A BUILDING BETWEEN THE PUBLIC AND PRIVATE REALMS OF …

the sufi convent of sokollu mehmed pasha in istanbul 159

Sokollu Mehmed Pasha, one of the most influentialand propertied men in Ottoman history, who from1565 until 1579 served three subsequent sultans asgrand vizier, justified his successful career in publicservice by his patronage of numerous public build-ings throughout the Empire, located in places as var-ied as Hungary and Syria. The public building complexhe commissioned in Kadærga (âdærÚa Lim¸næ), at thesouthern edge of the Hippodrome in Istanbul,1 standsout among Sokollu’s public building complexes be-cause of its physical proximity to his private realm.He built the mosque there in 1571–72, on the site ofan old church; his wife, Ismihan, legally endowed it.What remains of the group of Sokollu’s buildingsaround this mosque are a madrasa completed by 15702

and a dervish convent added around 1574 (fig. 1).The site of these buildings, however, was part of thelarge piece of land the grand vizier and his royal wifeowned in Kadærga, where their palace compound wasalso located (fig. 2).3 At the time the public build-ings were completed, the couple had abandoned thepalace (referred to as the “old palace,” d¸rüssa{¸de-i{atº_a, in Sokollu’s endowment deed)4 in favor of a“new palace” (d¸rüssa{¸de-i cedºde) on a site nearbywhere the Sultan Ahmed mosque now stands.5 Pre-sumably, the old palace then came to house the ex-tended household of Sokollu and Ismihan, given thatthe grand vizier’s endowment deed suggests no newpublic function for the compound.6

The Kadærga complex was supervised by Sinan, therenowned chief court architect, who codified what wenow know as “classical Ottoman” architecture. Likeseveral earlier public building complexes in Istanbul,the main element of the Kadærga complex is a loftymosque with a symmetrically-arranged madrasa sur-rounding its courtyard. Its architectural program dif-fers from Sinan’s earlier mosque-madrasa complexesonly by the incorporation of a Sufi convent locatedbehind the mosque (fig. 3). An exhaustive study of

the architectural features of the convent was accom-plished by Baha Tanman;7 in an effort to make knowna hitherto neglected building, Tanman presented itin isolation. I will now try to build on that study bysituating this building in space and time, investigat-ing the historical context that defined the relation ofthe convent to, and its difference from, the rest of thebuilding complex.

Primarily, this article investigates the historicalconstellation that accounts for the peculiar architec-tural program that brought together a madrasa anda Sufi convent. This combination was rare in sixteenth-century Istanbul, although it had been common inthe fourteenth and fifteenth centuries throughout theIslamic world. Around the beginning of the sixteenthcentury, the ties between Sufi groups and the Otto-man ulema were severed by the transformation of aSufi network into an openly anti-Ottoman state withclaims to world rule—namely that of the Safavids. Thecoupling of madrasa and Sufi convent in the archi-tectural program of major building complexes inIstanbul became problematic thereafter; the few knownsixteenth-century examples should therefore be treatedas a problem, which the present article seeks to ad-dress. Given the issue in question, this study is as muchabout the relation of a group of Sufis to “the madrasa”in the conceptual sense—or the religio-legal cadretermed {ilmiyye—as it is about the architectural andsocial relation of that particular Sufi convent to theadjacent madrasa.

I also question herein why the convent buildingin Kadærga deviates from the architectural paradigmthat was pursued by Sinan in the design of earliergrandee buildings in Istanbul and in the rest of theKadærga complex. Possibly the convent was in factdesigned, not by Sinan, who is recorded to have su-pervised a conspicuously small number of Sufi con-vents throughout his career,8 but by one of theyounger architects who worked for him. Still, there

ZEYNEP YÜREKL~

A BUILDING BETWEEN THE PUBLIC AND PRIVATE REALMSOF THE OTTOMAN ELITE: THE SUFI CONVENT OF SOKOLLU

MEHMED PASHA IN ISTANBUL

muq-20-2.pmd 9/2/2003, 4:23 PM159

zeynep yürekli.160

Fig. 1. Aerial view of the Sokollu Mehmed Pasha complex in Kadærga, Istanbul. From left: the madrasa, the mosque, and theconvent. (Photo: R. Günay, from Sinan’æn Istanbul’u [Istanbul, 1987])

Fig. 2. Map of Kadærga, Istanbul.

muq-20-2.pmd 9/2/2003, 4:23 PM160


Fig. 3. Site plan of the Kadærga complex and section through the central axis of the mosque. 1. convent 2. mosque 3.madrasa 4. cemetery. (After drawings in Kuban, “An Ottoman Building Complex”)

is room for an argument that goes beyond the iden-tity of the architect, and evaluates this building withits marginal position between the public-religious andprivate-residential idioms in Ottoman architecture.

THE SHAYKH OF THE PATRON

Sokollu Mehmed Pasha’s patronage of Sufism extendsacross most of the Ottoman territory. His long and

detailed endowment deed of 1574 lists three dervishconvents to be included in his public building com-plexes.9 In addition to the convent in Istanbul, therewas one in Szigetvár (Hungary),10 and another oneequidistant from Istanbul to the southeast, in Payas,near Iskenderun.11 This patronage also appears to havebeen carefully selective. The endowment deed stipu-lates for each convent that the shaykh should be “amystic with a heart alive” ({¸rif-i zinde-dil), without

muq-20-2.pmd 9/2/2003, 4:23 PM161

zeynep yürekli.162

mentioning any specific shaykh or Sufi lineage. Con-temporaneous and later sources recount, however, thatSokollu endowed the convent in Kadærga specificallyfor his own spiritual advisor, Nureddinzade MustafaMuslihuddin (d. 1574), a Halveti shaykh from Filibe/Plovdiv (Bulgaria).12

The shaykh Nureddinzade appears in contempo-raneous sources as a pious man of modest attire,

knowledgeable in Islamic sciences and profoundlycommitted to orthodox Islam. His commitment wassuch that he waged war against unorthodox Sufi groupswith “the sword of shari{a” (×erº{at _ælæcæn çalardæ).13

He and his shaykh, Sofyalæ Bali, are known for thesupport they lent to the Ottoman Sunni ideologyagainst certain dervish groups in the Balkans consid-ered heretical by the state, and also against the Safa-

Fig. 4. Floor plans of the dervish convent in Kadærga. 1. tev¥ºdÒ¸ne/ritual space 2. lower courtyard 3. upper courtyard 4.shaykh’s quarters 5. ÒalvetÒ¸ne/retirement cell (?) 6. kitchen and eating room (?) 7. bathroom-latrine 8. cemetery 9.shops. (Based on drawings in B. Tanman, “Tekkeler”)

muq-20-2.pmd 9/2/2003, 4:23 PM162


vids.14 Sofyalæ Bali was called by a biographer “thespy of the shaykhs” (çs¢sül-me×¸yiÒ),15 by virtue ofhaving warned the state against improper activitiesof shaykhs of some sub-orders of the Ùalvetiyye in theprovinces.16

When Nureddinzade came to Istanbul in the 1550sthe current ×eyÒülisl¸m, Ebussu{ud, was investigatinghis activities and those of Sofyalæ Bali based on accu-sations of heresy.17 According to the biographer {Ata}i,Nureddinzade survived the investigation thanks to anintentional encounter with Ebussu{ud at which heimpressed the ×eyÒülisl¸m immensely with an acutecomment on the interpretation of a passage from theQur}an, thereby convincing him that the accusationswere false.18 Upon Ebussu{ud’s recommendation, thecurrent grand vizier, probably Rüstem Pasha, offeredNureddinzade the convent of Küçük Ayasofya (sur-rounding the courtyard of the former Byzantine churchof Hagios Sergios, which had been converted to amosque),19 already known for its orthodoxy (see fig.2).20

Throughout the rest of his life at the Küçük Aya-sofya convent, Nureddinzade was recognized as a trust-worthy shaykh by the upper class of soldier-statesmenand court officers in Istanbul, who wanted to base theirsupport of Sufism on the views of an insider with ad-equate knowledge of religious sciences. They soughtNureddinzade’s opinion of other shaykhs with regardto their orthodoxy,21 and patrons of new Halveti con-vents concerned with the uprightness of their conventsconsulted him for the appointment of shaykhs. We readof a palace officer, for example, who endowed a newconvent for “upright Halvetis” (ªule¥¸}-i Ùalvetiyye) inTærhala/Trikkala (Greece) and asked Nureddinzadefor the appointment of one of his disciples there.22

Nureddinzade also negotiated for the appointment ofhis disciples to other positions. Such is the case of aformer madrasa professor, who was appointed with theshaykh’s mediation as the private teacher of the sonof a vizier and the assistant functionary of that vizier’spalace.23

Nureddinzade eventually became so influential inIstanbul and the Balkans that “the [Halveti?] orderbecame public in the land of Rum as the Nureddinzadeorder (N¢reddºnz¸de «arº_ati).”24 His influence emanatedfrom his extraordinary skill in preaching, a skill healso seems to have taught to his numerous disciples.25

He himself preached in several prestigious locationsin Istanbul, and his sermons were attended by promi-nent ulema. In the meantime, Sokollu Mehmed Pa-

sha became attached to him, and Sultan Süleymanbefriended the shaykh, inviting him to the palace oftenfor private spiritual conversation. In 1566, Nured-dinzade paid an unexpected night visit to the sultan,in order to tell him of his dream of the Prophet de-manding that the sultan resume ghaz¸h (religious war-fare);26 the Szigetvár campaign is supposed to havebeen the consequence. Nureddinzade joined Süley-man’s campaign to Szigetvár as the “army shaykh” (ordu×eyÒi), while Sokollu Mehmed Pasha, as chief com-mander, was in charge of the campaign.27 When Sü-leyman died during the campaign, Nureddinzade wasone of the select three men who took the deceasedsultan’s coffin to Istanbul for burial.28

Sokollu’s building complex in Szigetvár is describedin the endowment deed as “outside the castle of Szi-getvár, on the location where the honorable body ofthe deceased sultan Süleyman was temporarily put.”29

Since Süleyman died during the campaign in 1566,his body was temporarily entombed within the currentcampsite of the Ottoman army, at a certain distancefrom the castle of Szigetvár. Given that Nureddinzadejoined in the campaign as the “army shaykh,” the in-clusion of a Halveti convent there may have been theresult of his proposition to Sokollu. A disciple ofNureddinzade, probably suggested to Sokollu by him,was appointed as the shaykh of that convent.30

Little is known about Sokollu’s third convent inPayas. Fifty years after its construction Evliya Çelebidescribed the building as a madrasa,31 although Sokolluplanned it as a Sufi convent and prohibited its con-version into a madrasa in the endowment deed.32

Nureddinzade died in 1574, before the dervishconvent in Kadærga could be completed for him to takecharge of, and only a month or two before Sokollu’sendowment deed was legally activated.33 Before dying,Nureddinzade sent Sokollu his will concerning thechoice of a shaykh to replace him in Kadærga, namelyhis fellow-in-lineage (pºrd¸×) Mehmed b. {Ömer (d.1587–88), more popularly known as Kurd Efendi.34

Both shaykhs belonged to the part of the Ùalvetiyyethat was closest to the Ottoman bureaucracy, namelythe branch known as Cemali-Halveti. The name isderived from the so-called Cemali family, of which thefirst Halveti shaykh in Istanbul, Çelebi Halife (d. 1497–98), was a member. Sixteenth-century members of theCemali family included prominent shaykhs, {ilmiyyeofficers, and high-ranking bureaucrats. Early in thesixteenth century, two Cemalis rose to the highest ranksof the Ottoman religio-legal bureaucracy and admin-

muq-20-2.pmd 9/2/2003, 4:23 PM163

zeynep yürekli.164

istration: Zenbilli {Ali (d. 1525) became ×eyÒülisl¸m in1503, and his cousin Piri Mehmed Pasha (d. 1532–33) was appointed grand vizier in 1517.35 With suchfamilial relations to the bureaucracy, Halveti shaykhswhose spiritual lineage (silsila) went back to ÇelebiHalife were privileged for patronage in Istanbul.Among them, a certain line of shaykhs coming fromÇelebi Halife’s relative and disciple Kasæm Çelebi aredescribed in contemporaneous Ottoman sources asbeing particularly committed to adherence to theshari{a and to occupation with external (¬¸hirº) Islamicsciences as much as with esoteric (b¸«inº) sciences.Nureddinzade and Kurd Efendi came from that line,generally termed Cemali-Halveti after Kasæm Çelebi’sfamily name.

THE CONVENT OF THE SHAYKH

The convent is centered on a domed ritual space(tev¥ºdÒ¸ne) and two interconnected courtyards, onehigher than the other by one story and the two ac-cessible to each other by a staircase running behindthe tev¥ºdÒ¸ne (fig. 4). Surrounding the courtyards isa series of residential spaces, consisting of thirty singlerooms for dervishes (twenty-one on the entrance leveland nine on the lower level to the southwest),36 anda larger room on the south corner of the entrancelevel that probably was meant to serve as the quar-ters of Nureddinzade. The service areas consist of aspace directly under the shaykh’s quarters (possiblythe kitchen and eating room, given that the end-owment deed does not list a separate {im¸ret in thecomplex), and a bath-latrine under the tev¥ºdÒ¸ne,accessible from the lower courtyard. Finally, there isa windowless space on the east corner of the entrancelevel that is very likely to have been the retirementcell (ÒalvetÒ¸ne).

With that functional scheme, the convent in itselfappears as a self-sufficient microcosm, incorporatingall the elements that communal dervish life required.It is also separated from the mosque-madrasa com-plex by a wall that provides access to the mosquecompound from the lower courtyard of the conventthrough a single door. The convent might as well bean afterthought, as it was still in construction whenNureddinzade died in 1574 (about three years afterthe completion of the mosque). Nevertheless, we havetwo pieces of evidence suggesting that it could nothave been planned as a compound functionally iso-lated from the rest of the complex: First, the endow-

ment deed of Sokollu Mehmed Pasha states that thesermon after each Friday prayer service in the mosqueis to be given by the shaykh of the adjacent dervishconvent—a very important statement to which I willreturn below.37 Second, as Tanman has also pointedout, the tev¥ºdÒ¸ne is not suitable for the regular prayerservices that Halveti dervishes were expected to per-form, since what should have been the qibla wall iswhere the entrance to the tev¥ºdÒ¸ne is located. Alongwith the madrasa students and outsiders, the dervishesmust have used the main mosque for regular prayerservices, followed by additional prayers, five times aday, as requisitioned by the patron,38 and includingthe Friday service followed by a sermon delivered bytheir own shaykh. For the education in external Is-lamic sciences that Nureddinzade and Kurd Efendiwere reported to offer in their convents,39 the booksendowed by Sokollu Mehmed Pasha to the madrasamight have been used.40

With its lavish interior decoration and protrudingdome, the tev¥ºdÒ¸ne is the centerpiece of the convent(figs. 5–7). With the exception of the misplaced“mihrab,” the interior decoration is typical of a mosque(figs. 8 and 9).41 The painted medallions bearing thenames of the Four Caliphs and Hasan and Husayn wereplaced in the order that was customary in mosquedecoration, although they conflicted with the falsemihrab. This “mihrab” was placed where the shaykhwould stand to oversee the rituals (fig. 10); it thusbecame a literal architectural embodiment of thewidely used metaphor of the Sufi shaykh as a mihrabthat his dervishes should turn to for their spiritualdevelopment. The inscription above the gate into thetev¥ºdÒ¸ne bears a hadith suggesting that this space wasmeant for specific Sufi ritual: “The prophet of God...said that the highest-grade people according to Allahare those who perform dhikr (al-dh¸kir¢n)” (figs. 11and 12).

Communal sessions of dhikr (liturgically regulatedpractice of litany, with prescribed rules regarding wordsequence, respiratory rhythm, physical posture, andorder of those who join the session), as in many otherSufi brotherhoods, appear to have been taken veryseriously in the convent of Nureddinzade. {Ata}i re-lates that a disciple of Nureddinzade, called Nured-din-i Zakiri, so excelled in the science of dhikr thathis melodic creations “have been accepted as incon-testable by all knowledgeable masters [of dhikr] forover sixty years.” The discovery of this great dhikrmaster occurred by chance one day, when the dervishes,

muq-20-2.pmd 9/2/2003, 4:23 PM164


Fig. 5. Aerial view of the tev¥ºdÒ¸ne from the southeast, with the mosque in the background. (Photo: K. Görkay)

Fig. 6. View showing the domes of the entrance hall and the tev¥ºdÒ¸ne, with the mosque in the background. (Photo: Archiveof the Mülga Eski Eserler Encümeni, Istanbul Archaeological Museum)

muq-20-2.pmd 9/2/2003, 4:23 PM165

zeynep yürekli.166

Fig. 7. View of the tev¥ºdÒ¸ne from the upper courtyard. (Photo: Z. Yürekli)

Fig. 8. Southwest arch of the tev¥ºdÒ¸ne, with medallions bearing the names of Abu Bakr and {Ali on the pendentives. Thewooden women’s gallery is probably a nineteenth-century addition. (Photo: Archive of the Mülga Eski Eserler Encümeni,Istanbul Archaeological Museum)

muq-20-2.pmd 9/2/2003, 4:23 PM166


←Fig. 9. Interior of the tev¥ºdÒ¸ne, looking towards themisplaced “mihrab” from the west corner, showing themedallions bearing the names {Uthman (on the pen-dentive) and Hasan and Husayn (above the “mihrab”).The Kufic insciption below the medallions, probablya nineteenth-century addition, bears a fragment of aQur}anic verse (39: 53). (Photo: Archive of the MülgaEski Eserler Encümeni, Istanbul Archaeological Mu-seum)

Fig. 10. Halveti dhikr in the eighteenth century. (FromI. M. d’Ohsson, Tableau général de l’Empire othoman[Paris, 1787])↓

muq-20-2.pmd 9/2/2003, 4:23 PM167

zeynep yürekli.168

←Fig. 11. View from the entrance hall towards the gate intothe tev¥ºdÒ¸ne. (Photo: Archive of the Mülga Eski EserlerEncümeni, Istanbul Archaeological Museum)

Fig. 12. Inscription above the gate into the tev¥ºdÒ¸ne withthe hadith, “The highest-grade people according to Allah arethose who perform dhikr (al-dh¸kir¢n).” The painted inscrip-tion above bears the basmala. (Photo: Archive of the MülgaEski Eserler Encümeni, Istanbul Archaeological Museum)↓

muq-20-2.pmd 9/2/2003, 4:23 PM168


having formed the usual circle of unity (¥al_abeste-itev¥ºd) in the presence of Nureddinzade, were unableto start the session because the current dhikr-leader(z¸kirba×æ) was absent, and Nureddinzade told Nu-reddin-i Zakiri to start the dhikr. That the ritual couldnot start without a dhikr-leader implies that dhikr inNureddinzade’s convent was not only a prescribed wayof achieving ecstasy (as is also common in many otherSufi orders) but also a matter of expertise within thebrotherhood. The inscription mentioning “those whoperform dhikr” above the gate into the tev¥ºdÒ¸ne isin direct reference to this specific practice.

With the rise of the Safavid state in Ardabil earlyin the sixteenth century, religious scholars in theOttoman Empire became increasingly suspicious ofdervish groups. The orthodoxy and permissibility ofsuch Sufi ritual practices as communal dhikr sessionsthus became a matter of dispute among religious schol-ars. Zenbilli {Ali, the above-mentioned Cemali ×eyÒülis-l¸m of the early sixteenth century, issued a fatwa infavor of communal dhikr, stating that “if a personattempts to break up a gathering of dhikr-performers(z¸kirler), the order of the shari{a is that he should bekilled.”42 Zenbilli {Ali supported his fatwa with a trea-tise that deals with the accusations against dhikr byprevious religious scholars, and counters them with anumber of hadith similar to the one above the gatein Kadærga.43 A similar treatise by the Halveti shaykhSünbül Sinan (d. 1529–30), a disciple of Çelebi Halife,also makes use of such hadith to prove his point.44 Aconsiderable part of Sünbül Sinan’s treatise discussesthe permissibility of communal dhikr, as opposed tosilent individual dhikr, and concludes that the formeris not only permissible but also better, because itmultiplies the spiritual effect.

The fact that both treatises were written in Arabicindicates that their primary intended audiences weremadrasa students and graduates, that is, prospectiveor actual ulema. The Halveti shaykh Hulvi (d. 1654),a spiritual descendant of Sünbül Sinan, recounts thatthe latter wrote his treatise as a result of a debate withthe ulema and sent copies of it to prominent religiousscholars.45 Another Cemali shaykh, Ishak Karamani (d.1526–27) wrote a similar treatise in response to a let-ter sent by a religious scholar, the preacher Molla {Arab,to all Halveti convents in Istanbul. Ta×köprizade re-ports that not only did Ishak Karamani convincinglyprove the permissibility of Halveti dhikr, but he alsopinpointed Arabic-language mistakes in Molla {Arab’sletter.46

The hadith at the entrance to the tev¥ºdÒ¸ne inKadærga and the architectural separation of that spacefrom the mosque both appear as implicit responsesto a party of religious scholars who contested the per-missibility of communal dhikr on the one hand, andthe permissibility of such Sufi rituals within mosqueson the other. Zenbilli {Ali’s above-mentioned fatwa wasissued in answer to a question about the permissibil-ity of breaking up the dhikr gatherings that the Sufisheld “in the mosques, after performing their Sunnireligious duties (ed¸-i fer¸}i¾ ü sünen).” There shouldbe no doubt, then, that in the early sixteenth centurythere were indeed people who wished to have themosques free of such Sufi ritual. According to Hulvi,the above-mentioned debate between Sünbül Sinanand the ulema arose because of the shaykh’s customof holding dhikr rituals in the mosques of MehmedII and Ayasofya after Friday sermons.47 Sokollu Meh-med Pasha’s complex provided the shaykh and his der-vishes with a separate and protected space where theycould perform dhikr sessions away from the sight ofthe students at the madrasa.

That this protected space for dhikr (apparently notserving the function of regular prayer services at all)should imitate a domed mosque is a curious phenom-enon. The later Halveti convent in the Atik Validecomplex (examined below), for example, similarly hasa ritual room within the convent compound, but whilenothing is known of its original interior decoration,both its size and its position in the compound arecomparable to a madrasa lecture room (dersÒ¸ne)rather than a mosque. The tev¥ºdÒ¸ne of Sokollu’sconvent is unique in that it is architecturally a mosquewithout serving the function of a mosque. This phe-nomenon should be evaluated against the backgroundof the development of Halveti ritual in Istanbul.

The first Halveti convent in Istanbul, the KocaMustafa Pasha convent of the late fifteenth century(also examined below), was centered on a Byzantinemonastic church that was converted into a mosque tobe used for regular prayer services as well as Sufi ritu-als. The particulars of sixteenth-century Halveti ritualdeveloped in that architectural setting, while an el-evated dome like that of the Koca Mustafa Pashamosque came to be recognized as the suitable super-structure for a space to house the ecstatic and circu-lar Sufi ritual (sem¸{) held in great esteem by Halvetis.The following anecdote told by a disciple of SünbülSinan, the second shaykh of the Koca Mustafa Pasha

muq-20-2.pmd 9/2/2003, 4:23 PM169

zeynep yürekli.170

convent, implies a curious connection between thedome of the Koca Mustafa Pasha mosque and theexperience of ecstasy during the ritual:

My shaykh Sünbül Sinan Efendi was a man of sem¸{. Healso got up [to perform it]. Whenever the shaykh cameto the sem¸{, the dome of the mosque was elevated andwe saw the angels whirling in the air.48

The convent of Küçük Ayasofya, where Nureddinzadewas shaykh, did not have a separate ritual space ei-ther. Dhikr and sem¸{ sessions were held in the mosque,as was common practice at a number of Halveti con-vents in Istanbul.49 In endowing a new convent witha relatively protected ritual space for Nureddinzadeonly a few minutes’ walk from his old convent, Sokollu’sprimary concern might have been protecting his shaykhfrom exposure to criticism regarding the permissibil-ity of dhikr, without depriving him of the kind of ritualspace he and his dervishes were accustomed to. Thetev¥ºdÒ¸ne of Sokollu’s convent, with its mosque-likedecoration and domed structure, accords with theritual habits that developed through three generationsof Halvetis in the mosques of Istanbul. The separa-tion of the tev¥ºdÒ¸ne from the mosque looks reluc-tant here: in its structure and decoration, the tev¥ºdÒ¸nestill wants to be a small mosque.

{ILMIYYE-ÙALVETIYYE RELATIONS AND THEARCHITECTURAL PROGRAM OF GRANDEE

COMPLEXES IN ISTANBUL

The arrival of the first Halveti shaykh in Istanbulcoincided with the rise of the Ottoman {ilmiyye as anempire-wide religio-legal cadre in the aftermath ofthe conquest of the city in 1453. Mehmed II (r. 1444–46 and 1451–81) systematically transformed Istanbulinto the capital of an Ottoman empire as he under-stood it—an empire with, among other things, a cen-tralized religio-legal system.50 As part of his imperialproject, he systematized the madrasas under an im-perial classification in 1471. Highest-ranking in thenew classification were the Sahn-æ Seman madrasas,built as part of his complex in Istanbul (1463–70).51

All other madrasas in the empire were now subordi-nated to that center in order to form an empire-wideeducational network for the {ilmiyye.

Following Mehmed’s death in 1481, as his two sonsBayezid and Cem were fighting over the right to thethrone, the shaykh Çelebi Halife stepped into thepolitical scene in Amasya, a provincial capital near

the Black Sea, and actively supported the eventualwinner, Bayezid, whose princely residence had beenin Amasya.52 After defeating his brother with the helpof Çelebi Halife’s party, Bayezid gratefully invited theshaykh to Istanbul.53 Seventeenth-century Halveti tra-dition held that Çelebi Halife then saw in a dreamthat a torch of divine radiance coming out of Amasyasoon traveled to the utmost limits of the Islamic lands,filling them with a light that covered the entire world,particularly the land of Rum.54

The dream narrative reflects the conviction of itsseventeenth-century author that the institutionaliza-tion of the Ottoman Ùalvetiyye had started with ÇelebiHalife in Istanbul. Bayezid’s former boon compan-ion Kapucæba×æ Hacæ Mustafa, later known as KocaMustafa Pasha (d. 1512), endowed a new convent forÇelebi Halife between 1486 and 1491.55 The site cho-sen was the Byzantine monastery Hagios Andreas ente Krisei.56 The monastic church was converted intoa mosque to form the center of the building com-plex now known as Koca Mustafa Pasha. The ar-chitectural program incorporated a Sufi convent, amadrasa, a hospice ({im¸ret), and a bathhouse.57 Inan arrangement similar to that of the complex inKadærga, at one side of the church/mosque is themadrasa, and at the other side the convent for ÇelebiHalife (fig. 13).

The story of the transformation of the OttomanÙalvetiyye into a network begins here. The crucial stepin the process of the institutionalization of a Sufi or-der was the propagation of a shaykh’s teaching to alarger community of his disciples.58 Convents were thusconnected to each other by spiritual kinship and sharedritual; they were considered akin to each other in thesense that, at some point in their spiritual lineage, theircurrent or previous shaykhs were pºrd¸×, or disciplesof the same shaykh. In sixteenth-century Istanbuland the Balkans, the network of Halveti convents, at-tached to each other by their shaykhs’ spiritual lin-eage reaching back to Çelebi Halife, was particularlyinfluential.

The convent of Koca Mustafa Pasha was recognizedas the center of the Ottoman Ùalvetiyye, whence spranga number of suborders (of which the Cemali-Halvetiswere only one). Religious groups in the sixteenth-century Ottoman lands had to live with the increas-ing legal and administrative authority of the {ilmiyye.59

As becomes apparent from the fatwas of Ebussu{ud,the activities of Sufi convents were under the constantsurveillance of the ×eyÒülisl¸m to make sure their prac-

muq-20-2.pmd 9/2/2003, 4:23 PM170


tices did not deviate from Sunni Islam. The Halvetisin particular, in order to keep their distinguishedposition in urban social life, were to engage in debatesregarding the permissibility of their ritual practices.

The aforementioned letter sent by Molla {Arab di-rectly to the Halveti convents in Istanbul forced themto take a defensive position. As the shaykh of the al-leged center of the Ottoman Ùalvetiyye, Sünbül Sinanhad the mission of convincing the ulema that Halvetirituals were in accordance with the shari{a. A spiritualdescendant of Sünbül Sinan describes the situationthen as follows: “The ulema were divided into twoparties. Most of them were on the side of the shaykhSünbül, for no one could defeat the shaykh in a schol-arly discussion.” Sünbül Sinan was invited to such ascholarly discussion in the mosque of Mehmed II by

his opponents, a group of eminent religious scholarsled by Saræ Kürz (or Saræ Görez, d. 1521–23), the judgeof Istanbul, and Kürz Seydi, a professor at the Sahn-æSeman madrasas in the building complex of MehmedII. As the scholars investigated Sünbül Sinan, he suc-ceeded in rendering their accusations unconvincingby sarcastically reminding the judge of their studentdays together in a madrasa, and by recalling for theSahn-æ Seman professor the ritual sessions he used tojoin when he was the professor of the madrasa in theKoca Mustafa Pasha complex:

He said: “When [Koca] Mustafa Pasha was the grandvizier, you used to be the professor at our madrasa. OnFriday nights, the pasha used to come to join the com-pany of the dervishes. You also used to come, take offyour turban, take off your robe, and whirl along with

Fig. 13. Koca Mustafa Pasha complex, Istanbul. Site plan. 1. mosque 2. dervish cells 3. madrasa 4. tomb of the patron5. tombs (later additions) (W. Müller-Wiener, Bildlexikon zur Topographie Istanbuls [Tübingen, 1977])

muq-20-2.pmd 9/2/2003, 4:23 PM171

zeynep yürekli.172

the dervishes (sem¸{ iderdin). Did you not know of thisproblem then? Or did you know of it but do so in or-der to please the pasha…?” Then he ascended theminbar and expended so much truth and subtlety thatmost of the scholars pledged allegiance. Eighteen madra-sa students became dervishes and entered the order.Abandoning their student robes and turbans, they puton dervish cloaks and caps.60

The continuation of the debate between orthodoxreligious scholars and Halveti shaykhs can be observedthrough the fatwas issued by the ×eyÒülisl¸m Ebussu{udon questions concerning Sufi practices.61 Ebussu{udwas not against Sufi rituals altogether, but he advisedstrict control over their conditions, permitting com-munal dhikr as long as there was no dancing (ra_ª),whirling (devr¸n or sem¸{), or instrumental music in-volved.62 One of the questions he was asked was if itwas appropriate to display hostility against the Halvetisbecause they performed dhikr with devr¸n. Ebussu{ud’sresponse was that “there are decent people among theHalvetis,” whose rituals were religiously acceptable, andthat whoever displayed hostility against them shouldbe reprimanded.63

Among several Sufi groups that partook of the cul-tural milieu of the Ottoman sixteenth century, theCemali-Halvetis are notable for their close relation tothe {ilmiyye. Cemali-Halveti shaykhs like Nureddinzadeand his shaykh, Bali, “served” the religio-legal cadre,so to speak, appropriating an outspokenly Sunnistance.64 That branch of the Ùalvetiyye posited itselfnot only as a spiritual path, but also as a path thatcommitted ulema could choose if they were discon-tent with a career in the {ilmiyye. Çelebi Halife was amadrasa professor until, according to Hulvi, he no-ticed the materialist orientation of the {ilmiyye careerand turned to Sufism.65 Such a retreat to Sufism fromthe {ilmiyye is quite typical in the biographies of Cemali-Halveti shaykhs, including Nureddinzade. Nured-dinzade’s disciples were also usually madrasa graduates,capable of fortifying their sermons with their knowl-edge of external religious sciences.66 The primaryactivities in Nureddinzade’s convent were “scientificdiscussions” (müz¸kere-i {ilmiye) attended by “a selectgroup of righteous people,” and “education in reli-gious sciences” (ta{lºm-i {ul¢m-i dºniyye).67

The coupling of a madrasa with a convent in thearchitectural program should be evaluated against thebackground of the close and relatively unproblematicrelation of the Cemali-Halvetis to the {ilmiyye. We knowwith certainty of only one other dervish convent built

in a mosque-madrasa complex in Istanbul after theKoca Mustafa Pasha convent and before that of Kadær-ga: the convent in the Atik {Ali Pasha complex, en-dowed in 1509 by the grand vizier Hadæm {Ali Pasha(d. 1511).68 {Ali Pasha’s convent, now demolished, wasbuilt for Kasæm Çelebi (d. ca. 1520), who as previouslynoted was a relative and disciple of Çelebi Halife andthe source of the Cemali-Halveti branch.69 KasæmÇelebi sent several disciples to the Balkans; one of themwas Sofyalæ Bali, the shaykh of both Nureddinzade andKurd Efendi.

The above-mentioned debate regarding the permis-sibility of Sufi rituals seems to have complicated theinclusion of Sufi convents in mosque-madrasa com-plexes during the reigns of Selim I (1512–20), Süley-man (1520–66), and Selim II (1566–74). Selim I andthe ×eyÒülisl¸m Kemalpa×azade (d. 1534) appear in laterHalveti tradition as enemies of the Halvetis.70 We knowthrough piecemeal evidence that even then someHalveti convents were supported by the ruling elite,although they were hardly included in mosque-madrasacomplexes in Istanbul. The Cemali grand vizier PiriMehmed Pasha endowed two convents in Istanbul. Oneof them was built for his uncle, the above-mentionedshaykh Ishak Karamani.71 Süleyman’s mother HafsaSultan (d. 1533) endowed a convent in 1522 as partof her building complex in Manisa (which also in-cluded a madrasa), and left the choice of a shaykh toSünbül Sinan, who installed his disciple Merkez Efendithere.72 Selim I’s daughter Øah Sultan, who was at-tached (and, according to some, married) to MerkezEfendi, endowed two convents in Istanbul (in Eyüpand Yenikapæ), with the condition in the endowmentdeed that both shaykhs should be disciples of MerkezEfendi.73 Hafsa Sultan’s building complex in Manisaand Øah Sultan’s in Eyüp were both attached to thepalaces of their patrons,74 like Sokollu and Ismihan’sin Kadærga. The patronage of Sufism thus seems to havebeen largely a private matter at that time.

Under Murad III (1574–95), himself a Halveti sym-pathizer, courtly patronage of Sufism spread to otherSufi orders as well. Among a number of Halveti con-vents endowed then, the convent in the Atik Validecomplex in Istanbul (1582–83), designed by Sinan, isnoteworthy for its relative monumentality. The com-plex was endowed by Nurbanu, who was Murad III’smother and Sokollu’s mother-in-law. The extant build-ings form a program similar to that of the Kadærgacomplex, consisting of a mosque, a madrasa, and aHalveti convent (figs. 14 and 15), although the Atik

muq-20-2.pmd 9/2/2003, 4:23 PM172


Fig. 14. Atik Valide complex, Istanbul. Site plan. 1. mosque 2. madrasa 3. dervish convent. (Müller-Wiener, Bildlexikon)

Valide complex also included additional buildingsaccording to Nurbanu’s endowment deed.75 The firstshaykh of the convent was Vi×ne Mehmed, a shaykhfrom a different branch of the Ùalvetiyye than wasNureddinzade. Also in the 1580s, the court officersØemsi Ahmed Pasha and Mehmed A¯a endowed aconvent each for a disciple of Vi×ne Mehmed, namelyYayaba×æzade Hæzær.76 Both shaykhs appear to have beenquite influential in palace circles of Murad III’s time.

In the seventeenth century, however, the relationsbetween a part of the ulema and the Halvetis becameuneasy because of the so-called Kadæzadeli movement.77

The theological stance underlying that assault was aSelefi one, which denounced any innovation (bid{a)

introduced into religious practice after the time ofthe Prophet and thus had an anti-Sufi direction. Thehistorical fact of the Kadæzadeli movement preventsus from readily assuming that dervishes and madrasastudents could peacefully live in a complex withouttension based on professional ambition or theologi-cal dispute, the two of which more often than notwent in tandem.

THE CONVENT VERSUS THE MADRASA INKADIRGA

A telling instance of the full-fledged bureaucratizationof the Halveti order in the seventeenth century is found

muq-20-2.pmd 9/2/2003, 4:23 PM173

zeynep yürekli.174

Fig.15. Atik Valide complex, Istanbul. Aerial view from the west: the mosque, the madrasa in the foreground, and the der-vish convent to the left. (Photo: A. Güler and M. Niksarlæ, from A. Kuran, Sinan the Grand Old Master of Ottoman Architecture[Washington, D.C., 1987])

in the diary of a Halveti dervish in Istanbul, who re-lates therein that his appointment as the shaykh of aconvent in 1665 had to be approved by the ×eyÒülisl¸m,probably because being the shaykh was almost auto-matically coupled with serving as Friday preacher ina mosque.78 As mentioned before, the endowmentdeed of Sokollu Mehmed Pasha stipulates such acombined post of shaykh and Friday preacher. Simi-larly, the endowment deed of Nurbanu states that the

shaykh of the convent in the Atik Valide complex ischarged with preaching in the mosque on all sacreddays.79 The shaykh chosen for the convent was there-fore expected to be “a preacher and adviser to theworshippers.”80

In the second half of the sixteenth century, theemployment of Halveti preachers was not restrictedto the mosques beside their convents. NumerousHalveti shaykhs were appointed to the most prestigious

muq-20-2.pmd 9/2/2003, 4:24 PM174


Friday preacher positions in major mosques of Istanbul,including the mosques of Süleymaniye and MehmedII, the two highest-ranking mosque-madrasa complexesin the Ottoman {ilmiyye. {Ata}i mentions in passing thatit was common knowledge that the position in MehmedII’s mosque was held by “saints who have completedtheir spiritual development” («arº_den f¸rig. olan e{izze),that is, shaykhs.81 In the relative freedom enjoyedby Sufis under Murad III, some Halveti preachersalso combined Friday sermons with dhikr rituals inmosques. Müniri-i Belgradi, for example, tells of afriend who went to listen to the Friday sermon byYayaba×æzade in the mosque of Sütlüce, which, as theburial place of the aforementioned proponent of dhikr,Ishak Karamani, must have been a popular Halvetisite.82 But when Yayaba×æzade started a dhikr sessionwith his dervishes before the sermon, Müniri’s friendsaid to himself in disappointment: “I came to listento this holy man. He could have had the dhikr ses-sion in his convent.”83

The Friday posts in the major mosques of Istanbulwere highly prestigious, enabling their holders tocommunicate their ideas to the masses and to adver-tise the principles of the theological stance they ad-vocated. In the 1630s, the pulpits of these mosquesbecame the loci for a theological dispute generatedby Kadæzade Mehmed, the initiator of the Kadæzadelimovement, and his major adversary, the Halveti shaykhAbdülmecid Sivasi. According to Madeline Zilfi’sstudy (see note 77), of all the appointments to Fridaypreacher posts in the five major mosques in Istanbulbetween 1621 and 1685, a great majority was of Halvetishaykhs. It is no surprise, then, that Halveti conventswere the major targets for Kadæzadeli attacks duringthat time. Much as this was based on a theologicaldispute, it also incorporated professional competition,as a group of the ulema who considered themselves“true preachers of the preacher path («arº_-i v¸{i¬¸n)were forced to share their mosques with Sufis turnedpart-time preachers—and this at a time when thenumber of mosques in the city was not substantiallyincreasing.”84

From the late sixteenth century on, not only theFriday preacher posts but also the entire {ilmiyye nolonger provided a substantially flourishing career pool.Under Murad III (r. 1574–95), when the expansionof imperial territories came to a relative halt, thenumber of madrasa graduates seems to have exceededthe number of available {ilmiyye positions. In the sec-ond half of the sixteenth century, madrasa students

repeatedly engaged in uprisings as well as relativelypeaceful demonstrations complaining of their situa-tion.85

If there was such frustration on the part of themadrasa students in the Kadærga complex as well, onewonders how they perceived their neighbors, the der-vishes with whom they shared the mosque and theshaykh who regularly preached to them. The modestarchitecture of the Sufi convent in comparison to themadrasa belies the fact that dervishes were probablythe majority among the inhabitants of the entire com-plex. On the assumption that one room was meant toaccommodate one person, the madrasa could houseseventeen people, including the müderris and his as-sistant, whereas the convent accommodated up to thirtydervishes in addition to the shaykh.86

Madrasa professors rotated frequently, whereasbeing the shaykh of a convent was usually a lifelongposition. Within the first six years of existence of themadrasa in Kadærga, four different people were ap-pointed as its müderris, with a maximum duration oftwo years,87 while Kurd Mehmed Efendi stayed thereas shaykh for thirteen.88 As opposed to the insecuritythat haunted professors’ thoughts about their futurecareers, the shaykh, with his stable position and hisregular sermons in the mosque, enjoyed the status ofa host. The students at the madrasa were also likelyto stay there for a much shorter time than did thedervishes. The {ilmiyye system encouraged madrasastudents, like their professors, to rotate to madrasasof increasing rank,89 whereas submission to a Sufishaykh was oftentimes a lifelong commitment.

Compared to dervishes, the students and their pro-fessor at the madrasa did not have much to complainabout in terms of stipend. The endowment deed allo-cates fifty aspers per diem for the müderris, twenty-fivefor the shaykh, five per diem for each madrasa student,and only three for a dervish.90 Shaykhs were occasion-ally granted large amounts of money by grandees whosought their spiritual guidance, however, and theemployment of a shaykh by multiple endowmentswas also not uncommon. Kurd Efendi, for example,also preached in the mosque of Mehmed II, receivingan additional salary from that endowment.91 Fur-thermore, the totals of salaries allocated in the endow-ment deed of Sokollu Mehmed Pasha for conventmembers (the shaykh and the dervishes) and formadrasa members (the professor, his assistant, and thestudents)—again assuming one person per room—were more or less the same. To the endower, the Sufi

muq-20-2.pmd 9/2/2003, 4:24 PM175

zeynep yürekli.176

convent thus seems to have been at least as importantas the madrasa.

THE RELATION OF THE CONVENT TOIMPERIAL ARCHITECTURAL IDIOMS

The Sufi convent in Kadærga was designed in a man-ner more modest in scale and architectural refinementthan the rest of the complex. The main architecturalguidelines that contribute to the monumental appear-ance of the mosque-madrasa in Kadærga—namely sym-metry, leveling of the ground, and elevation from thenatural topography—were not adopted in the adja-cent convent. On the contrary, the convent seemsalmost hidden behind the mosque and within theexisting topography (fig. 16).

In that respect, the Kadærga convent is also quitedifferent from that in the Atik Valide complex. Thelatter consists of dervish rooms and a ritual spacebehind a continuous portico surrounding a leveledcourtyard (fig. 17).92 The madrasa-like character ofthe Atik Valide convent, with its courtyard and ashlarmasonry and a ritual room that is hard to distinguishfrom a madrasa lecture room, becomes yet more evi-dent when the structure is compared to the madrasain the same complex (see fig. 15). Abandoned andstripped of its decoration today, the convent wouldhardly have been identifiable without the eighteenth-century inscription placed above the gate into the ritualroom.93

Since the architect apparently knew from the be-ginning who the user of the Kadærga convent wouldbe, Nureddinzade himself may have guided the de-sign process to some extent. He is recorded to havebeen “free of ostentation in his food and clothing”(ta{¸m ve lib¸sda tekellüfden {¸rº).94 His fellow, KurdEfendi, also favored avoiding unnecessary preoccupa-tion with outer appearance. In a treatise delineatingthe etiquette of dervishes, he stated that he consid-ered it unacceptable for a convent (his own, we mightassume) to house a dervish who washes his clothesunless they are canonically unclean, or who brusheshis beard.95 In accord with these ideas, the outer wallsof the convent lack even the simplest decoration.

Yet the fact that the inside of the tev¥ºdÒ¸ne waslavishly decorated makes it doubtful that the conventwas designed solely on the basis of the shaykhs’ exal-tation of modesty. Possibly Sokollu did not want toreveal the amount of money he had spent on a der-vish convent, and by keeping its outer appearance

modest in scale, embellishment, and monumentality,he would also have helped his shaykh preserve a repu-tation of modesty.

The contrast between modest outside and lavishinterior could also be read in terms of the inward-oriented nature of the Halveti “course of conduct” (seyr-i sül¢k). The entrance wall is a plain surface interruptedsolely by the single door into the convent and fourirregular windows, two of which seem to be later ad-ditions (fig. 18). The west wall is enlivened only byfour small shops on the lower level and a series of sevenwindows on the upper (fig. 19). Six of the dervish cellsand the shaykh’s quarters excepted, the rooms of theconvent have windows that look only towards the court-yard. In contrast to the madrasa, where all rooms butone have windows with views to the outside (and allbut the two corner rooms also have windows facing

Fig. 16. Sokollu Mehmed Pasha complex in Kadærga, Istanbul.Aerial view from the north: the mosque, the madrasa in theforeground, and the dervish convent behind the mosque.(Photo: A. Güler and M. Niksarlæ, from Kuran, Sinan)

muq-20-2.pmd 9/2/2003, 4:24 PM176


Fig. 17. Aerial view of the convent in the Atik Valide complex from the north. (Photo: VGMA; from Tanman, “Tekkeler”)

Fig. 18. The entrance wall and the south corner of the convent in Kadærga. (Photo: K. Görkay)

muq-20-2.pmd 9/2/2003, 4:24 PM177

zeynep yürekli.178

the courtyard), the convent compound is a secludedplace. In addition, the madrasa shares its courtyardwith the mosque, while the convent courtyard is re-served for the brotherhood. The inward-oriented ar-chitecture and embellishment of the convent—itssparseness of windows and lack of exterior decoration,which we also see in the Atik Valide convent—is inkeeping with the idea of seclusion from the outer world(Òalvet), so crucial an element in the Halveti courseof conduct that the order derives its name from thatconcept.

Nevertheless, the difference between the conventand the rest of the complex is not only in its modestand secluded appearance. As opposed to the presetdesign paradigm reflected by canonical mosque-ma-drasa, the architectural arrangement of the convent

relies on the topography of the site. For the construc-tion of the mosque-madrasa, natural topography wasoverpowered by leveling the construction ground,which is not the case with the convent. Neither dothe surrounding streets seem to have been dictatedby the boundaries of the convent building. On thecontrary, the design of the convent appears to havebeen entirely determined by natural topography andits outline defined by existing boundaries.

The distorted symmetry one notices in the con-vent plan is even more pronounced in the three-dimensional reality of the building; to the mind ofa sixteenth-century Ottoman court architect, whoseknowledge of geometry was by professional defini-tion very elaborate, it wouldn’t have constituted sym-metry at all.96 The southwest part of the building restson lower ground, as dictated by natural topography,and is therefore designed as a two-story building sur-rounding a lower courtyard accessible by a stairwayrunning behind the tev¥ºdÒ¸ne (figs. 20 and 21). It isnot the asymmetry per se but the preference for to-pographical determination over symmetry that reflectsa different approach to architectural design from theone utilized in the mosque-madrasa.

While the mosque-madrasa courtyard imposed anidiom that reflected the empire-wide institution of“madrasa” in its conceptual sense, the convent court-yard was reserved for the brotherhood; it was definedby the topography of its site and is closer to residen-tial architecture in scale and outlook. The architec-tural difference is due to the different ends that thetwo institutions served: the madrasa was designed foran unknown series of professors in rotation, and forstudents educated under an imperial managerial sys-tem for the main purpose of becoming members ofthe religio-legal cadre of that system, while the con-vent was created for a particular user, in this caseNureddinzade.

As much as the architectural program of the Kadærgacomplex suggests a convergence of the Ùalvetiyyewith the {ilmiyye, it also echoes their segregation, inthat the convent is assigned an autonomous space inthe mosque-madrasa complex. Likewise, seen againstthe architecture of the adjacent madrasa, that of theconvent building also suggests both convergenceand segregation. Although the convent buildingdeviates from the classical Ottoman idiom for reli-gious architecture in terms of the basic design crite-ria employed, its architecture does incorporate someaspects of that idiom, especially in the mosque-like

Fig. 19. The west wall of the convent in Kadærga. (Photo: K.Görkay)

muq-20-2.pmd 9/2/2003, 4:24 PM178


Fig. 20. The southwest portico of the lower courtyard. (Photo: Archive of the Mülga Eski Eserler Encümeni, Istanbul Ar-chaeological Museum)

Fig. 21. View from the lower courtyard towards the staircase leading to the upper courtyard. The protruding wooden con-struction on the left is probably a later addition. (Photo: Archive of the Mülga Eski Eserler Encümeni, Istanbul Archaeologi-cal Museum)

muq-20-2.pmd 9/2/2003, 4:24 PM179

zeynep yürekli.180

interior decoration of the tev¥ºdÒ¸ne.The convent building thus refers to the imperial

idiom in religious architecture, but the reference isimprecise. The aforementioned architectural differ-ences between the convent and the madrasa, as wellas the differences between the Kadærga convent andthe Atik Valide convent, might indicate that we shouldconsider the convent in Kadærga as much an exten-sion of the palace of Sokollu Mehmed Pasha as a partof the mosque-madrasa complex. Sokollu MehmedPasha’s relation to the madrasa as an architecturalpatron was primarily in his administrative persona asthe grand vizier: he endowed a madrasa for the em-pire. His patronage of the Halveti convent, however,was of a personal nature: he endowed this conventspecifically for his shaykh, with the continuation andenhancement of a personal relationship in mind.

The give-and-take between Sufi ideal and imperialimage that one can observe in the architecture of thedervish convent is striking. Ideally, a Sufi would seekknowledge by experiencing his own mental state (¥¸l),and not what scholarship told him (_¸l). Thus, again

ideally, every Sufi pursued his own singular path andsought his own truth. The Kadærga convent is architec-turally singular as a result of its configuration in accordwith the topography of the site, much like many otherSufi convents. Yet on the other hand, it was built at atime when the Halveti order was highly institutional-ized and bureaucratized, when the singularity of thespiritual paths of Halveti dervishes was mixed with theirrole in imperial politics or their positions in the {ilmiyyeas Friday preachers. Particularly the disciples of SofyalæBali, who was praised by Ta×köprizade for being “a manof _¸l and a man of ¥¸l at the same time” (hem ehl-i _¸lhem ehl-i ¥¸l),97 were trained to partake of Ottomanreligious scholarship, which they did.98 Accordingly,the architecture of the convent combines imperialfeatures with the topography of the site. It incor-porates classical architectural elements—a domedritual space decorated as a mosque, a spacious colum-nar anteroom covered by cross-vaults (fig. 22), regu-lar bands of stone and brick masonry on all visibleouter surfaces, stone as the prevailing surface mate-rial on the inside, and lead-covered roofs (fig. 23)—

Fig. 22. The northwestern bay of the columnar anteroom with the tev¥ºdÒ¸ne on the left. (Photo: Z. Yürekli)

muq-20-2.pmd 9/2/2003, 4:24 PM180


Fig. 23. Aerial view of the convent from the southeast, show-ing the dome over the entrance hall, the cross vaults overthe columnar anteroom, and the dome over the tev¥ºdÒ¸ne,all covered with lead. (Photo: K. Görkay)

in what would otherwise have been a vernacular ar-chitectural rendering of local topography.

Given its proximity to the palace of Sokollu andIsmihan (see fig. 2), perhaps the convent building wasarchitecturally conceived as part of their extendedhousehold, which would explain its apparent adher-ence to a palatial idiom. The interconnected asym-metrical courtyards, seclusion from the outer world,limited use of domes to highlight only the entranceaxis and the ritual space (unlike the madrasa, wheredomes are used in each unit), and lack of the basicelements that contribute to monumentality in Otto-man religious architecture all recall the harem sec-tion of the Topkapæ Palace. In Ottoman and Persian

literature, it was common practice to refer to the con-vents of esteemed shaykhs and to palaces with the samewords (such as derg¸h and ¸sit¸ne). Here, this meta-phoric association finds architectural expression in theconvent’s allusions to a palatial idiom.

CONCLUSION: THE EMPIRE AND THE SELF

The Kadærga complex is one of the very few buildingcomplexes in sixteenth-century Istanbul that combinea Sufi convent with a madrasa. In the sixteenth-cen-tury Ottoman context, madrasas were places where acrucial sector of imperial policy operators, namelythe {ilmiyye class, was educated. Sunni Islamic thoughttraditionally did not separate {ilm (literally, “knowl-edge”: a word that had religious connotations untilsecularist thinking weakened them) from politics any-way. The Sunni Sufi convents that established closerelations with the {ilmiyye, on the other hand, servedin part as an alternative for those who sought a formof {ilm other than that taught in the madrasa, or wholacked access to the institution. Thus some Sufi or-ders particularly served the spiritual, intellectual, andpolitical interests of those who were denied such ac-cess; the case of women is perhaps most easily no-ticed. Interestingly, other Ottomans denied madrasaeducation included the greater part of the courtlyelite who occupied the highest-ranking offices: slavesoldiers including grand viziers like Sokollu MehmedPasha, as well as members of the inner court, includ-ing the sultans and their first-degree relatives.

Sunni Sufism was not necessarily isolated, eithertheologically or politically, from the {ilmiyye, as therehad never been a clear-cut distinction between them.The fact that part of the sixteenth-century ulema criti-cized Sufi rituals and part of the seventeenth-centuryulema did not want to see Sufis in the pulpits doesnot mean that we should conceive of Sufism and the{ilmiyye as two different universes, since this wouldignore the many members of the {ilmiyye who attendedSufi convents. Some convents also provided alterna-tive {ilmº communities for many “insiders” of themadrasa. But the debate between a part of the {ilmiyyeand the Sufis in the sixteenth century did complicatethe inclusion of Sufi convents in mosque-madrasacomplexes in Istanbul. Until Murad III’s time, excep-tions were made only for Sufi shaykhs recognized andtrusted for their commitment to the shari{a.

The architectural program of the Kadærga complexis one such exception, bearing witness to the conver-

muq-20-2.pmd 9/2/2003, 4:24 PM181

zeynep yürekli.182

gence of the Cemali-Halveti community with the Ot-toman {ilmiyye: a convergence we have observed in theideological support that some Cemali-Halveti shaykhs,such as Nureddinzade, lent to Ebussu{ud, or in thevery high number of Cemali-Halveti shaykhs, again suchas Nureddinzade, who had madrasa educations, com-posed scholarly treatises, and held Friday preacherposts. The architecture of Sokollu’s dervish conventin and of itself reflects the convergence of Sufism withimperial politics by combining elements of both.

Certainly, we are not dealing only with imperialpolitics. After decades of scholarship assuming thatthe idea of “the self” is an idiosyncrasy of Westernthought, Cemal Kafadar relatively recently made usaware that “the self” is actually not missing from theintellectual life of the Ottomans.99 Among other things,he drew attention to the capacity of dervish circles tooffer an environment where individuals could copewith their anxieties. Accordingly, the Sufi method ofaspiring to personal development seems to have of-fered many soldiers, officers, and courtiers a retreatfrom the administrative and economic afflictions ofthe second half of the sixteenth century. The hybridnature of the Kadærga convent, with its simultaneousarchitectural references to public-religious and private-residential idioms in Ottoman architecture, reflectsthe capacity of a particular Sufi circle to infiltrate si-multaneously the public realm of the empire and theprivate realm of the ruling elite.

Harvard UniversityCambridge, Massachusetts

NOTES

Author’s note: This article is indebted to the invaluable help andconstructive criticism of Gülru Necipo¯lu and Cemal Kafadar atevery stage, and to their unconditional willingness to share theirimmense knowledge of Ottoman primary sources with their stu-dents, of whom I am very happy to be one. This article devel-oped from my Ph.D. Qualifying Paper, which itself grew out ofthe research paper I wrote for Prof. Necipo¯lu’s graduate stu-dent seminar on Ottoman architecture at Harvard University in1998. A shorter version of that paper was then submitted in August1999 at the Eleventh International Congress of Turkish Art inUtrecht, Netherlands, where I benefited from the insightfulcomments of Machiel Kiel, Zeynep Ahunbey, and Baha Tanman.The present version of the article, however, would not have beenpossible without dissertation research fellowships from the KressFoundation and the American Research Institute in Turkey in2000–2001, which made possible my further research in the manu-script libraries in Istanbul and in the Archives of the Director-ate of the Pious Endowments (hereafter VGMA) in Ankara. I

would also like to thank Derin Terzio¯lu and Aslæ Niyazio¯lu fordirecting me to some of the primary and secondary sources atthat stage, David Roxburgh, Selim Kuru, and Jeff Spurr for readingand commenting on an earlier version of the article, and BahaTanman for providing me with old photographs of the build-ing. Tanman has also generously permitted me to redraw hisdrawings with minor amendments for publication in this article.Throughout this article, the term “Cemali” refers to a family, whilethe term “Cemali-Halveti” refers to a particular Sufi lineage thatis explained in the article.

1. See Do¯an Kuban, “An Ottoman Building Complex of theSixteenth Century: The Sokollu Mosque and its Dependen-cies in Istanbul,” Ars Orientalis 7 (1968): 19–39.

2. Its first müderris, A¯azade Mehmed Efendi, was appointedin 1570 and taught there until 1572. Cahid Baltacæ, XV.–XVI.Asærlarda Osmanlæ Medreseleri (Istanbul, 1976), pp. 422–23 andp. 590.

3. Tülay Artan, “In the Tracks of a Lost Palace,” Ninth Interna-tional Congress of Turkish Art (Istanbul, 1991), pp. 197–202;idem, “The Kadærga Palace: An Architectural Reconstruction,”Muqarnas 10 (1993): 201–11; idem, “The Kadærga PalaceShrouded by the Mists of Time,” Turcica 16 (1994).

4. The waqfiyya of Sokollu Mehmed Pasha, ms. VGMA, Defter572, pp. 27–63 (hereafter Waqfiyya), p. 32.

5. Waqfiyya, p. 33. See also Tülay Artan, “Sokollu Mehmed PashaSarayæ,” ~stanbul Ansiklopedisi.

6. In the 1580s, the palace was inhabited by Sokollu andIsmihan’s son, Ibrahim Han; see Tülay Artan, “Esma SultanSahilsarayæ,” ~stanbul Ansiklopedisi. Ibrahim Han is buried inthe cemetery of the Kadærga complex; see J. Bacqué-Gram-mont, H. Laqueur, N. Vatin, Stelae Turcicae: Cimetières de lamosquée de Sokollu Mehmed Pa×a à Kadirga Limani, de BostanciAli, et du türbe de Sokollu Mehmed Pa×a à Eyüb (Tübingen, ca.1990).

7. Baha Tanman, “Sinan}æn Mimarisi: Tekkeler,” Mimarba×æ KocaSinan: Ya×adæ¯æ Ça¯ ve Eserleri, ed. Sadi Bayram, vol. 1 (Istan-bul, 1988), pp. 311–32, on pp. 313–19.

8. See the list of buildings in Tu¥fetül-Mi{m¸rºn; Zeki Sönmez,Mimar Sinan ile ~lgili Tarihi Yazmalar, Belgeler (Istanbul, 1988).

9. Waqfiyya, ms. VGMA, Defter 572, pp. 27–63. This endow-ment deed seems to have been the main source for anothertext listing “the buildings of the public foundations of thedeceased Mehmed Pasha”: Ebniye-i Ùayr¸t ü Ýasen¸t, ms. FatihMillet Library, {Ali Emiri Tarih 933, fols. 18v–23v (hereaf-ter Ebniye). The patronage of Sokollu Mehmed Pasha willbe examined in detail in a forthcoming book on Ottomanarchitectural culture in the age of Sinan by Gülru Necipo¯lu,who has brought both texts (Waqfiyya and Ebniye) to myattention. Ebniye was included in a compilation of texts re-lated to endowments, apparently meant as a handbook forauthors of endowment deeds. Considering it unnecessaryto copy the extremely long document in its entirety, thecompiler seems to have reworked it as a text of reference.Ebniye, either from public knowledge or from other endow-ment documents, occasionally adds details missing from theoriginal endowment deed, but it omits other parts, especiallythose that list income-generating facilities and conditionspertaining to the employees of the endowed public institu-tions.

muq-20-2.pmd 9/29/2003, 12:35 PM182


10. Waqfiyya, p. 29 (location), p. 55 (conditions). Ebniye curi-ously omits the building complex in Szigetvár.

11. Waqfiyya, p. 29 (location), p. 58 (conditions). While theoriginal endowment deed clearly identifies this convent asa Sufi lodge, and separately lists the conditions of a hospice({im¸ret) in that complex, Ebniye (fol. 19v) mentions only aconvent (z¸viye) with the primary function of providing foodfor travelers, a function ascribed to the hospice in the originalendowment deed.

12. Nev{izade {Ata}i, Ùad¸}i_u’l-Ýa_¸}i_ fº Tekmºleti×-Øa_¸yi_,Øakayi_-i Nu{m¸niye ve Zeyilleri, vol. 2 (Istanbul, 1989), pp.212–14 and p. 323. See also Hafæz Hüseyin Ayvansarayi,Ùadº_atül-Cev¸mi{ (Istanbul, 1304/1886–87), p. 194.

13. Müniri-i Belgradi, Silsiletül-Mu_arribºn ve Men¸_æbül-Mütta_ºn,ms. Süleymaniye Library, Øehid Ali Pa×a 2819/3, fol. 113v.

14. V. Minorsky, “Shaykh Bali-Efendi on the Safavids,” Bulletinof the School of Oriental and African Studies 20 (1957): 437–51.

15. Müniri, Silsiletül-Mu_arribºn, fol. 119v.16. See Nathalie Clayer, Mystiques, état et société: Les Halvetis dans

l’aire balkanique de la fin du XVe siècle à nos jours (Leiden, 1994),pp. 79–80.

17. According to Müniri (Silsiletül-Mu_arribºn, fol. 113r), Nured-dinzade went to Istanbul to defend himself, while accord-ing to {Ata}i (Ùad¸}i_, p. 212) it was also to defend his shaykh.Nureddinzade heard of the investigation in a timely man-ner and consulted a Zeyni shaykh in Sofia on what to do.The shaykh, {Ali b. Sinan, advised Nureddinzade to go toIstanbul and defend himself, assuring him that he wouldsurely end up as the director of a convent there: {Ata}i,Ùad¸}i_, p. 80.

18. {Ata}i, Ùad¸}i_, pp. 212–13.19. {Ata}i, Ùad¸}i_, p. 213. For the endowment from 1507, see

Ö. L. Barkan and E. H. Ayverdi (eds.), 953 (1546) TarihliIstanbul Vakæflaræ Tahrir Defteri (Istanbul, 1970), pp. 16–17;Semavi Eyice, “Kapu A¯asæ Hüseyin A¯a’næn Vakæflaræ,” AtatürkÜniversitesi Edebiyat Fakültesi Ara×tærma Dergisi 9 (1978): 170–85. On the convent, see Baha Tanman, “Küçük Ayasofya Tek-kesi,” ~stanbul Ansiklopedisi.

20. Ta×köprizade relates in Øa_¸}i_-i Nu{m¸niye that SultanSüleyman granted an earlier shaykh of the Küçük Ayasofyaconvent, {Abdülkerim Kadiri (d. 1544–45), the permissionto issue fatwas like a ×eyÒülisl¸m (×eyÒülisl¸m gibi Òal_a fetv¸virmeye ruÒªat buyurdælar), and receive a salary for this. Øa_¸}i_-i Nu{m¸niye ve Zeyilleri, vol. 1 (Istanbul, 1989), pp. 517–18.

21. For example, the Bayrami shaykh {Osman, investigated onaccount of his spiritual descent from a shaykh, namely HamzaBali, who had been executed on charges of heresy, appliedto Nureddinzade, became his dervish, and thus survived theinvestigation. {Ata}i, Ùad¸}i_, p. 463. According to Müniri,the execution of Hamza Bali had been put forward byNureddinzade; Silsiletül-Mu_arribºn, fol. 114r.

22. Ata}i, Ùad¸}i_, pp. 595–96. Upon Nureddinzade’s appoint-ment, his disciple Mehmed (d. 1604–5), described by {Ata}ias a shaykh with perfect attention to the shari{a, took chargeof this new convent endowed by the court herald (çavu×){Ali.

23. {Ata}i, Ùad¸}i_, pp. 151–52. The disciple is Nureddin (d.1571–72 or 1573–74) and the vizier is {Ali Pasha, possiblySemiz {Ali Pasha (d. 1565), who was a vizier from 1554 un-til 1561, when he became the grand vizier.

24. Müniri, Silsiletül-Mu_arribºn, 114a. Müniri always uses the term“land of Rum” (R¢m diy¸ræ or diy¸r-i R¢m) for the Europeanpart of the empire, i.e., Rumili; this is unlike some otherOttoman sources, where this term may denote Anatolia orall the Ottoman lands.

25. E.g., {Ali Kemali of Tærhala, the shaykh of the convent ofPiri Mehmed Pasha between 993 and 1012, who preachedin the Ayasofya mosque and became known as Preacher ofAyasofya ({Ata}i, Ùad¸}i_, p. 470); Ibrahim of Crimea, shaykhof the Küçük Ayasofya convent following Nureddinzade, whopreached in several sultanic mosques in Istanbul ({Ata}i,Ùad¸}i_, p. 370); Mehmed, mentioned in note 22 above, whopreached in the {Osman×ah mosque in Tærhala ({Ata}i,Ùad¸}i_, pp. 595–96); Mehmed of Crimea, described by {Ata}ias an influential and priceless preacher in Varna (Ùad¸}i_,p. 599); and Hasan Nefa}isi, who preached in the KethüdaKadæn mosque in Istanbul ({Ata}i, Ùad¸}i_, p. 600).

26. {Ata}i, Ùad¸}i_, p. 213.27. According to Müniri, Silsiletül-Mu_arribºn, the choice of

Nureddinzade as the army’s shaykh was Süleyman’s: Silsiletül-Mu_arribºn, fol. 113v. A number of accounts of Halveti shaykhsjoining in the Ottoman campaigns in Rumili are recountedby Müniri: fols. 78v ff. See also Clayer, Mystiques, état, et société,pp. 113–41.

28. {Ata}i, Ùad¸}i_, p. 100.29. Waqfiyya, p. 28.30. {Ali Efendi (d.1006/1597–98), a disciple of Nureddinzade,

“became famous with the title of convent-keeper (z¸vºye-d¸r)and shaykh of the tomb (türbe ×eyÒi) at the tomb and con-vent (Ò¸n_¸h) built for Sultan Süleyman in Szigetvár.” ({Ata}i,Ùad¸}i_, pp. 465–66) {Ali Efendi was appointed there bySokollu, according to Müniri, Silsiletül-Mu_arribºn, fol. 96r.{Ali Efendi was the author of several religious works that {Ata}ifound “totally accepted, respected, and worthy of praise.”

31. Evliya Çelebi b. Dervi× Mehemmed Zælli, Evliya Çelebi Seya-hatnamesi, III. Kitap (Istanbul, 1999), p. 31; ms. TopkapæPalace Library, Ba¯dad 305, III, fol. 20r.

32. Waqfiyya p. 58: Òan_¸h-i mezb¢r medreseye tebdºl olunmayupd¸}im¸ bu va¾{æ üzere isti{m¸l olæna.

33. Nureddinzade died on 22 Awa}il Dhº l-Ka{da 981 (March 16,1574) according to Müniri, Silsiletül-Mu_arribºn, fol. 114r. Theendowment deed was activated on aw¸}il Dhº l-Ýijja 981(April13–22, 1574) according to Waqfiyya, p. 63.

34. {Ata}i, Ùad¸}i_, p. 313. Kurd Efendi was the first shaykh ofthe convent according also to Zakir Øükri Efendi, ~stanbulÙ¸n_¸hlaræ Me×¸yiÒi, ed. Turgut Kut (Cambridge, 1995), fol.7r; ms. Atatürk Kitaplæ¯æ K.75, also published by M.S. Tay×iand K. Kreiser as Die Istanbuler Derwisch-Konvente und ihreScheiche: Mecmua-i Tekaya (Freiburg, 1980).

35. For their biographies, and on the Cemali family, see YusufKüçükda¯, II.Bayezid, Yavuz ve Kanuni Devirlerinde Cemali Ailesi(Istanbul, 1995). Ottoman sources vary on the nature of therelation of Piri Mehmed Pasha to Çelebi Halife. WhileKüçükda¯ argued on the basis of several sources that PiriMehmed was Celebi Halife’s son, the sixteenth-centuryHalveti shaykh Yusuf Sinan, writing in 1581, presents bothZenbilli {Ali and Piri Mehmed as Çelebi Halife’s uncles; YusufSinan, Tezkire-i Ùalvetiyye, ms. Süleymaniye Library, EsadEfendi 1372–2, fol. 9v. The text was published as: Men¸_æb-

muq-20-2.pmd 9/2/2003, 4:24 PM183

zeynep yürekli.184

i Øerºf ve Þarº_atn¸me-i Pºr¸n ve Me×¸yiÒ-i Þarº_at-i {Aliyye-iÙalvetiyye (Istanbul, 1290/1873–74).

36. Ebniye (fol. 20r) also describes the building as a convent withthirty rooms for dervishes; this information is missing fromWaqfiyya.

37. Waqfiyya, p. 50: Ò¸n_¸h yanændaki çmi{de va{¬ icün va¾{ itdügikürsºde ba{de ªalav¸til-Cum{a müslüm¸nlara va{¬ ü naªº¥at eyleye.Similar conditions applied to the shaykhs of the other twoconvents endowed by Sokollu. The shaykh of the Szigetvárconvent was charged with Friday sermons in the mosque,while the shaykh of the Payas convent was charged with twoweekday sermons. Waqfiyya, pp. 55, 58.

38. Waqfiyya, p. 50, states that in return for their salaries fromthe endowment, the dervishes should pray after each regu-lar prayer service for the spirits of the Prophet and the patronand for the continuation of the rule of the Ottoman sul-tan.

39. {Ata}i, Ùad¸}i_, pp. 213, 313.40. Waqfiyya, p. 29, states that the books endowed to the madra-

sas by the patron were to be listed together in the appen-dix of the endowment deed. This appendix is missing fromboth Waqfiyya and Ebniye. Probably because there was nomadrasa in the Szigetvár complex, books were given as en-dowments to the dervish convent and were also to be listedin the (missing) appendix: Waqfiyya, p. 55. There is nomention of books for the Payas convent.

41. Today the tev¥ºdÒ¸ne is used as the classroom of a Qur}anicschool. The walls and the dome of the tev¥ºdÒ¸ne were com-pletely whitewashed during a major restoration around 1975that converted the convent into the present school, and theynow bear no traces of original decoration. Most of the deco-ration seen in old photographs was dated stylistically to theoriginal date of building by Tanman, in “Tekkeler,” p. 318.

42. Mufti {Ali Çelebi, Õ¢ret-i Fetv¸, ms. Süleymaniye Library, EsadEfendi 1761, fols. 57v–58r.

43. Mufti {Ali Çelebi, Ris¸la fº l-Dhikri l-Ja¥rº, ms. SüleymaniyeLibrary, Esad Efendi 1761, fols. 54r–57r.

44. Sünbül Sinan, Ris¸la fº l-Dhikr, ms. Süleymaniye Library, EsadEfendi 1761, fols. 20v–27v. In Halveti traditions, the trea-tise is also known as Ris¸le-i Ta¥_º_iyye. For the nearest-to-contemporaneous biography of the author available, writ-ten in the late sixteenth century, see Yusuf Sinan, Tezkire-iÙalvetiyye, fols. 17v–24v.

45. Hulvi, Leme¬¸t, ms. Süleymaniye Library, Ali Emiri Øeriye(Millet) 1100, fol. 162r. Modern Turkish translation of thetext has been published as: Mahmud Cemaleddin Hulvi, Leme-zat-i Hulviyye ez Lemezat-i Ulviyye, ed. Mehmet Serhan Tay×i(Istanbul, 1993).

46. Ta×köprizade, Øa_¸}i_, pp. 373–74.47. Hulvi, Leme¬¸t, fol. 161v.48. Yusuf Sinan, Tezkire-i Ùalvetiyye, fol. 22r–22v.49. On mosque-tev¥ºdÒ¸nes built by Sinan in Istanbul, see Tan-

man, “Tekkeler,” pp. 322–23.50. On Mehmed II’s imperial vision and how it informed his

architectural patronage in the new capital, see Gülru Neci-po¯lu, Architecture, Ceremonial, and Power: The Topkapæ Palacein the Fifteenth and Sixteenth Centuries (New York and Cam-bridge, 1991), pp. 3–30.

51. On this imperial classification of Ottoman madrasas and areclassification under Süleyman after the completion of the

madrasas in the Süleymaniye complex in 1557, see Cevat ~zgi,Osmanlæ Medreselerinde ~lim, vol. 1 (Istanbul, 1997), pp. 35–37.

52. On the political role of Çelebi Halife at this historical mo-ment, see Hans Joachim Kissling, “Aus der Geschichte desChalvetijje-Ordens,” Zeitschrift der Deutschen MorgenländischenGesellschaft 103 (1953): 233–89.

53. Yusuf Sinan, Tezkire-i Ùalvetiyye, fol. 11v.54. Hulvi, Leme¬¸t, fols. 153v–157v.55. For the endowment, see Barkan-Ayverdi, Tahrir Defteri, pp.

366–69. A compilation of documents related to the endow-ment is ms. VGMA, Defter 654/42.

56. According to Yusuf Sinan, Hacæ Mustafa asked Sultan Bayezidfor this monastery (_æzlar kilis¸sæ) in order to make an en-dowment for Çelebi Halife; Tezkire-i Ùalvetiyye, fol. 12v. Seealso Semavi Eyice, “~stanbul’da Koca Mustafa Pa×a Camii veOsmanlæ-Türk Mimarisindeki Yeri,” Türkiyat Dergisi 5, 8(1953): 152–182.

57. Yusuf Sinan, Tezkire-i Ùalvetiyye, fol. 12v; Hulvi, Leme¬¸t, fols.153v–157v. In 953 (1546), the complex was recorded as con-sisting of a mosque, a hospice, a madrasa, a dervish con-vent, and a primary school, with a bathhouse, shops, andgardens nearby; Barkan-Ayverdi, Tahrir Defteri, pp. 366–69.

58. Ahmet T. Karamustafa, God’s Unruly Friends: Dervish Groupsin the Islamic Later Middle Period 1200–1550 (Salt Lake City,1994), pp. 88–89.

59. On this subject, see also Ahmet Ya×ar Ocak, Osmanlæ Toplu-munda Zændæklar ve Mülhidler (15.–17. Yüzyællar) (Istanbul,1998).

60. Yusuf Sinan, Tezkire-i Ùalvetiyye, fol. 23r–23v; also quoted in{Ata}i, Ùad¸}i_, p. 360.

61. See fatwas in Ertu¯rul Düzda¯, Øeyhülislam Ebussuud Efendi’ninFetvalaræna Göre Kanuni Devrinde Osmanlæ Hayatæ (Istanbul,1998), pp. 131–39. On Ebussu{ud’s position within the dis-cussions among the sixteenth-century ulema, culminatingin the so-called Kadæzadeli conflict of the seventeenth cen-tury, see Derin Terzio¯lu, “Sufi and Dissident in the Otto-man Empire: Niy¸zº-i Mæsrº (1618–1694),” Ph. D. Diss.(Harvard University, 1999), pp. 220–33.

62. In addition to the fatwas in Düzda¯, see Ebussu{ud, Z$ ikr esn¸-sænda def ve s¸}ire çalmanæn ç}iz olup olmadæ¯æ, ms. Süleyma-niye Library, Yazma Ba¯æ×lar 2144, fols. 60v–61v.

63. Düzda¯, Øeyhülislam Ebussuud, p. 134, no. 347; Fetv¸ Mecmu{asæ,ms. Beyazæt Library 2757, fol. 323r.

64. See Clayer, Mystiques, état, et société, pp. 70–90.65. Hulvi, Leme¬¸t, fols.153v–154r.66. See references in note 25, above.67. {Ata}i, Ùad¸}i_, p. 213.68. The complex endowed in 1509 was recorded in 1546 as

consisting of a mosque, a madrasa, a dervish convent, anda hospice: Barkan-Ayverdi, Tahrir Defteri, pp. 67–71. Thedervish convent does not exist anymore. It was later namedafter its second shaykh, Karaba× Ramazan: see Zakir Øükri,~stanbul Ý¸n_¸hlaræ Me×¸yiÒi, fol. 29v.

69. Ta×köprizade, Øa_¸}i_, p. 375; Yusuf Sinan, Tezkire-i Ùalve-tiyye, fol. 16v; Hulvi, Leme¬¸t, fols. 157v–158r. See also Küçük-da¯, Cemali Ailesi, pp. 115–21.

70. Hulvi, Leme¬¸t, fols. 159v–163r.71. Yusuf Küçükda¯, Vezir-i Azam Piri Mehmed Pa×a (Konya, 1994),

pp. 134–39.

muq-20-2.pmd 9/2/2003, 4:24 PM184


72. Hulvi, Leme¬¸t, fols. 164v–168r; Yusuf Sinan, Tezkire-i Ùalvetiyye,fols. 24v–30v; Ç. Uluçay and I. Gökçen, Manisa Tarihi (Istan-bul 1939), p. 99.

73. The waqfiyya of Øah Sultan (1570), ms. VGMA, Defter 1993,pp. 11–17; Yusuf Sinan, Tezkire-i Ùalvetiyye, fols. 24v–30v, 33r–33v; Tahsin Yazæcæ, “Fetihten Sonra ~stanbul’da Ilk HalvetiØeyhleri: Celebi Muhammed Cemaleddin, Sünbül Sinan veMerkez Efendi,” ~stanbul Enstitüsü Dergisi 2 (1956): 87–113;E. Esin, “Merkez Efendi ile Øah Sultan Hakkænda bir Ha×iye,”Türkiyat Mecmuasæ 19 (1980): 65–92. For other courtly pa-trons of Halveti convents, see Clayer, Mystiques, état et société,p. 110, note 135.

74. Uluçay-Gökçen, Manisa Tarihi, p. 99; the waqfiyya of ØahSultan, p.14.

75. The waqfiyya of Nurbanu Sultan, ms. VGMA, Defter 138. Thewaqfiyya also lists a hospital, a hospice, a d¸r al-¥adºth, anda d¸r al-qurr¸}.

76. On the convent of Nurbanu and Vi×ne Mehmed, see {Ata}i,Ùad¸}i_, p. 361; Zakir Øükri, ~stanbul Ù¸n_¸hlaræ Me×¸yiÒi, fol.46r. The chief court architect Mehmed A¯a was also a discipleof Vi×ne Mehmed Efendi: Ca{fer Efendi, Ris¸le-i Mi{m¸riyye:An Early Seventeenth-Century Ottoman Treatise on Architecture,ed. Howard Crane (Leiden, 1987), fols. 10r–13r, 17v–18v.On Yayaba×æzade, see {Ata}i, Ùad¸}i_, pp. 464–65; Müniri, Silsi-letül-Mu_arribºn, fols. 95v–97r; Hulvi, Leme¬¸t, fols. 207v–208r.

77. Madeline C. Zilfi, “The Kadæzadelis: Discordant Revivalismin Seventeenth-Century Istanbul,” Journal of Near EasternStudies 45, 4 (1986): 251–69.

78. Cemal Kafadar, “Self and Others: The Diary of a Dervish inSeventeenth-Century Istanbul and First-Person Narratives inOttoman Literature,” Studia Islamica 69 (1989): 121–50, p.140.

79. Ms. VGMA, Defter 138, p. 126.80. Ms. VGMA, Defter 138, p. 124.81. {Ata}i, Ùad¸}i_, p. 372. The endowment deed of Mehmed II

does not, however, impose such a condition regarding theFriday sermon in the mosque. For a list of all known manu-scripts of the endowment deed and a facsimile of what isprobably the oldest one in Arabic (Türk ve Islam EserleriMüzesi 1872), see Osman Ergin, Fatih ~mareti Vakfiyesi (Istan-bul, 1945). For the facsimile of a later Turkish translation,see Fatih Mehmed II Vakfiyeleri (Ankara, 1938).

82. N. Esra Di×ören, “Çavu×ba×æ Camii,” ~stanbul Ansiklopedisi. Themosque is also known by the name of Mahmud A¯a, whohad the mosque rebuilt in 1538 by the architect Sinan.Ta×köprizade (Øa_¸}i_, p. 373) also states that Ishak Karamaniis buried in Sütlüce.

83. Silsiletül-Mu_arribºn, fol. 96v: Bu müb¸regi diñlemege geldüm.Zikrullahæ z¸viyesinde de iderdi.

84. Zilfi, “The Kadæzadelis,” p. 267.85. See Mustafa Akda¯, Celali isyanlari 1550–1603 (Ankara, 1963),

pp. 153–282. Akda¯ focuses on uprisings, while we also knowof relatively peaceful demonstrations, like the one stagedby lower-level madrasa students who failed to gain admis-sion to higher-level madrasas, during the circumcision fes-tival that Murad III organized for his son Mehmed in 1582:see Derin Terzio¯lu, “The Imperial Circumcision Festivalof 1582: An Interpretation,” Muqarnas 12 (1995): 84–100,p. 87.

86. Waqfiyya does not give the numbers of madrasa students anddervishes. Ebniye (fol. 50r) simply describes the conventbuilding as having thirty dervish cells (otuz hücreli bir z¸viye).

87. Baltacæ, Osmanlæ Medreseleri, pp. 422–23.88. Zakir Øükri, ~stanbul Ù¸n_¸hlaræ Me×¸yiÒi, fol. 7r; {Ata}i,

Ùad¸}i_, p. 313.89. On a document delineating the rotation system in 1576, see

Izgi, Osmanlæ Medreselerinde ~lim, p. 59.90. Waqfiyya, p. 50. Ebniye (fol. 20r) only mentions three aspers

per diem for each of the thirty rooms in the convent.91. {Ata}i, Ùad¸}i_, p. 372.92. See Tanman, “Tekkeler,” pp. 319–22.93. Tanman, “Tekkeler,” p. 312.94. {Ata}i, Ùad¸}i_, p. 213.95. Kurd Efendi, Ris¸le-i ûrd Efendi, ms. Süleymaniye Library,

Izmir 795/7, fol. 12r.96. In 1570, abstract geometry was apparently the core of ar-

chitectural theory in the palace workshops. The biographyof the chief court architect, Mehmed A¯a, relates that hewas drawn to the craft of mother-of-pearl inlay when he heardan apprentice read aloud a book on geometry in the pal-ace gardens. Ca{fer Efendi, Ris¸le-i Mi{m¸riyye, fols. 13r–17r.The fact that he took up the arts of architecture and mother-of-pearl inlay together, the latter being a craft that requireda highly developed skill in geometrical abstraction, suggeststhat these two crafts had a common theoretical basis.

97. Ta×köprizade, Øa_¸}i_, p. 522.98. For the religious scholarly writings of Nureddinzade and Kurd

Efendi not mentioned in this article, see {Ata}i, Ùad¸}i_, pp.214 and 323 respectively. The masterpiece of Kurd Efendi,which {Ata}i does not mention by name, is his two-volumeMurshidu ’l-an¸m ila d¸ri ’l-sal¸m fº sharº{ati ’l-Isl¸m, ms.Süleymaniye Library, Yusuf A¯a 258 (vol. 1) and Yusuf A¯a260 (vol. 2).

99. See Kafadar, “Self and Others.”

muq-20-2.pmd 9/2/2003, 4:24 PM185

Documents

A BUILDING BETWEEN THE PUBLIC AND PRIVATE REALMS OF …