qubba-i zumurrudīn) is not used to describe the building’s actual dome but rather to metaphorically evoke al-qubbat al-khaḍrāʾ (lit. “the green dome”) – the term denoting the heavens in the Islamic tradition, as Jonathan Bloom notes.56 By the fifteenth century, the concept had become standard poetic imagery in both Arabic and Persian.57
In addition, the text contains references that are meaningful in the context of experiencing spaces in a multisensory way, addressing not only vision but also sound and touch. Specifically, the reference to designs of būqalamūn (nuqūsh-i būqalamūn) is relevant in this regard, since this polyvalent term can refer to a range of materials (both natural and man-made) that specifically appeal to notions of wonder (ʿajab) and are often marked by shimmering color.58 As Matthew Saba has shown in his insightful study of ninth-century Abbasid luster ceramics from the caliphal palace city of Samarra in Iraq, the term būqalamūn could refer to a range of objects with changing color effects, from peacock feathers to textiles.59 In the inscription of the Çinili Köşk, the term is specifically used to describe the “courtyard floor”; currently, the floor inside the building is paved evenly with large rectangles of grey marble, a modern addition. What did the original floor look like? Was it a floor like the one preserved in the seventeenth-century Shaykh Lotfollah Mosque in Isfahan, made of rectangular turquoise tiles? Could the inscription refer to floor coverings? Carpets would certainly have been present in a palatial building like the Çinili Köşk, and they might have had the flower designs and changing colors to which the inscription refers. In the inscription, the floor is specifically connected to nature’s beauty, an emphasis not on man-made aspects of the building but rather on allusions to natural things that elicit wonder and admiration. The cosmological motifs in the first part of the inscription tie into this kind of comparison as well: the beauty of the structure and its admirable features are compared to natural phenomena such as the firmament, the constellations, and the golden sky – a phrase that perhaps refers to dawn, which is invoked in the second inscription. Ultimately, by comparing the building’s beauty to natural phenomena, the Çinili Köşk is equated with divine creation and, by extension, its patron is approached to God. (Although as we shall see, the text of the smaller inscription includes a reference to the ephemerality of worldly matters and humans.)
The smaller inscription in verse, which is woven into the larger one at the top edge of the inscription band, contains two separate rubāʿiyyāt (a short, formulaic poem in Persian).60 It contains further comparisons between the building and natural beauty within a cosmological framework:
In this second inscription, the emphasis shifts slightly: a cosmological theme appears in the last verse, where the sun, dawn, and sky are evoked. The preceding verses praise the building’s luxury as a setting for the court, although a reminder of the transitory nature of worldly and imperial glory appears in the use of the term ramīm (decayed), referring to a Qur’anic passage.63 The verse obliquely evokes the famous lines that, according to Tursun Beğ, Mehmed II uttered when seeing the ruined nature of some of Constantinople’s monuments: “The spider is holding the curtains in the palace of Khusraw. The owl is holding watch on the citadel of Afrasiyab.”64 Thus the inscription also contains a small hint at the fact that the world passes and that even the (here unnamed) royal patron’s life is transitory. A related sentiment is expressed two centuries earlier in the Sırçalı Madrasa (1243) in Konya, where a medallion in the tile work, paired with another one containing the signature of Muhammad al-Tusi, praises the work and notes that the building will survive its maker, but that it too is ephemeral.65
As Paul Losensky notes, complex relationships can emerge between poetry and architecture in the sense that “verbal texts contribute an audible voice to mute buildings, while architectonic imagery lends material substantiality to the airy stuff of poetry.”66 Here the skillful material artifice of the two intertwined inscriptions in tile mosaic, which are easily legible despite the complex calligraphy translated into tile, mirrors the poetic artifice of the inscriptions.
The Tile Decoration of the Çinili Köşk
In addition to the double inscription above the entrance, the īwān frame around the entrance of the building is adorned with more tile mosaic: two interlocking floral scrolls that intersect with square kufic script alternately reading “allāh” and “akbar” (see Figure 8). These motifs – “part abstract, part mimetic”67 – connect the design to Timurid tile work of the late fourteenth century onward. Altogether, the façade is the most intricate example of tile mosaic that has survived in Ottoman architecture. The tile mosaic also differs from earlier examples that survive in central Anatolia, especially Konya and Sivas, in that it includes such Timurid-style motifs, as well as tiles in yellow, expanding the earlier color palette.
It is important to note, as David J. Roxburgh emphasizes, that full-size paper models were needed for complex designs in tile mosaic. These models would be used to cut the tiles, assemble them, and prepare finished panels that could be installed on a monument.68 In the case of the Çinili Köşk, such models, employed in assembly on the ground, would have been needed for the design just described on the frame around the īwān, and certainly for the two intertwined inscriptions above the doorway. The geometric patterns on the surfaces above and below the inscription, made out of square and rectangular tiles rather than cutouts, might have been assembled directly on the monument.
Certainly the Çinili Köşk’s façade evokes the kinds of architectural decoration created in Timurid Samarqand beginning in the 1380s, when tile mosaic in flat panels began to dominate, and the earlier use of carved tiles, along with features such as non-load-bearing columns made of tiles, began to give way to sleek surfaces produced in mosaic.69 The growth of the Timurid kitabkhāna, with its capacity to produce designs for use in a range of media, also fostered the use of tile mosaic in architecture and the creation of calligraphic models that could be either scaled up for architectural use or kept at a small scale in manuscripts and albums.70 In technical terms, tile mosaic had the advantage that accidents of glazing and firing did not affect the finished product.71 Financial loss was not so great if the monochrome pieces of tile from which the mosaic shapes would be cut were marred by misfiring. If this happened to a painted tile with several colors and a complex design, however, that was a different story, especially when colored glazes ran due to unsuitable firing temperatures. In that case, a decision needed to be made as to whether making a new tile should be attempted (again with the risk of misfiring) or whether a tile could be used anyway; a tile might still be usable, for instance, if the problem at hand was simply a color that had run because of inadequate firing temperature. In an Ottoman context, misfired underglaze-painted blue-and-white tiles can, for instance, be seen on the cenotaph of Sitti Hatun (ca. 1450), in the mausoleum of Mehmed I in Bursa; these misfired tiles might have been originally intended for the Muradiye Mosque (1435–36) in Edirne.72
The interior decoration of the Çinili Köşk has been only partially preserved, with blue and turquoise tiles, some with overlaid gold decoration (restored in places; Figures 11 and 12). Unlike the tile mosaic of the façade, which clearly belongs to an Iranian or Central Asian context, such monochrome tiles, both with and without gold overlay, can be found in earlier Ottoman buildings, including ones in Bursa (see Figure 48), as well as Saljuq monuments of the thirteenth century, such as the Karatay Madrasa in Konya (see Figure 69). To make matters more complex, representations of similar tiles also appear in Timurid paintings – for instance, a depiction of “Bahram Gur and the Indian Princess in the Dark Palace,” copied in Herat in the 1430s.73 This is just one of many examples in the fifteenth-century Ottoman context where references can be polyvalent: thus the tiles with gold overlay can be a reference to three periods: the Saljuq past of Anatolia, a period that is evoked in the early fifteenth century to overcome the trauma of Timur’s invasion; earlier Ottoman architecture in Bursa, the center of dynastic memory; and contemporary architectures in central Anatolia, Iran, and Central Asia. Simply based on stylistic analysis, it is impossible to decide which of these references prevailed in the mind of fifteenth-century Ottoman builders and patrons. Considering the complex material politics of the time, it is also possible that the multiple intersecting references were intended and consciously applied.

11. Interior view of room with blue and turquoise tiles with restored gold decoration, Çinili Köşk, 1472, Istanbul, Turkey.

12. Interior, detail of blue and turquoise tiles with unrestored gold decoration, Çinili Köşk, 1472, Istanbul, Turkey.
Although little has been preserved in Karaman itself, one example of Karamanid tile work, the mihrab of İbrahim Bey İmaret (1432), was removed from its original location and transferred to the Istanbul Archaeological Museum in the early twentieth century. While black-line tile covers most of its surface (this technique is discussed in the next chapter), it is noteworthy here that the muqarnas niche is filled with monochrome dark purple tiles. Such tiles, along with rare fourteenth- and fifteenth-century examples of tile mosaic in Konya, were the tail end of the kinds of tile work produced in central Anatolia under Saljuq and Ilkhanid rule, although the later programs are much less complex in their motifs and scale than the thirteenth-century ones.74 In its design and style, the tile mosaic of Çinili Köşk is much closer to late fourteenth- and fifteenth-century examples in Iran and Central Asia than to thirteenth-century ones within Anatolia, pointing to a case of transregional exchange. With these issues in mind, the question of why Mehmed II and his architects chose exterior tile work that is much closer in style to Timurid examples to represent Karaman looms large. Was this a matter of opportunity, of a group of tile workers familiar with the techniques and motifs used in eastern Iran already being present in Istanbul? Was it a conscious choice inspired by the beauty of this kind of work and perhaps by an impulse to compete with the Aqqoyunlu at a moment when the conflict with Uzun Hasan was not yet settled?
A further challenge in expanding the discussion of possible references to the contemporaneous context of Anatolia beyond the realms of Karaman is the state of architectural evidence. In Chapter 3, I discuss the relationship between the architecture of the western Anatolian beyliks from the late fourteenth century onward and Ottoman building projects. I argue that close connections are present between these structures and Ottoman patronage in Iznik and Bursa, but that these happen at the level of stonework, rather than tile. (And indeed tile was rarely used in the western Anatolian beyliks). For central and eastern Anatolia, evidence is more limited. Thus nothing remains of the patronage of the powerful Qadi Burhaneddin Ahmad, who ruled from his center of Sivas from 1380 until the Ottomans defeated him in 1398.75 The previous ruler of the region, Eretna (r. 1336–52), and his successors left behind buildings such as the Güdük Minare (1347–48), a monumental funerary structure built of brick with tile insets (a technique known as bannāʾī).76 This technique had been used in Sivas since the early thirteenth century – for instance, on the minaret of the Great Mosque (1212–13), in the Şifaiye Madrasa (1216–17), and in the Çifte Minareli Madrasa and the Gök Madrasa (both 1270–71).77 Another Eretnid monument, the so-called Köşk Madrasa (1339), originally located on the outskirts of Kayseri but now absorbed by the expanding city, is built of ashlar masonry like most thirteenth- and fourteenth-century buildings in the region.78 While the entrance to the complex has drop-shaped stones forming its doorway and a frieze above, the muqarnas base of the mausoleum at its center is a direct reference to the Mahperi Hatun Mausoleum (ca. 1237–38) in the center of Kayseri.79 Here too we are faced with the continuation of a local architectural idiom created in the twelfth and thirteenth centuries far into the post-Mongol period and up until the Ottoman conquest of the city in 1474. Tile work is not used in any of these buildings. For that, one has to look to the southeast, to the Zeynel Bey Türbe in Hasankeyf, built for a son of Uzun Hasan who fell at the battle of Otlukbeli between Ottomans and Aqqoyunlu in 1473. The building built of brick and tile survives today, but was moved to a new site in May 2017 to save it from the waters of the Ilısu Dam that have swallowed historic Hasankeyf.80 Bannāʾī tile work, rather than tile mosaic, adorns this building.
Overall, tile mosaic is highly unusual both within Ottoman architecture and in fifteenth-century Anatolia more broadly. In Karaman, limited evidence from the late fourteenth century survived in the Hatuniye Madrasa (1381–82) until at least the 1960s.81 The technique of tile mosaic was most widespread in the region of Konya in the mid-thirteenth century. Examples include the Sırçalı Madrasa, built in 1243 (Figure 13), and the Karatay Madrasa (Figure 14), built in 1251–52.82 Beyond Konya, the Gök Madrasa in Tokat, an undated building that can be placed in the 1270 to 1280s, had extensive tile mosaic decoration in its courtyard, now fragmentary but better preserved on the qibla īwān (Figure 15).83 In all of these examples, where breakage of tiles has occurred over the centuries, it is clear that the tiles were installed on top of the underlying walls (built of brick or stone, depending on the building) as premade mosaic panels, in the same manner as on the Çinili Köşk. Thus high levels of skill in producing tile mosaic had been achieved in Konya, but also in rarer examples in Tokat and Sivas. If the technique was no longer used on a large scale after the 1280s, this was due to the long-term effects on patronage of the Mongol conquest of Anatolia in 1243. Expensive and time-consuming techniques such as tile mosaic introduced at the height of Saljuq power in the 1220s and 1230s were still used in the first decades after the Mongol conquest, as notables who now filled in the role of intermediary between the largely powerless Saljuq sultans and the Mongol overlords became sufficiently wealthy to commission large complexes for themselves – including examples such as the Karatay Madrasa and the Sırçalı Madrasa, along with many more buildings constructed in central and eastern Anatolia between 1243 and circa 1280.84 With the closer administrative integration of Anatolia into the Ilkhanid empire after 1277, however, patronage became increasingly local and small-scale, while larger construction projects were more likely to be completed in the Ilkhanid centers in Iran, especially Tabriz and Sultaniye.85 In Ottoman architecture, large-scale tile programs first appear in Mehmed I’s mosque-zāviye complex in Bursa, begun in 1419, and in the work connected to the Masters of Tabriz, a group of tile makers thought to have come from Tabriz in the existing scholarly literature. The work on which these makers’ collective name appears – that is, the mihrab in Mehmed I’s mosque-zāviye – refers to then-current fashions in tile decoration in Iran and parts of Central Asia.86 Notably, though, tile mosaic was rarely used in the monuments connected to the Masters of Tabriz as Chapter 2 shows.

13. View of main īwān with tile mosaic, Sırçalı Madrasa, 1243, Konya, Turkey.

14. Interior view Karatay Madrasa, 1251–52, Konya, Turkey.

15. Detail of tile mosaic on the qibla īwān, Gök Madrasa, ca. 1270–90, Tokat, Turkey.
Ottoman architecture in the fifteenth century showed a conscious and sophisticated engagement with the past and present architectures of both its own realm and neighboring regions under Muslim rule. At the Çinili Köşk, the presence of complex references to a range of stylistic realms – Byzantine and Timurid first and foremost, but perhaps also Karamanid – is handled in two main ways. First, these stylistic references are embodied in the building’s plan, elevation, and exterior decoration, where they seamlessly combine into a new whole, expressing the skill of Ottoman building workers in creating artifice and wonder as backdrops for the court’s activities. Second, the poetry applied to the building’s façade deploys some of the same ideas of rulership and cosmology as those expressed in Timurid contexts, but also notions of natural beauty embodied in the artifice of architecture. Originally, the garden landscape would have enhanced these effects with elements such as plants and reflections in pools. Thus architecture and poetry merge together in the creation of an aesthetic experience and its presentation to the court.
We can observe this close relationship between building and text in other monuments: in later chapters, I show how such dynamics worked in the first half of the fifteenth century, as the Ottomans reconsolidated their realm and later relaunched the expansion into Anatolia that Bayezid I had started. At the present moment, what matters is the fact that those responsible for Ottoman buildings (architects, tile makers, stone carvers, and also poets and calligraphers) in the fifteenth century engaged with complex references to multiple cultures and time periods. It is also important to point out that this sort of engagement began before Mehmed II’s conquest of Constantinople in 1453, as Necipoğlu has argued.87 Thus the Çinili Köşk with its multiple references also connects to that Ottoman past, and its architects did not reinvent the wheel in the name of commemorating Mehmed II’s conquest of Karaman, although the renewed references to Byzantine architecture were an additional element peculiar to Istanbul’s shaping as the Ottoman capital. Within the palace context, as Kafescioğlu argues, the Çinili Köşk and its no-longer-extant companion pavilions were part of a world where forms could be mixed freely within that consciously cosmopolitan aesthetic framework.88
The Çinili Köşk is the only example of tile mosaic that was produced in Istanbul in this period. A petition in the Topkapı Palace Archive suggests that tile cutters from Khurasan (kāshī-tarāshān-e Khurasān, as they refer to themselves) worked in Istanbul in the mid-fifteenth century and, as they fell on hard times, begged the sultan for more work. While the Çinili Köşk is not mentioned in the petition, Faik Kırımlı and Necipoğlu have both argued that it was after completing this very project that the tile cutters found themselves out of work.89 Kırımlı even goes so far as to suggest that the same group also created the tile decoration of the Mahmud Pasha Mausoleum. However, we shall see that, from a technical standpoint, the tile decoration on that monument was made in a different and rarely used technique, suggesting that another workshop might have been responsible and that knowledge of the complexities of Timurid-style tile mosaic may not have been necessary. Expertise in creating complex geometrical patterns in stone, however, was a must.
Istanbul beyond the Çinili Köşk
The Çinili Köşk was part of the larger project of turning Constantinople into an Ottoman city. In her analysis of the addition of mosque complexes under the patronage of Mehmed II’s grand viziers, Kafescioğlu argued that these new construction projects operated as a way to structurally mark Istanbul with Ottoman presence and as a Muslim-ruled city.90 In the architecture of these and other monuments of the time, which show close negotiation between Byzantine, earlier Ottoman, Anatolian, and Timurid models, one sees the same juxtaposition of stylistic modes that we just observed in the Çinili Köşk, and it is characteristic of Mehmed II’s patronage. Thus the Rum Mehmed Pasha Mosque, built C. 1472 and towering on a hill on the Asian side of Istanbul, has the silhouette of a Byzantine church, with brick dominating its exterior (Figure 16). In the interior of the mosque, however, the surviving fragments of decoration (Figure 17) show that the dome was painted in scrolls and leaves that point to a use of Eastern motifs, the kinds of rūmī patterns associated with Saljuq architecture and its successors.91

16. Üsküdar shoreline with Rum Mehmed Pasha Mosque, ca. 1472, Istanbul, Turkey.

17. Detail of painted designs inside the dome, before restoration, Rum Mehmed Pasha Mosque, ca. 1472, Istanbul, Turkey.
In terms of fifteenth-century architectural decoration and the many stylistic references that are combined, it is worth considering Mehmed II’s mosque in Istanbul, built in 1463–70. There, two tile lunettes survived the collapse of the mosque in an earthquake in 1766, after which much of the building was rebuilt.92 These two panels are located over the windows in the bays of the courtyard portico where it joins the façade of the prayer hall (Figure 18). In an attempt to place these panels in context, Sandra Aube has suggested that the same workshop made both these tile panels and the ones produced for the Üç Şerefeli Mosque (1438–47) in Edirne, commissioned by Murad II, but this point implies the continuity of a single, unproven workshop tradition.93 In fact, notable differences in types of decoration between the two panels on Mehmed II’s mosque are revealed upon close observation, even though the overall effect when viewed from afar is similar.

18. View along arcade on south side of courtyard, toward location of underglaze tile panel on west side of courtyard, Mosque of Mehmed II, 1463–70, Istanbul, Turkey.
In the panel located on the east side of the courtyard (Figure 19), the frame is divided into fields by interlinked white bands. The larger fields are filled with peonies from the khiṭāyī repertoire combined with leaves akin to the saz leaf that would appear in later decoration, although perhaps not quite as pointed.94 In the central lunette, two overlapping inscriptions are placed on a rūmī scroll background; the larger, cursive text reads: “[and] the preservation of them both [Heaven and Earth] tires Him not, and He is the Most High, the Great” (“lā yāʾduhu ḥifẓuhumā wa huwa l-ʿaālī l ʿaẓīm”; a passage from the Throne Verse, Qur’an II:255).95 While both the cursive and kufic inscriptions are in white on a blue-and-turquoise background, the frame has some yellow accents in addition to these colors.

19. Underglaze painted tile panel, east side of courtyard, Mosque of Mehmed II, 1463–70, Istanbul, Turkey.
In the tile panel located on the west side of the courtyard (Figure 20), a white band runs along the entire length of the frame, twisting itself into regular shapes that alternate between complex geometric knots and a motif whose form seems to be, for lack of a better term, a mix between a fleur-de-lys and a merlon; alongside it runs a thin scroll that carries cloud bands from the khiṭāyī repertoire and saz leaves. The inscription panel has a large basmala in white kufic script, and in yellow kufic the phrase “There is no compulsion in religion” (lā ikrāh fī l-dīn; Qur’an II:256). As in the eastern panel, some yellow appears in the frame, but the color is even more prominent here because of its presence in the central field.

20. Underglaze painted tile panel, west side of courtyard, Mosque of Mehmed II, 1463–70, Istanbul, Turkey.
The central fields in both lunettes include the combination of khiṭāyī and rūmī forms noted by Tacizāde Cafer Çelebi, a madrasa-educated member of the Ottoman court, in the poem cited earlier in this chapter. In it, he describes windows in Mehmed II’s mosque in these lines:
Cafer Çelebi’s poem, composed in 1493, is a panegyric about Istanbul, an example of the genre of şehrengīz or masnavī-shahrāshūb, which took hold first in Persian and then in Ottoman-Turkish poetry, becoming widespread in both literatures by the sixteenth century.97 Titled Vaṣf-ı ḫıṭṭa-i İslâmbol – that is, “Description of the Region of İslâmbol,” here designating Istanbul with the Ottoman expression that referred to the city’s Islamization, as it literally means “full of Islam” – the poem begins with praise for Mehmed II as a conqueror. It then moves on to an engagement with the built environment of late fifteenth-century Ottoman Istanbul, including monuments such as the Hagia Sophia, the madrasas belonging to the mosque complex of Mehmed II, and Topkapı Palace.98 The poetic description of parts of the city includes a wide range of cosmological motifs, including allusions to the Pleiades (süreyyâ) and the mythical spring of paradise (Arabic kawthar, Ottoman Turkish kevser).99 Related to these are expressions of the astonishment and wonder the viewer experiences when contemplating the buildings (temâşâ etmek) from inside.100 The poet expresses bewilderment (ser-gerdânık) as he visits the palace and is left amazed (ḥayrân) by the sight of the designs and images (nakş u nigârı).101 In terms of materials mentioned, marble is central: it appears in descriptions of a hammam within the palace and of the mosque of Mehmed II.102 Textiles make scarcer appearances, only in a reference to tent screens (serâ-perde) and curtains (perde) in the palace.103 The Çinili Köşk is mentioned specifically, called sırça-serâ – another name with the same meaning, “tiled pavilion.”104 The building’s beauty is exalted in the text with an emphasis on the pool of water lying in front of it and on its monumentality – a comparison is made to the dam that Alexander the Great built against the menace of Gog and Magog.105
The poem’s description of the mosque of Mehmed II praises the building highly.106 In addition to the verse emphasizing khiṭāyī and rūmī designs cited earlier, the beauty of the designs (again, the phrase nakş u nigârı) of the woodwork (kündekâri) on the doors is described, along with a range of architectural elements and decorative features within the building.107 Materials are at times evoked – in the passage cited earlier, we see glass mentioned twice, once as colored glass in general and then specifically as window glass. The poem also describes the vast, marble-paved courtyard with its water features at the center, comparing the latter to kevser, the spring of paradise.108 The praise lavished on the building is in tune with the building’s importance in the Ottoman reshaping of Istanbul. The mosque of Mehmed II was part of a competition that Ottoman architects engaged in with the Hagia Sophia – its large central dome in particular. This competition would mark imperial architecture for centuries to come. The symmetrical arrangement of the mosque and subsidiary buildings, Necipoğlu argues, may be a nod to the principles of Renaissance architecture.109 From this point of view, the mosque of Mehmed II presents a synthesis between the various aesthetic modes that late fifteenth-century Ottoman designers (and patrons) had at their disposal and is a built expression of the “visual cosmopolitanism” of Mehmed II’s court.110 Thus the Çinili Köşk was not unique in its combination of elements that evoke wide-ranging cultural references. That building, the palace, and Mehmed II’s mosque were all part of a longer tradition of complex engagements with architectural pasts and presents within the Ottoman context that had already begun by the late fourteenth century. We now turn to another case from fifteenth-century Istanbul that features references that might be read as Timurid: the mausoleum of grand vizier Mahmud Pasha.
Mahmud Pasha’s Patronage
Mahmud Pasha Angelović (d. 1474), a Serbian aristocrat turned Ottoman notable, served as grand vizier from 1456–68 and 1472–73, a career that ended in his execution after his second dismissal.111 As a patron of architecture, he sponsored a mosque and related buildings in Istanbul beginning in 1463 (the date of the mosque112), two no-longer-extant mescids in the same city, fountains, and a palace.113 His patronage also extended to Thrace and the Balkans. In Edirne, Mahmud Pasha built the Taşluk Camii, demolished in 1939, and a hammam.114 A mosque in Hasköy near Edirne and a palace in Skopje no longer survive, while a mosque in Sofia currently houses the city’s archaeological museum.115 The Fidan Han in Bursa, dated 866 AH / 1462 CE, has also been preserved.116 Unlike some of his contemporaries, such as Zaganos Pasha, İshak Pasha, and Gedik Ahmed Pasha, who built mostly in the Anatolian cities that were their bases of power, Mahmud Pasha focused on Istanbul and the Balkans.117
The Mahmud Pasha Mosque (see Figure 21, center), built in 1463 near Istanbul’s Grand Bazaar, is constructed on a variant of the inverse T-plan, a type of structure established in Ottoman architecture in the fourteenth century.118 Small side domes and a porch enclose the larger domes of the prayer hall.119 The mosque’s exterior is covered with stone cladding devoid of decoration. Analyzing the mosque within the context of late fifteenth-century Ottoman architecture, Kafescioğlu points out that the plan expands the notion of the mosque-zāviye, perhaps inspired by additions to existing churches that had been enlarged with subsidiary (often funerary) spaces, and thus clearly referred to the Byzantine architecture of Constantinople.120 In adapting such a plan for an Ottoman mosque-zāviye, its builders used the areas wrapped around the main space as a way to “segregate the main prayer area (now functioning as a congregational mosque) from the later spaces of the convent.”121 This transformation also reflects changing attitudes toward such buildings and the dervish groups that traditionally used them, beginning in the late fifteenth century.122 In Chapters 2 and 5, I discuss the role of mosque-zāviyes in the Ottoman context and their transformation into Friday mosques beginning in the second quarter of the sixteenth century. In Mahmud Pasha’s complex, the mosque-zāviye served as the centerpiece of a larger intervention into the urban fabric of a busy commercial district, located between the grand bazaar and the shore of the Golden Horn (Figure 21). In addition to the mosque-zāviye, the complex consisted of the patron’s mausoleum, a double hammam, khan, fountain, palace, public kitchen, and mektep. Of these, the mausoleum, khan, half of the hammam, and the fountain survive.123

21. Inclined map of the Mahmud Pasha complex, 1461–74, Istanbul, Turkey.
A strong contrast to the sober, stone-clad exterior of the mosque is offered in the patron’s mausoleum located behind the mosque. Its exterior decoration of turquoise tile inlaid into a marble covering, forming geometric patterns (see also Figures 27 to 29) lead us to examine techniques of stone carving and tile making. The difference in external appearance between the mosque and the mausoleum is striking and rare. The hammam belonging to the complex, which is located at some distance away from the mosque and mausoleum, is architecturally more akin to the mosque, with a sleek and monumental stone façade. As we shall see in Chapter 2, a similar mode of decorative variation between the exteriors of mosque, mausoleum, and madrasa is employed in Mehmed I’s funerary complex in Bursa.
With the tile decoration of Mahmud Pasha’s mausoleum, the question emerges whether the vizier was trying to emulate Mehmed II with it, perhaps emphasizing ambitions stoked by his participation in the conquest of Karaman in 1468. While this conquest took place after the construction of the mosque, the mausoleum was built in 1473–74, after the Ottomans conquered Karaman and a year after the construction of the Çinili Köşk. In effect, the two buildings are the only extant examples of exterior tile use in Istanbul – and rare examples within Ottoman architecture as a whole, together with Mehmed I’s mausoleum (1421–24) in Bursa, the Alaca Türbe (1470s?) in Skopje, and the Yeşil Cami (1378–91) in Iznik.
Unlike his mausoleum, Mahmud Pasha’s mosque and hammam do not stand out aesthetically among their fifteenth-century Ottoman counterparts. These buildings are in line with other monuments built by Mehmed II’s viziers in Istanbul – for instance, those of Rum Mehmed Pasha, Has Murad Pasha, and Ishak Pasha, all part of the larger project of turning Constantinople into the new Ottoman capital.124 This observation leads to the question of whether the “eastern” style of the mausoleum was an expression of the patron’s taste. Was it a way to connect to Ottoman elite tastes as a Serbian prince? Could it have been one of the variants found in the repertoire of cosmopolitan Ottoman courtly aesthetics? And was the use of tile – a rare and somewhat exotic material in Istanbul at this time – a mark of status in and of itself? A closer look at the mausoleum’s decoration and comparative examples will provide initial insights that also relate to other elements of the fifteenth-century Ottoman architectural project.
The Mahmud Pasha hammam (Figure 22), built in 1461, is partially preserved a short distance from the mosque and mausoleum. It was originally a double hammam, but the women’s section fell into ruins by the late nineteenth century and the former men’s section currently houses a number of small clothing stores.125 Over the entrance, a poetic inscription (Figure 23) in Arabic praises the building’s beauty, making a paradisiacal comparison: “It is a building filled with beauty and its marble [gleams] in the light like a lamp; the gardens of Eden: their rivers flow. Its date: it is the repose of spirits.” This inscription includes a chronogram whose last phrase (hiya rāḥat al-arwāḥ) gives the date of construction: 871 AH / 1461 CE.126 While no marble has been preserved in the interior of the monumental entry hall, this common material for the walls and floors of hammams appears in the caldarium.127 The allusion to the interior space in the inscription over the entrance allows the reader to envision what one might have originally encountered upon entering the building. The references to paradise enhance the impression of beauty and the mention of the rivers of the gardens of Eden connects to the water that would certainly be present within the building, given its function. The chronogram furthermore emphasizes another sensory component of the monument: the relaxing environment of the hammam – “repose of spirits,” as the inscription says – which would be enhanced by the sound of water flowing over marble within the building. At the same time, the hammam is built on a monumental, imposing scale, particularly visible from the outside. In the interior, although the original features of the hammam are no longer preserved, one would have had the sense of an intimate space where people gathered both for the cleansing rituals of bathing and to converse and share gossip as they lingered.128

22. Mahmud Pasha Hammam, 1464, Istanbul, Turkey.

23. Foundation inscription, Mahmud Pasha Hammam, 1464, Istanbul, Turkey.
The mausoleum (Figures 24 and 25) is located in a cemetery behind Mahmud Pasha’s mosque. Due to restoration of the mosque, ongoing since 2014, the cemetery is not currently accessible. The mausoleum, which was not included in the restoration project as of July 2019, the last time I was able to visit due to the COVID-19 pandemic that began in February 2020, is visible from a street running behind the eighteenth-century Nuruosmaniye Mosque down to the entrance of the Mahmud Pasha Mosque, and also from a parking lot behind two apartment buildings on Şeref Efendi Sokak in the vicinity of the Grand Bazaar. The octagonal building is decorated in a similar way on each of its sides: the bottom third of the wall is covered with grey marble paneling (Figure 26), while in the upper portion pieces of tile are set into geometrical strapwork composed of the same marble (Figure 27). Windows are set into the lower parts of the walls beneath carved stone lunettes on seven of the mausoleum’s sides, while on the eighth – the north side pointing toward the mosque – the entrance is located.129 Another higher set of windows appears in the upper section of the building among the tile work.

24. Street view of Mahmud Pasha Mausoleum, 1474, Istanbul, Turkey.

25. View from parking lot, with lower part obstructed by railing, Mahmud Pasha Mausoleum, 1474, Istanbul, Turkey.

26. View from parking lot, Mahmud Pasha Mausoleum, 1474, Istanbul, Turkey.

27. Detail of stone and tile decoration, Mahmud Pasha Mausoleum, 1474, Istanbul, Turkey.
In her study of late fifteenth-century Istanbul, Kafescioğlu argues that the “Timurid style geometric patterns” were used as a way of carrying the Persianate visual idiom, otherwise restricted to the palace, where monument such as the Çinili Köşk stood, into the urban sphere beyond the royal precinct.130 The visual connection between palace and city, extended by way of a patron closely associated with Mehmed II, is certainly an element that emerges in the mausoleum’s tile decoration. Perhaps surprisingly, however, the mausoleum’s façade does not simply repeat the tile mosaic of the Çinili Köşk. Rather, small pieces of tile are set into carved stone, a technique that does not have its roots within the Timurid context and that is indeed exceedingly rare among extant monuments within the Ottoman context and beyond.
The overall composition of the tile decoration follows the same model on each of the building’s sides, with simple carved stone molding following the lateral and upper edges (Figures 28 and 29). The same type of molding is used to form narrower rectangular frames that rise from the bottom of the structure, along the sides of the lower windows, up to about fifty centimeters below the roofline. The upper windows are placed above, within these same inner frames. The upper portion of the frame is filled with geometric tile patterns while stars and medallions are located in the spandrels of the upper windows, which form pointed arches. The space between the two sets of stone moldings is also filled with tile patterns, distinct from those surrounding the windows. In a rectangular frame surrounding each of the upper windows, girih (knot) motifs alternate in tile and stone.131

28. Detail of upper corner moldings, Mahmud Pasha Mausoleum, 1474, Istanbul, Turkey.

29. Detail of pattern of girih tiles around window, Mahmud Pasha Mausoleum, 1474, Istanbul, Turkey.
Such patterns based on girih designs were a long-standing tradition in Islamic art, reaching back to at least the so-called Sunni Revival in early twelfth-century Syria, when they emerged strongly.132 Following the Mongol conquest of much of the eastern Islamic world in the second and third quarters of the thirteenth century, girih patterns continued to be used, often in combination with the new khiṭāyī patterns imported from East Asia via the Mongol channels of cultural exchange and trade. By the fifteenth century, such patterns had become fully integrated into the arts of the eastern Islamic world (and, to a lesser extent, the Mamluk realm), and geometric and floral patterns could be freely combined.133 While this was not done in the tile work of the Mahmud Pasha Mausoleum, which sticks fully to geometric girih designs, the tile work of the Çinili Köşk to some extent shows this type of integration with the presence of square kufic medallions in the floral patterns and with the fully geometric designs underneath the roofline of the porch and in the fragments on interior arches (see Figures 5 and 6).
On each side of the mausoleum, patterns are repeated. One pattern is created with five-pointed stars arranged to form stacked polygons, with small kite-shaped quadrilaterals of dark-blue tile arranged in eight-pointed stars at their centers. In the rectangular panels beneath the windows, carved dodecagons are overlapped in such a manner as to form a stone grid of negative space that presents six-pointed stars, girih shapes, and badge-shaped hexagons that are filled in with tiles (see Figure 27). In the tile work, dark blue and turquoise alternate. Where tiles on the building have been broken, it is clear that they are made of red clay rather than fritware, which is off-white or white. (This will be relevant when discussing the possible sources of these tiles and their elusive makers.) In sections where tile elements have fallen out (see Figure 27), it becomes clear that the Mahmud Pasha Mausoleum is an achievement of stone carvers more than tile makers. The carefully arranged geometrical patterns are carved out of the stone with minute detail. The small pieces of tile would then have had to be cut to measure and inserted into the empty spaces. The workers who cut these tile insets were most likely familiar with the technique of tile mosaic, but only a limited number of simple geometric shapes were needed in this case, and their work would thus have been rather limited. Rather than creating a sleek surface, the technique highlights the materiality of both stone and tile, each appearing in sharp opposition to the other. Perhaps we are looking here at a one-off association between stone carvers and a group of tile cutters. The designs might have been drawn on paper, akin to the patterns found in the fifteenth-century Topkapı Scroll, which shows “a repertory of geometric designs for wall surfaces and vaults.”134 However, the actual types of templates that would presumably have been used have not been preserved.135
As Necipoğlu demonstrates in her extensive study of the Topkapi Scroll, the drawings contained in that source were not the actual designs used by architects on site, but rather “mnemotechnic devices that assured the preservation and transmission of architectural knowledge over the generations.”136 The same is perhaps true for earlier traces of architectural design, such as the famous Ilkhanid stone slab from Takht-i Sulayman and a carving from medieval Armenia bearing what appear to be patterns for muqarnas designs.137 The actual paper templates themselves, however, do not survive precisely because they were designed to be cut up and used on site during the process of producing the tile work. Thus the materiality of the template – which, as anthropologist Tim Ingold notes, is an artifact that is made from specific materials, just like the object it serves to produce – is no longer accessible to us.138 Initial, small-scale drawings and calligraphies that would eventually be used in architectural contexts may still exist in albums (which I discuss in Chapter 5), but they are most likely not recognizable as such, as preserved designs could have been used at a range of scales and in various media.
Since nothing comparable to the tile work on the Mahmud Pasha Mausoleum exists in Istanbul, the question is how this monument fits within the larger context of fifteenth-century Islamic architecture within and beyond the Ottoman realm. To what extent is this an Ottoman experiment? How did this monument’s makers use architecture to express the Ottoman endeavor to carve out a place among the post-Mongol empires in the Islamic world? While Kafescioğlu is right to state that Timurid visual culture was a relevant reference in this context, the makers of the Mahmud Pasha Mausoleum also drew on a wider range of architectural practices in the Islamic world.
More than evoking Timurid architecture – in which stonework hardly played a role and tile mosaic had been the dominant form of decoration since the late 1380s – the patterns here draw from an aesthetic that harks back to the Saljuq monuments of central Anatolia. At first sight, the geometric motifs created with strapwork carved in stone and tile insets in the empty spaces evoke the geometric designs on the portals of thirteenth-century Saljuq monuments in central Anatolia or the tile work on certain other buildings of the same period. At the same time, the reference is one that does not find direct comparisons: thirteenth- and fourteenth-century examples are either carved solely in stone, such as on the Sultan Han (ca. 1230–34) near Kayseri, or created in tile mosaic, such as in the Gök Madrasa in Tokat (see Figure 15). The stone-and-tile combination is a new, creative solution, perhaps a case Jonathan Hay evokes in noting that at times, a memory will be enough for transmission of motifs.139 In evoking these references, the Mahmud Pasha Mausoleum is unique within Istanbul, although examples in earlier Ottoman architecture are discussed in subsequent chapters.
Transfer to Skopje: Tile Work on the Move
Another example of the rare exterior use of tile in Ottoman architecture appears on the so-called Alaca Türbe (Figure 30) in Skopje. This mausoleum does not have an inscription, but Maximilian Hartmuth suggests that it was built in the 1470s, possibly for Paşa Beğ, one of the sons of İshak Beğ (d. 1445), the ucbeyi of Skopje.140 It is located behind the zāviye that İshak Beğ built in 1438–39 as part of a larger complex; the madrasa and the hammam do not survive, but the Sulu Han, one of two hans (commercial buildings) connected to the foundation, still stands.141 On this building, blue and turquoise tile elements are placed around the drum of the dome (Figure 31), forming a geometric pattern set into the stone. Star- and flower-shaped medallions of tile are also set into the spandrels of the window arches (Figures 32 and 33).

30. Alaca Türbe, 1470s, Skopje, North Macedonia.

31. Detail of dome and tile decoration on drum, Alaca Türbe, 1470s, Skopje, North Macedonia.

32. Detail of upper-level window with tile medallion, Alaca Türbe, 1470s, Skopje, North Macedonia.

33. Detail of tile medallion, Alaca Türbe, 1470s, Skopje, North Macedonia.
Hartmuth argues that the tile decoration may have been created by workers sent from Edirne or Istanbul.142 This argument may have its merits, considering that there really are no other examples of fifteenth-century tile work in the Ottoman Balkans. At the same time, it also falls into the pattern of assuming movement from imperial centers – Istanbul and Edirne – to frontier peripheries. In the case of the Alaca Türbe, a further question emerges – namely, whether tile workers would have had to be present on site to create the decoration found on it. Would not pieces of tile, either precut or ready to be cut on site, have been enough? The geometric pattern around the drum of the dome is relatively simple, with irregular diamond shapes arranged to form hexagons. The spaces in between are filled with what looks to be copious amounts of modern mortar or concrete, so it is harder to understand how the tiles were originally installed in the stone structure. The medallions in the window spandrels are done differently, with the tile pieces set into spaces carved into the stone. The shapes are relatively simple here as well – namely, triangles of various sizes. Since none of the tiles are currently missing, perhaps thanks to restoration, it is impossible to tell how deep the cuts into the stone are; indeed, it seems that some of the tile triangles are not quite flush with the stone, perhaps an indication of unequal depth. Most importantly, however, this kind of decoration could easily have been completed by local stoneworkers who were provided with tile pieces that might have been shipped from anywhere – including Istanbul, where tiles were presumably produced at a relatively significant scale for the decoration of the Çinili Köşk as well as the no-longer-extant tile work on the Old Palace. That the stone carvers who were involved in building the Alaca Türbe were capable of precise work is clear from the carving of the moldings on the mausoleum and the muqarnas cornice just below the tile panel. Thus it is absolutely conceivable to imagine a scenario involving a design made in one place, tiles being requested from a place of production and shipped to Skopje, and the installation of the tiles there by the same workers who completed the stone carving. A moving workshop of tile makers – one of the big topics of debate in the context of much of fifteenth-century Ottoman production – is not necessary in this specific case.
From these observations emerge two larger topics of discussion. First, the question of drawings in the process of architectural production in the fifteenth-century Islamic world needs to be addressed. The most important starting point is Necipoğlu’s discussion of the Topkapı Scroll, which raises the question of the relationship between the theoretical and practical elements of these and similar drawings. This also relates to a larger discussion in the history of sciences in the medieval Islamic world of the relationship between ʿilm (theoretical knowledge) and ʿamal (the practical expertise and skill in doing or making something). These are crucial in the realm of art and architecture, where we often see the results of ʿamal – namely, the buildings and objects created by its means – but know little how ʿilm feeds into it. The fifteenth-century manuscript Fī tadākhul al-ashkāl al-mutashābiha aw al-mutawāfiqa (Paris, Bibliothèque nationale de France, Ms. Persan 169, fols. 180r–199r) is a crucial text to advance our understanding of these issues with regard to architecture because it sits at the intersection of theoretical and practical applications of geometry.143 A copy of a treatise probably compiled by Abu Bakr in the early fourteenth century in an Ilkhanid context, the manuscript presents “the unresolved implications concerning the intersection, or lack thereof, between mathematical knowledge and artisanal expertise.”144 With regard to the mathematical exactness or absence thereof of the geometrical figures, Jan P. Hogendijk notes that the manuscript may reflect some of the practical solutions used in the context of making art and architecture, rather than more complicated theoretical ones. Moreover, considering the fact that some figures are presented without textual explanations or only vague ones, the manuscript also assumes the presence of a teacher along with students, within an oral tradition of instruction rather than a written one of self-study.145 Such a collaborative milieu, in which relationships between teachers and students or between members of a workshop were crucial, connects to the kitabkhāna environment.
The second question is the related issue of local traditions versus migrating artists, to use the terms from the subtitle of Michael Meinecke’s book on that topic.146 We should not necessarily assume stasis in these workshops and collaborations: considering the changing nature of building sites, in particular, we should imagine environments in which individuals could move around and changing sets of workers could collaborate. In the absence of detailed records about artists, their workshop genealogies, and often even their names, how can we understand if a building is the work of locals, a migrating group, or a combination of both? Is style always a trustworthy criterion? Which elements can deceive? To what extent are we to believe attributions that emerge from the recording of artists’ names, particularly nisbas that seem to indicate place? These questions will reemerge along with some answers in the next two chapters.
Mid-Fifteenth-Century Aesthetics: Istanbul and Beyond
In this chapter, I have examined how in Istanbul during the reign of Mehmed II, the aesthetic impact of the eastern Islamic world – the larger (post-)Timurid realms of Iran and Central Asia, but also contemporary Karamanid and thirteenth-century Saljuq Anatolia – became apparent in Ottoman-sponsored monuments, side by side with references to Renaissance Italy and the Byzantine Empire. Clearly these elements were consciously combined with ones that had been present in Ottoman architecture prior to the conquest of Constantinople.
In the case of the Çinili Köşk, the ideological intent behind the monument is to reflect and possibly surpass the monuments of the city of Karaman, as Necipoğlu has demonstrated. At the same time, the artistic means through which this is done are more varied than a simple copy of Karamanid architecture. Although the Ottomans destroyed the Karamanid palaces, what survives of the Karamanids’ patronage in religious monuments suggests that the kind of extensive tile mosaic seen in the Çinili Köşk had largely fallen out of use by the mid-fifteenth century. In fact, as Meinecke shows, most extant examples of such rich tile mosaic in central Anatolia date to the thirteenth century, with the occasional example appearing in a fourteenth-century monument.147 Thus the argument that the tile cutters from Khurasan with their desperate petition were indeed the makers of the Çinili Köşk’s tile mosaic is quite convincing, particularly considering the specific expertise that would have been needed in order to complete a relatively large-scale project in this complex technique. The interwoven inscriptions of the portal would certainly have required the involvement of competent calligraphers, who were at hand in the nakkaşhane in Istanbul, an increasingly centralized workshop led during the reign of Mehmed II by the artist Baba Nakkaş, whom we encounter in Chapter 5. Meanwhile, the poetic content of these inscriptions, as well as poetic engagement with this and other buildings, displays the extent to which cosmological motifs and notions surrounding artistically produced beauty – hence the process of artifice – were appreciated and embedded in the architecture.
The case of the Mahmud Pasha Mausoleum is more complicated, for this is not a work of tile mosaic making as much as one of stone carving. While large-scale drawings could certainly have been made of the patterns, these would have been needed to serve the carving of the stone to create spaces for tile fill rather than to produce panels of tile mosaic for installation on the monument. The tile shapes used on this monument are limited to a few types, all fairly simple and all having straight edges, with none of the complicated curved shapes of the Çinili Köşk’s inscriptions. In this case, tile cutters would not have been particularly challenged and could even have supplied the shapes while working off site from carefully measured templates. The stone carvers who completed the work on the Mahmud Pasha Mausoleum, on the other hand, would have needed to work meticulously to create the stone strapwork into which tiles were inserted and to make sure that it was uniform on all sides of the monument. Perhaps the fact that this technique did not catch on and was not repeated is a sign of how challenging this project proved. The Alaca Türbe in Skopje, as we have seen, has tile inserts on a much smaller scale and in far simpler designs, and it was not necessarily derived from the Çinili Köşk or made by the same group of tile cutters.
Finally, the tile lunettes in the mosque of Mehmed II in Istanbul point toward the future direction that Ottoman building ceramics would take – namely, increasingly complex underglaze-painted fritware that would culminate with the highly centralized production of Iznik in the mid-sixteenth century. In the second half of the fifteenth century, this was by no means determined as of yet: both vessels and tiles in a range of styles were produced at various sites including Istanbul and Iznik.148 These developments would continue during the reign of Bayezid II, as the last chapter of this book shows, a period when designs created on paper may have taken precedence, leading to an increasing overlap in the decorations of manuscripts, tiles, pottery, textiles, and metalwork. At the same time, a dialogue with the pasts of lands now under Ottoman rule would increasingly be created in chronicles produced during Bayezid II’s reign to present a carefully crafted Ottoman past that aligned with the empire’s present.
The following chapters, however, explore how some of the features visible in the Çinili Köşk and the Mahmud Pasha Mausoleum have precedents in earlier fifteenth-century Ottoman architecture. I examine elements such as tile decoration with origins in the eastern Islamic (Timurid, but also Qaraqoyunlu and Aqqoyunlu) world and in the Saljuq architecture of Anatolia; dialogue in stonework with the Mamluk monuments of Greater Syria and Egypt; and the sensory experiences created by the combination of these elements. Chapter 2 focuses on the mosque complex of Mehmed I in Bursa, where such elements were introduced on a substantial scale, carrying meaning in a period of dynastic reconstruction that called for buildings that projected kingship as much as wonder.