CMYK CMYK CMYK CMYK CMYK CMYK CMYK CMYK CMYK CMYK

ISSN: 1307-8852

Anadolu ve Çevresinde
ORTAÇA⁄

Anadolu ve Çevresinde ORTAÇA⁄

3

AKVAD (Anadolu Kültür Varl›klar›n› Araflt›rma Derne¤i)

Ankara, 2009
3
2009

CMYK CMYK CMYK CMYK CMYK CMYK CMYK CMYK CMYK CMYK
 The Mosque-Hospital Complex in Divriği:
A History of Relations and Transitions
OYA PANCAROĞLU

The name of Divriği is virtually synonymous with the remarkable early thirteenth-century
building complex located on a hillside nearly halfway between the citadel and the lower town (fig.
1)1. Host to the most effusive and distinct examples of stone carving on medieval Anatolian
architecture, this monument consists of two institutions: a mosque and a hospital built in 1228-29
respectively by Aḥmadshāh and Tūrān Malik, two members of the regional Mengujekid dynasty2. An
associated bathhouse at a short distance away, now in ruins, was possibly conceived as part of this
socio-religious complex3. This institutional configuration, juxtaposing the accommodation of
communal Muslim prayers with the dispensation of social services, is a familiar one from the
medieval Islamic world and would have a particularly long history in Anatolia and the lands of the
Ottoman Empire4. The plan of the mosque with five aisles consisting of five bays each and a high
dome in front of the mihrab renders a type which evolved in twelfth- and thirteenth-century Anatolia
(fig. 2). The hospital, like its counterparts in other parts of the medieval Islamic world, is a variation
of the familiar madrasa plan with rooms and vaulted halls disposed around a covered court with an
arcade. One of the rooms of the hospital, designated as the tomb chamber, is accentuated by a high
dome and provided with a window opening to the interior of the mosque on its qibla wall.
This relatively standard institutional configuration and basic architectural typology contrasts
significantly with the main decorative program of the complex. Radically different from all known
contemporary examples, the decoration—concentrated especially on the monumental portals of the
two buildings—consists most notably of stylized vegetal motifs in high-relief stone carving of truly
astounding quality and variety (figs. 3-4). This decoration has no known close counterparts on
architecture although echoes of some of the components can be discerned in a limited and disjointed
manner on other media such as stucco, woodwork, and book illumination. By contrast, on the eastern
wall of the mosque, there is an opening framed by a more typical portal design consisting of a
muqarnas niche and a rectangular surround with low-relief geometric decoration of the type seen in
numerous contemporary buildings of the Seljuk dynasty especially in central Anatolia (fig. 5). This
feature may have served either as a window or, more likely, as a door giving restricted access to the

1 The main studies on the complex include Önge et al. 1978, Kuban 1997 and Sakaoğlu 2004. Numerous articles
and book chapters have also been written which address particular aspects of the monument’s structure or
decoration; these can be found in the Bibliography and will be cited in the footnotes where relevant. It may be
noted that Önge et al. 1978 also contains reprints of earlier studies up to 1978.
2 In this article, personal names will be fully transliterated according to the current academic conventions with
diacritics used in the transcription of the Arabic letters into English (with the addition of “ı” instead of “i” in
certain cases, e.g., Qılıch Arslān or Yaghıbasan). For place names, the most current modern spellings will be used.
Names of dynasties are also not subject to full transliteration. For a note on Tūrān Malik’s name, see footnote 12.
3 The bath, known as Hamam-ı Bâlâ or Bekirçavuş Hamamı, is in ruins but it was excavated in 2003 and a plan has
been published. Sakaoğlu 2004, 424-6 and Önge 1978, 33 see it as part of the mosque-hospital complex but this
opinion is not shared by Özbek 2004 who suggests it may have been built towards the end of the thirteenth
century.
4 Pancaroğlu forthcoming/2010.
170 OYA PANCAROĞLU

interior of the mosque via the wooden gallery in the southeast corner of the mosque known as the
mahfil5.
In addition to the problem of the visual contextualization of the decoration, the difficulties
surrounding the interpretation of the monument are compounded by structural issues which can be
traced at least to the sixteenth century when a minaret was built in the northwest corner above a bulky
cylindrical corner buttress. It seems clear that at least a portion of this façade along with all five bays
of the westernmost aisle (and probably three northern bays from the adjacent aisle) of the mosque had
collapsed and were rebuilt at a later, though as yet undetermined, date. The western portal of the
mosque is therefore assumed, for structural as well as stylistic reasons, not to be original to the
building (Kuban 1997, 49, 57-8, 127-32). This assumption is also supported by the particular
formulation of the inscription placed above the entrance which contains a reference to the “initial
construction” of the mosque (see Appendix). On the other hand, it must be noted that a number of
spolia with notably figural elements—including single- and double-headed eagles—were
asymmetrically incorporated into the wall around this western portal and may well have been
harvested from another Mengujekid period building or from the collapsed original western façade.
Although the western portal of the mosque does raise important questions about later restorative and
artistic practices, the apparent absence of any visual or textual clues to its date currently hinders
meaningful historical speculation about this particular feature of the monument.
A detailed endowment (waqf) deed must have certainly been drawn up for the complex, outlining
its endowed income and regulating the dispensation of the funds from those endowments for the various
expenses and salaries associated with the running of the whole complex. An extant endowment deed
bearing the date of 15th of Muḥarram 641 (July 4, 1243) appears to be a somewhat condensed and later
redaction of the original document and is concerned only with the mosque6. Nevertheless, some sense of
the original endowment can be gleaned from Ottoman tax registers (tahrir defterleri) for the district of

5 Because the building is constructed into the hillside, the eastern opening which is at ground level from the exterior is
situated well above the interior floor level, hence the function of the mahfil. Opinions have varied on the originality
of the wooden mahfil construction, with implications on whether the opening on the eastern wall is interpreted as a
window or a portal. For the various positions on this issue see Crowe 1972 and Kuban 1997, 133-5 (for the window
interpretation dismissing the originality of the mahfil) and Yavuz 1978 and Sakaoğlu 2004, 275-9, 309-11 (for the
portal interpretation affirming the originality of the mahfil). The portal interpretation appears to be the more logical
one based on the exterior design which is normally applied only to entrances as well as the particular articulation and
decoration of the vaults above the mahfil (Yavuz 1978). Moreover, the rather sordid story of the apparently stolen and
dispersed wooden parapet panels (Önge 1969 and Sakaoğlu 2004, 275-9) strongly suggests that the mahfil was indeed
an original and highly decorated gallery probably intended for the seating of the ruler. A recent dendrochronological
study of the remaining wooden beams has also revealed at least two of them to date to around 1240 (Kuniholm 2000,
107-8) which is the same year as the wooden minbar of the mosque which will be discussed further below. The eastern
opening, therefore, most probably functioned as a restricted access entrance for the ruler and his entourage. It is most
likely that this mosque entrance on the eastern façade was also more convenient for the royal party presumably
arriving from the palace in the citadel, rather than the main north portal which would require proceeding further down
and circling around the whole complex. In this regard, it is reminiscent of the siting of the late thirteenth-century
Arslanhane Mosque in Ankara which is also situated on a slope so that the main entrance opposite the mihrab is at a
higher elevation than the floor level of the mosque and is served by a wooden gallery.
6 Sakaoğlu 2004, 327-38; Kayaoğlu 1978; Berchem and Edhem 1917, 82-3 and 107-10. A critical take on the veracity
of this document is found in Sakaoğlu who suspects it to be a seventeenth- or eighteenth-century redaction; most
others who reproduce or mention it (including Gülsoy and Taştemir 2007, 91) are generally unconcerned about its
authenticity and do not seem to have studied the document itself. Berchem and Edhem 1917, 82-3 stated that they
only saw a copy of a document reported to be preserved in Divriği at the time. According to Sakaoğlu 2004, 338, n.
1, this document is a later summary redaction of the original and is now kept in the Directorate of Pious Endowments
in Ankara. Evliya Çelebi, the seventeenth-century Ottoman traveler, noted that the date and endowments of the
mosque had been written on the “door and wall” of the mosque on account of the seven (years worth of?) Rūm taxes
which had been spent on it (cited in Sakaoğlu 2004, 447-8). Given that Evliya Çelebi also identified the builder of
the mosque as the Seljuk sultan ‘Alā al-Dīn Kayqubād, this statement may be the product of some degree of
misconception or imagination on his part.
 The Mosque-Hospital Complex in Divriği: 171

Divriği from the years 1519 and 15307. From these records, which mainly constitute verifications of the
existing endowed properties, it is also possible to reconstruct a picture of the salaried posts in the mosque8
and to note that, probably sometime in the Mongol period and definitely by the early Ottoman period, the
hospital, now identified as Medrese-i Kübrâ (“Grand Madrasa”), had ceased to offer medical services
and/or instruction and functioned instead as an institution staffed for the teaching of law9.
Much has already been said and written about the Mosque-Hospital Complex of Divriği in
recognition of the unique status it occupies in the medieval cultural heritage of Anatolia. It should be
noted that issues of conservation have reached an acute phase and the future of the monument is cause
for grave concern, hinging precariously on the resolution of divergent interests and opinions surrounding
its preservation.10 It is hoped that further research on aspects of the monument and its physical and
historical contexts may contribute to better considered efforts for its preservation. To date, the
architecture and the decoration have been extensively studied, albeit mainly from a formalistic
perspective, and a rudimentary historical contextualization has already been achieved. The aim of this
essay, therefore, is neither to add to the current pool of information nor to provide an entirely new
interpretation but rather to undertake a historically contextualized assessment of this remarkable

7 Gülsoy 1995, 111-14, 118-21 and Gülsoy and Taştemir 2007, L-LIII, LV-LVI, 212-15, 220-21.
8 The 1519 Ottoman tax register lists the following salaried posts in the mosque: a hatîb(preacher), a nâzır (supervisor
of endowments), an imam, a muezzin, a ferrâş (caretaker), a sermahfil (an aide to the muezzin or hatîb), a Shafi‘i
imam, and a mu‘arrif (functionary in charge of leading invocations); see Arıkan 1991, 56-7 and Gülsoy 1995, 112-3
(who omits mention of the sermahfil). The extant redaction of the Arabic endowment deed dated 1243 does not list
the separate post of the Shafi‘i imam but adds the posts of juz’khwān (Qur’an reader), qubbadār (functionary in charge
of the tomb chamber which is in fact located within the hospital) and murammim (repairman?); see Berchem and
Edhem 1917, 107-10 and Sakaoğlu 2004, 328.
9 It is mentioned in the 1519 tax register that the original endowed properties of the “medrese” had been lost (zayi)
during the Mamluk period (1401-1516 for Divriği) and that its original endowment deed could not be found (Gülsoy
1995, 119 and Arıkan 1991). However, as recorded by both tax registers, two substantial sets of additional
endowments were made in the fourteenth century (Gülsoy 1995, 118-20; Arıkan 1991, 59-62; and Gülsoy and
Taştemir 2007, LV-LVII, 220-3) which suggests that the loss of the original endowed properties may have occurred
not in the Mamluk period but rather sometime in the period of the suzerainty of the Mongols and/or of their successors
in east-central Anatolia, the dynasty of Eretna (after 1343) followed by that of Qāḍī Burhān al-Dīn (after 1380). It
may even have occurred as early as 1276-7 when the Īlkhān Abaqa, on his way to meet the Mamluks further south in
Elbistan, passed through Divriği and ordered its walls to be destroyed (Balgalmış 1994, 452). The earlier of the
fourteenth-century endowments mentioned in the sixteenth-century tax registers is dated 724 A.H. (1323-4) and
attributed to one Fāṭima bint Ḥasan said to be the nâ’ibe (deputy or delegate) of Tūrān Malik. Fāṭima bint Ḥasan’s
designation as nâ’ibe suggests that she may have enjoyed a status as legal proxy or heir to Tūrān Malik who must have
died sometime before 1300. The 1519 tax register concerning Fāṭima bint Ḥasan’s endowment (which she apparently
made with the provision that she be buried in the tomb chamber) makes reference to the “medrese” suggesting that
the institutional transformation of the hospital had been realized by the early fourteenth century. Her generous
endowment provided, among other things, for the stipends of students and the salaries of jurists as would be expected
in a madrasa dedicated to the teaching of law. The tax registers also record that, in 797 A.H. (1394-5), further
properties were endowed by one Tāj Malik bint Malik Meḥmed Beg ibn Bahrāmshāh, who, judging by her name,
appears to be a niece of Tūrān Malik (probably the daughter of her brother Muẓaffar al-Dīn Muḥammad, the ruler of
Koloneia/Şebinkarahisar between 1225-8 who was later exiled to Kırşehir; see footnote 16). Tāj Malik’s endowment
made provisions for the employment of four fukahâ’ (jurists), two huffâz (Qur’an memorizers), and one ferrâş in
addition to specifying the amounts to be spent on tedrîs (i.e., instruction by a müderris [instructor]), i‘âdet (i.e.,
repetition of lessons by a mu‘îd [assistant instructor]), and nezâret (supervision by a nâzır). Despite the deviation from
the original institutional function of hospital (dār al-shifā’) envisioned by the founder and specified in the foundation
inscription, the trusteeship of the former hospital continued to remain in the hands of Tūrān Malik’s descendents. Two
such women are specified in the tax registers for the so-called “Medrese-i Kübrâ.” One of them was Tāj Malik and
the other was Isfahān Shāh bint Pir Meḥmed, identified in the 1530 tax register as a descendent of Tāj Malik and as
the current trustee of the endowment (Gülsoy and Taştemir 2007, LV-LVI, 221). The mention of these two women in
the tax registers and their continued use of royal names indicate the prestige and status enjoyed by the descendents of
the Mengujekids, long after the demise of the dynasty. Together with Fāṭima bint Ḥasan, they also provide evidence
for the continuation of women’s involvement in the life of an institution first founded by a woman.
10 For a most recent and astute perspective on these issues, see Arel 2009.
172 OYA PANCAROĞLU

monument which is the product of a particular conjunction of geopolitical motivations and cultural
resources in medieval Anatolia.

Peerless Portals
The main entrance to the mosque is through the northern portal that is directly opposite the mihrab (fig.
3). The exuberance of the high-relief carving on this portal is evident especially in the garland-like
arrangement of highly stylized vegetal motifs such as leafs, split palmettes, intersecting stems, and blossoms
as well as large medallions in crescent surrounds which join each other to embellish the grand rectangular
frame (figs. 6-8). On each side, this rectangular frame is edged by a slender engaged column topped with a
large capital incorporating muqarnas elements from which another column-like feature with a base and a
bundled shaft rises towards the upper edge of the frame. Predominantly geometric motifs—though quite
unlike the interlacing star-and-polygon patterns typical of the period in central Anatolia and also exemplified
in the eastern portal of the mosque and on its wooden minbar—decorate the recessed areas of the niche
around and above the doorway which is crowned by a large hexagonal motif11. Two inscriptions—one
contained in a band above the hexagonal motif and one just below the tip of the large niche—identify the
builder and the date of the building (see Appendix). Aspects of the decoration of this portal are echoed in
the apparently unfinished decoration of the grand mihrab composition which is visible from the north
entrance. This is evident especially in the two clusters of large palmette motifs on either side of the mihrab
niche similar to the motifs of the portal garland and in the bundled pilasters on the outer frame of the large
rectangular mihrab surround recalling the fantastic outer engaged columns of the north portal.
Both the geometric and the vegetal motifs in the decoration of the north portal of the mosque find their
counterparts in the decoration of the hospital portal on the western façade (fig. 4). However, these elements
are used far more sparingly on the hospital portal which is especially remarkable for the seemingly
gothicizing profile of its grand projecting arch, giving a much different impression than the mosque’s north
portal. A highly decorated mullion window above the entrance forms the visual focus of the entire
composition. Not as densely decorated as the north portal of the mosque, the hospital portal nonetheless
displays motifs executed in similarly high-relief carving. These include stylized vegetal motifs, large
projecting medallions in crescent surrounds filled with star motifs as well as a pair of stylized plant forms
surmounted by human heads (with their facial features mutilated at an unrecorded time) towards the outer
edge of the portal (figs. 9-10). Above these, large engaged capitals suspended without columns lend a
fanciful quality to the design, not unlike that discerned on the mosque portal and mihrab. The flat recess of
the entrance also incorporates low-relief carving of vegetal scrolls although some of this was clearly left
unfinished. The three-line foundation inscription of the hospital is located just below the mullion window
(see Appendix).

The House of Mengujek

While much remains puzzling about the origins of such decorative motifs and their compositional
configurations, the identity of the patrons, recorded in the foundation inscriptions of the mosque and the
hospital, makes the historical contextualization a less tricky task than the artistic contextualization. The
foundation inscription of the mosque states that it was built by Aḥmadshāh, son of Sulaymānshāh, in the
year 626 (1228-9). Aḥmadshāh was one of the latter-day rulers of the Mengujekid dynasty which was
established at the end of the eleventh century. Further above is a visually less impressive and less formal
inscription—written almost as an afterthought—which acknowledges the reign of ‘Alā al-Dīn Kayqubād
(r. 1219-37), the Seljuk sultan who was the contemporary of Aḥmadshāh. As for the hospital (dār al-
shifā’), its foundation inscription declares that it was built by the “just queen” (al-malika al-adila) Tūrān

11 Kuban 1997, 102-3, has noted the curious structural composition of this portal with its inner niche constructed in a
trabeate technique, unusual in Anatolia.
 The Mosque-Hospital Complex in Divriği: 173

Malik,12 the daughter of Fakhr al-Dīn Bahrāmshāh in the year 626 (1228-29). Fakhr al-Dīn Bahrāmshāh (r.
1162 or 1165 – 1225) was the most celebrated of all Mengujekid rulers in Erzincan, with an extraordinarily
long reign lasting some six decades. It is often assumed that Tūrān Malik was the wife of Aḥmadshāh
although neither the foundation inscription nor any other contemporary written source actually confirms a
marital connection between the two. Instead, it is her kinship to her late father that is pronounced in the
inscription13.
The history of the Mengujekid principality is little known beyond a basic outline of dynastic succession
and instances of encounter with outside forces14. The establishment of the dynasty is attributed to its
eponymous but rather mysterious founder known from the sources as Mangūjak Ghāzī15 who seems to have
been a Turkmen military commander in the service of the Great Seljuk sultan and who settled in the region
of the upper Euphrates basin sometime around 1080, in the aftermath of the Battle of Manzikert in 1071.
Based primarily in the town of Kemah, Mangūjak Ghāzī’s territories appear to have included the upper
Euphrates catchment region between Erzincan in the east and Divriği in the west16. In 1142, after the death
12 Tūrān Malik’s name is very frequently spelled in modern Turkish writing as Turan Melek (rather than Turan Melik), giving
the impression that the second half of this royal name carries the meaning of “angel.” Although common as a woman’s name
in (late) Ottoman and modern Turkish, Melek/Malak (which is written in the Arabic script simply as m-l-k without vowels)
in the sense of angel is not compatible with medieval naming customs. There are two other possible options for the
vocalization of m-l-k: Mulk/Mülk meaning “dominion” or “sovereignty” and Malik/Melik meaning “king” or “ruler.” The
preference for the second option is based on a number of factors, not the least of which is that medieval female royal names
in the Turco-Persianate world often do use such a construct, adding a normally male title such as malik, shāh, khān, or sulṭān
after a place name or concept. Examples can be cited from Mengujekid dynastic history itself—Tāj Malik and Isfahān Sh-
āh (see footnote 9)—as well as from later Turco-Mongol-Iranian dynastic history; on Jahān Malik the fourteenth-century
poet-princess of the Injuid dynasty in Shiraz, see Brookshaw 2005. Malik in such a construct is not so much a title (for which
khāṭūn and malika are the most common) as it is an integral part of the royal name. Thus, in her Divriği inscription, the royal
name of Tūrān Malik is preceded by her title al-malika al-‘ādila (“the just queen”).
13 The ambiguity of the relationship between the two patrons is explored in Berchem and Edhem 1917, 82 and Sakaoğlu, 2004,
166-8. The fact that the foundation inscription is silent about Tūrān Malik’s marital status is entirely in keeping with the
majority of the foundation inscriptions documenting female patronage in medieval and early modern Anatolia where
women’s vertical kinship (to her father or son) rather than matrimonial relationship was expressed. Thus, it is not impossible
thatAḥmadshāh and Tūrān Malik were not only members of the extended Mengujekid family but also married to each other.
The evidence, however, is silent on this matter. To complicate the matter further, the summary endowment deed dated 1243
(which may in fact date from a later period) makes mention of a woman named Fāṭima Khāṭūn identified as Aḥmadshāh’s
mother and joint patron of the mosque, but with no mention of the hospital. Whether this Fāṭima Khāṭūn should be identified
in turn with Tūrān Malik is not apparent from this particular document which makes no mention of the latter. The 1519 and
the 1530 tax registers mention yet another Fāṭima (see footnote 9 above): this is Fāṭima bint Ḥasan, said to be the nâ’ibe
(deputy) of Tūrān Malik. This Fāṭima made a very generous endowment in 1323-4 not only for the madrasa (formerly the
hospital) but also for the muezzin of the mosque and for the preparation and distribution of food on holy days. As the same
tax registers make no mention of a Fāṭima Khāṭūn, the supposed mother and co-patron of the mosque (the foundation of
which the registers assign solely to Aḥmadshāh), it may be suggested that the endowment deed dated 1243 is indeed a much
later redaction which obfuscated the identity and role of Fāṭima bint Ḥasan.
14 On the history of the Mengujekids, see Sakaoğlu 2004, Sümer 1990, 1-14, Turan 2001, 73-97, and Cahen 1988, passim.
15 The spelling of the name given in the medieval sources in Arabic script may be transliterated as either Mangūjak or
Mangūjuk.
16 Although the well-fortified Şebinkarahisar (the medieval Koloneia/Kughūniya) to the northwest of Erzincan is mentioned
in some of the secondary sources as being part of the twelfth-century Mengujekid territories (for example Sümer 1990, 7
and Turan 2001, 73), it is more probable that it actually passed into Mengujekid hands only around 1201-2 in the course of
the dissolution of the Saltukid dynasty (based in Erzurum) by the Seljuks under sultan Rukn al-Dīn Sulaymānshāh
(Mordtmann 1978, 578). On the history and monuments of Koloneia, see Bryer 1985, vol. 1, 145ff. The Byzantine Greek
historian Niketas Choniates mentions Koloneia as being in the hands of the Saltukids at the end of the twelfth century when
the future emperor of Byzantium, Andronikos Komnenos, passed through during his exile from Constantinople. Choniates’
ascription of Koloneia to the Saltukids at this time is disputed by Minorsky who suggests, without corresponding evidence,
that it must have been in the hands of the Mengujekids instead (Minorsky 1945, 555-6). It appears that the town changed
hands in turn between the Byzantines, Danishmendids, Saltukids, and Mengujekids between 1071 and 1202 before being
finally annexed by the Seljuks in 1228. The rule of Fakhr al-Dīn Bahrāmshāh (about whom see further below) probably
extended to this town as of 1201-2 and passed to his son Muẓaffar al-Dīn Muḥammad around the time of Bahrāmshāh’s
death in 1225 if not somewhat earlier. Muẓaffar al-Dīn was exiled by the Seljuks to Kırşehir in central Anatolia in 1228
when his brother Dāwūdshāh (the successor to Bahrāmshāh in Erzincan) was exiled to Ilgın and most of the Mengujekid
terrioties (apart from Divriği) was annexed. On the legacy of Muẓaffar al-Dīn in Kırşehir, see Sakaoğlu 2004, 110-8. On the
endowment his daughter made to Hospital of Tūrān Malik, see footnote 9.
174 OYA PANCAROĞLU

of Mangūjak Ghāzī’s son, Malik Isḥāq, sovereignty in these territories was divided among the latter’s
three sons, based in Divriği, Erzincan and Kemah, in accordance with the historically recurrent Turkic
custom of geographic power-sharing by members of the ruling family. Within a decade or two, Kemah
came to be ruled by Erzincan which soon rose to relative political prominence while Divriği, though ruled
by an uninterrupted line of Mengujekids for the next one hundred years, curiously failed to capture the
attention of medieval historians. Indeed, the existence of the rule of the Divriği Mengujekids is known
only from inscriptions in the town and from their modest numismatic record17.
The Erzincan branch of the Mengujekid house prospered especially during the long reign of Fakhr
al-Dīn Bahrāmshāh who ruled from Erzincan and guaranteed the security of his principality through
diplomatic relations with the Seljuks who were quickly becoming the most powerful dynasty in Anatolia,
especially after the annexation of Danishmendid territories between 1174 and 1178 by the Seljuk sultan
Qılıch Arslān II (r. 1155-92). During their last phase, the Danishmendids—one of the first Turkmen
dynasties to be established in central Anatolia in the aftermath of Manzikert—had been embroiled in a
difficult contest for power which saw them undertake risky and variable political schemes involving the
Byzantines, the Seljuks, and the Zangids of Syria18. The gradual eradication of the Danishmendid state
by Qılıch Arslān II picked up pace after the death of the emir Niẓām al-Dīn Yaghıbasan of Sivas in 1164.
Coinciding with the accession of Fakhr al-Dīn Bahrāmshāh to the throne in Erzincan, these developments
undoubtedly influenced the geostrategic pragmatism of the young Mengujekid ruler whose father, ‘Alā
al-Dīn Dāwūd, appears to have been killed during an incursion into eastern Anatolia led by the same
Yaghıbasan in 1162 or 1163. Accordingly, it is reported in the anonymous Tārīkh-i Āl-i Saljūq that, after
Qılıch Arslān II annexed the Danishmendid territories in 1174 (with the exception of Malatya which was
taken in 1178), [Bahrāmshāh] the ruler of Erzincan, along with the [Saltukid] ruler of Erzurum, declared
his submission to the Seljuks19. This natural entente was cemented (either previously or subsequently) by
the marriage between Bahrāmshāh and a daughter of Qılıch Arslān II who is known as ‘Iṣmatī Khāṭūn20.
Later in the early thirteenth century, Bahrāmshāh gave one of his own daughters in marriage to Qılıch
Arslān’s grandson, the sultan ‘Izz al-Dīn Kaykāwūs I (r. 1211-19)21. This was one of Tūrān Malik’s
sisters, known by the royal name of Saljūq Khāṭūn. In addition to giving a nuptial framework to the
political strategy of deference and cooperation, these significant marriages contracted with the Seljuk
royal family provided the occasion in which the Mengujekid house could make a spectacle of the wealth
and fineries they commanded and presented to the increasingly more powerful Seljuks in the form of
sumptuous dowries and court ceremonial emanating from or destined for Erzincan.
Recorded in history as an active patron of the arts and charitable public works, Fakhr al-Dīn
Bahrāmshāh was the dedicatee of Makhzan al-Asrār (“Treasury of Mysteries”), a didactical masnavī
composed by the poet Niẓāmī of Ganja22. A poet of great renown in all of the Persian-speaking lands
between Central Asia and Anatolia, Niẓāmī lived in his hometown of Ganja in the Arran district of

17 For the first and most comprehensive documentation of the inscriptions and coinage of Divriği (as well as some
documents), see Berchem and Edhem 1917, 55-114. Sakaoğlu 2004 encapsulates much of this information,
sometimes with variant readings, and is currently the most extensive compilation on the history of the Mengujekids
and their monuments.
18 Melikoff 1965, 110-1 and Turan 2002, 197-205.
19 Anonymous 1999, 82. From this brief mention, it is only possible to detect the adoption of a general policy of
deference to the Seljuks rather than to extrapolate the precise nature of the political relationship between the
Mengujekids of Erzincan and the Seljuks in the aftermath of the dissolution of the Danishmendids.
20 Sakaoğlu 2004, 70. ‘Iṣmatī Khāṭūn’s name (and her royal name Tāj Malik Khāṭūn) is given in the fourteenth-century
Mevlevi hagiographies thanks to her patronage of the family of Mawlānā Jalāl al-Dīn Rūmī (Lewis 2000, 69).
21 The ceremonial of this wedding is described in some detail in Ibn Bībī 1956, 172-82 (Turkish translation: Ibn Bībī
1996, 192-201). See also footnote 30.
22 Nizami Ganjavi 1374/1995 (English translation: Nizami Ganjavi 1945 which should be used with some caution for
its omission of some lines).
 The Mosque-Hospital Complex in Divriği: 175

Transcaucasia (now in modern Azerbaijan), receiving patronage mainly from regional Muslim rulers
who emerged in the area of northwest Iran, Transcaucasia, and eastern Anatolia after the breakdown
of the Great Seljuk sultanate in the late twelfth century (Chelkowski 1995). Makhzan al-Asrār was
the first of Niẓāmī’s five epic poems, known collectively as the Khamsa. Although its exact date is
not known, Makhzan al-Asrār was almost certainly composed in the early part of Bahrāmshāh’s reign,
possibly as early 1165-66 and certainly before 118823. In the section on the praise of the prince,
Niẓāmī expresses Bahrāmshāh’s sovereignty in dual attributes: presenting him as the king of Armenia
(malik-i Arman) and the emperor of Rum (shāh-i Rūm); relating his throne to the sultanate (salṭanat
awrang) and his seat to the caliphate (khilāfat sarīr); claiming him as the conqueror of Rūm (Rūm-
sitānanda) and the captor of Georgia (Abkhāz-gīr) (Nizami Ganjavi 1374/1995, 21-2). These paired
attributes, which are modeled on contemporary titles favored especially by the Turkmen rulers of
Anatolia, show the particular geopolitical matrix within which Bahrāmshāh was beginning to assert
his own identity and orientation24. Accordingly, the glorification of Bahrāmshāh’s political figure was
embedded in a shared Anatolian geography (and demography) that extended from Byzantium (Rūm)
to Armenia (Arman) and Georgia (Abkhāz) and articulated by a multi-referential political framework
that made use of both Persian and Arabic titles (shāh and malik) and both secular and religious sources
of Muslim sovereignty (salṭanat and khilāfat).
Elements of this construction of a political persona are also reflected in Bahrāmshāh’s issues of
copper coinage, albeit in a more muted manner, with titles such as malik al-umarā’ (“king of emirs”),
shāh-i ghāzī (“warrior emperor”), and naṣīr amīr al-mu’minīn (“helper of the commander of the
faithful”)25. It may be noteworthy that none of the extant coins of Bahrāmshāh acknowledge the Seljuk
sultan which seems to indicate that Bahrāmshāh enjoyed some degree of political independence (or,
rather, the semblance thereof) in Erzincan and Kemah while assuring the Seljuks of his allegiance and
loyalty by his actions. Thus, in 1203, Bahrāmshāh participated in the Seljuk sultan Rukn al-Dīn
Sulaymānshāh’s failed campaign against Georgia during which the Mengujekid ruler was captured.
Taken to the presence of the Georgian queen Tamar, Bahrāmshāh seemingly reaped the benefits of his fine
reputation as a ruler and was honored by the queen who released him after the payment of ransom26. This
incident encapsulates the remarkable nature of Bahrāmshāh’s rulership which not only coped with forces
both to the east and west of the Mengujekid domains but also turned these to the advantage of a dynasty

23 On the difficulties of dating Makhzan al-asrār any more precisely, see De Blois 1997, 439-40. The fantastic
suggestion that Bahrāmshāh may have personally met Niẓāmī in July 1163 (Sha‘bān 558 A.H.) turns on the
proposition that the Mengujekid prince may have joined a coalition of Turkmen forces from eastern Anatolia and
Azerbaijan under the leadership of the penultimate Great Seljuk sultan Arslānshāh b. Tughrul (Sakaoğlu 2004, 67-8).
This coalition is said to have met in Ganja before a successful move against the Georgians in retaliation for the
Georgian victories of 1161 and 1162. As the year of Bahrāmshāh’s accession to the throne is not known with certainty,
it is unclear under which title he may have joined such a campaign (if he did indeed join). It must be noted that this
engagement of the Mengujekids against the Georgians is not confirmed by all; see especially Peacock 2006, 129, 133
and 135 who mentions that most eastern Anatolian dynasties were at one time or another involved against the
Georgians but finds it unlikely that the Mengujekids joined any campaign against the Georgians before 1200.
24 On the construction of titles, see Shukurov 2001 who explains the conceptual underpinnings of the use of the toponym
Rūm by Turkic rulers in Anatolia: “Rum/Rhomania in Turkoman titles signified at the same time both the territory of
the Roman empire in the general sense and [the Anatolian lands] conquered by the Turks … as a specific, ‘Muslim’
segment of the ‘Roman’ space” (Shukurov 2001, 267). Hence, “[i]t was not a question of legitimacy but of self-
identification. Turks styled themselves Romans/Rumi since they mastered the lands of ΄Ρωμανία (bilad al-Rum,
diyar-i Rum) and became residents of it together with Christian Greeks and Armenians” (Shukurov 2001, 270). This
definition also sheds light on the use of other toponyms such as Arman and Abkhāz.
25 Artuk and Artuk 1970, 387-8 and Sakaoğlu 2004, 82-6.
26 Kaya 2006, 88 and Turan 2002, 259. According to the Georgian sources, however, Bahrāmshāh was not a willing
partner in the Seljuk campaign of 1203 and, in its aftermath, the Mengujekids were saddled with having to pay tribute
to the Georgians for some time (Shukurov 2005, 102).
176 OYA PANCAROĞLU

in an increasingly more pressurized political environment27. The longevity of Bahrāmshāh’s reign should
undoubtedly be ascribed to his personal pragmatic vision and is indicative of a highly receptive diplomatic
policy which is also echoed in the cultural life of the Mengujekids.
Bahrāmshāh’s distinct profile in the realm of cultural production was firmly established early in his
career with the renown he gained as a result of being the patron of Niẓāmī’s Makhzan al-asrār. His
reputation in this regard was also recognized at some length by the Seljuk historian Ibn Bībī who outlined
in detail the generous reward which Niẓāmī received for his poem from Bahrāmshāh. Indeed, Ibn Bībī
did not hesitate to sing the praises of Bahrāmshāh and to uphold him as a rare model of the just and
beneficent ruler28. Already just a couple of years after his Georgian captivity, the historian Rāwandī
extolled Bahrāmshāh for his conduct as a valorous commander in the 1203 campaign (Peacock 2006,
135). The reverberations of his exceptional standing in the Seljuk court and beyond served to substantiate
Niẓāmī’s glorification of his patron in geopolitical terms, extending to the east and the west. It is this
model of generous and pious patronage in the competitive political context of medieval Anatolia which
Bahrāmshāh embodied and exemplified to a greater degree than most rulers of his day.
These qualities were also espoused by members of his family, including his Seljuk wife ‘Iṣmatī
Khāṭūn who, together with her husband, is said to have extended a warm welcome to Bahā’ al-Dīn Walad,
the renowned preacher-scholar from Central Asia and the father of Mawlānā Jalāl al-Dīn Rūmī, as
recorded in Aflākī’s hagiography, Manāqib al-‘ārifīn29. Thus, it is related that ‘Iṣmatī Khāṭūn and Bah-
rāmshāh, upon hearing of Bahā’ al-Dīn Walad’s passage from Malatya to central Anatolia, caught up with
him in the town of Āqshahr west of Erzincan and enticed him to stay by agreeing to his request to build
a madrasa for him in that very location30. According to Aflākī, Bahā’ al-Dīn Walad stayed and taught in
the Mengujekid realm for four years, before moving to Larende (modern Karaman) in central Anatolia,
purportedly upon the death of his Mengujekid patrons. This madrasa at Āqshahr, known as the ‘Iṣmatiya,
is said to have been built and sponsored by ‘Iṣmatī Khāṭūn who also attended the lessons herself.
Although Bahā’ al-Dīn Walad apparently insisted on being accommodated in Āqshahr on account
of the large number of “bad persons” he alleged were living in Erzincan,31 this Mengujekid capital

27 It was just in advance of this campaign, around 1202, that the Seljuk sultan Rukn al-Dīn Sulaymānshāh annexed the
territories of the Saltukids immediately to the east of the Mengujekids in the region of Erzurum and installed his
brother Mughīth al-Dīn Tughrulshāh (formerly the malik of Elbistan) as the regional Seljuk ruler (Kaya 2006, 76-9).
Between 1202 and 1230, Erzurum was effectively a semi-independent branch of the Seljuks ruled first by Mughīth al-
Dīn Tughrulshāh and later by his son, Rukn al-Dīn Jahānshāh. The Seljuk decision to put an end to the Saltukid
dynasty is somewhat puzzling (Peacock 2006, 133-5). As to the Mengujekids’ avoidance of Seljuk territorial
expansion in this instance, it may perhaps be attributed to Bahrāmshāh’s personal marriage alliance with the Seljuks
(as son-in-law to Qılıch Arslān II and brother-in-law to Rukn al-Dīn Sulaymānshāh). Somewhat later, a nuptial
alliance was also contracted with the Seljuks of Erzurum on the eastern flank of the Mengujekids with the marriage
of a daughter of Bahrāmshāh to Rukn al-Dīn Jahanshāh; an inscription indicating this princess (but without providing
her royal name) as the builder of a tower in Bayburt is dated 1212-3 (Sakaoğlu 2004, 78-9). Although it is not possible
to ascertain the year in which this marriage was contracted, it is worth noting that it was balanced, either just
previously or subsequently, by the marriage of another daughter, known as Saljūq Khāṭūn, to the Seljuk sultan ‘Izz al-
Dīn Kaykāwūs I. Ibn Bībī reported that it was ‘Izz al-Dīn Kaykāwūs I and his circle of advisors who sought the hand
of the Mengujekid princess in this instance as an ideal matrimony for the newly-minted Seljuk sultan in Konya; see
Ibn Bībī 1956, 172-82 (Turkish translation: Ibn Bībī 1996, 192-201).
28 Ibn Bībī 1956, 70-2 (Turkish translation: Ibn Bībī 1996, 91-3). In this passage Ibn Bībī also relates the well-known
story of Bahrāmshāh’s endowment for the provision of food for birds and wild animals during the harsh winters of
Erzincan.
29 Aflākī 2002, 19-20. See also Lewis 2000, 69-70.
30 Not to be confused with Akşehir located west of Konya, this Āqshahr west of Erzincan probably corresponds to the
modern village of Akşar, to the southeast of Suşehri (ancient Nikopolis) in Sivas province. For a medieval geographic
localization of this town, see Le Strange 1905, 157.
31 Aflākī 2002, 19 and Lewis 2000, 71. Although Aflākī does not elaborate on the nature of the putative offense,
Erzincan appears, as reported by the contemporary geographer Yāqūt, to have had a reputation for the conspicuous
public consumption of wine by the majority of its residents who were Armenians and Bahā’ al-Dīn Walad may have
wanted to avoid such an association (Yāqūt 1977, 150).
 The Mosque-Hospital Complex in Divriği: 177

nonetheless attracted other distinguished (or less fussy) visitors who must have recognized the cultural
prestige of Bahrāmshāh’s court. The reputation of Erzincan as a significant center for patronage is
demonstrated by the arrival in the 1220s of the prominent scholar and physician ‘Abd al-Laṭīf al-
Baghdādī (1162-1231) who had made a name for himself in the Ayyubid lands as a polymath, teaching
in esteemed institutions in such cities as Cairo, Damascus, and Jerusalem before presenting himself at the
Mengujekid court after 1220 where he was apparently sponsored by Bahrāmshāh’s son and, after 1225,
successor, ‘Alā al-Dīn Dāwūdshāh (r. 1225-28)32. ‘Abd al-Laṭīf al-Baghdādī dedicated a number of his
important works to Dāwūdshāh who was known as a liberal and learned patron who composed poetry and
took an active interest in the natural and physical sciences in which he had profound knowledge33. ‘Abd
al-Laṭīf departed from Erzincan when Dāwūdshāh’s reign came to an abrupt end with the annexation of
the Mengujekid territories in Erzincan and Kemah by the Seljuk sultan ‘Alā al-Dīn Kayqubād in 1228.
It appears that the political strategy which Bahrāmshāh had meticulously followed during his reign
with regard to the Seljuks had now outrun its course in the face of an irresolvable political impasse
generated at least in part by the dual threat of the arrival of the Mongols in Transcaucasia and the
incursions of the army of the fugitive-invader from Central Asia, the Khwārazmshāh Jalāl al-Dīn
Mangūbartī, in eastern Anatolia34. These troubling developments on the doorstep of the Seljuks, coupled
with the soaring ambitions of the new sultan‘Alā al-Dīn Kayqubād and the rising tension with the
Ayyubids in Syria, almost certainly had an impact on the security—or, at least, the self-assurance—of the
Mengujekids. Nevertheless, the Seljuk historian Ibn Bībī represented Dāwūdshāh’s search for a new set
of regional alliances in this situation as an unacceptable act of infidelity to the Seljuks which apparently
justified ‘Alā al-Dīn Kayqubād’s annexation of the Mengujekid territories of Erzincan and Kemah and
the exile of Dāwūdshāh to the spa town of Āb-i Garm (Ilgın) in Seljuk central Anatolia where he was
granted an iqṭā’ (a revenue-generating land grant made by the state) to live out his days35.
With his latest patron clearly on the losing side in the new order forced by ‘Alā al-Dīn Kayqubād,
‘Abd al-Laṭīf al-Baghdādī left Erzincan and made his way back to his native Baghdad via Erzurum,
Kemah, Divriği, Malatya and Aleppo. It is rather tempting to see more than just a coincidence between
the itinerary of ‘Abd al-Laṭīf al-Baghdādī’s return journey around 1228-9 which places him in Divriği
between April and June of 1229 (Toorawa 2004, 65) and the construction of Tūrān Malik’s hospital in
Divriği within the same year. It is not impossible that the Iraqi scholar-physician considered the
opportunities for further patronage in the remaining Mengujekid domain of Divriği or, alternatively, that
Tūrān Malik (together with Aḥmadshāh) sought to entice ‘Abd al-Laṭīf al-Baghdādī to stay on in Divriği
in an attempt to sustain the cultural momentum which had attained an admirable level with the patronage
of her family in Erzincan. Although it is not possible to corroborate Tūrān Malik’s own motivation in this
regard, the 1530 Ottoman tax register also credits her with a madrasa (no longer extant) apparently known
as the Lâlâiye (Arıkan 1991, 62-3). The name of this madrasa is indicative of the patronage of an
unknown lālā (a tutor in a royal household), for whom it was probably built and endowed. As
Aḥmadshāh’s partner-in-patronage for the Mosque and Hospital project, Tūrān Malik’s interest in
providing institutional support for the pursuit and application of knowledge in Divriği appears to have
been a matter of dynastic and perhaps personal significance.

Mengujekid Architecture before Aḥmadshāh and Tūrān Malik

The links between Bahrāmshāh and Niẓāmī, ‘Iṣmatī Khāṭūn and Bahā’ al-Dīn Walad, and
Dāwūdshāh and ‘Abd al-Laṭīf al-Baghdādī exemplify Erzincan’s notable place in the network of

32 For the career and patronage of ‘Abd al- Laṭīf al-Baghdādī, see Toorawa 2004, Stern 1960 and Kaya 1988.
33 Ibn Bībī 1956, 358 (Turkish translation Ibn Bībī 1996, 367).
34 Turan 2002, 347-57 and Cahen 1988, 78-89.
35 Ibn Bībī 1956, 345-52, 354-8 (Turkish translation: Ibn Bībī 1996, 355-62, 364-7).
178 OYA PANCAROĞLU

patronage in a world of multiple centers of power, each vying for its own sources of cultural production
and prestige. Erzincan, the city identified most closely with the long and celebrated reign of Bahrāmshāh,
must have been the primary stage on which the Mengujekid royal image was constructed in the form of
public and palatial buildings. As such, the buildings of Mengujekid Erzincan must have epitomized the
architectural orientation of the dynasty within the network of political and cultural rivalry which defined
the modus operandi of the successor states of the Great Seljuk empire which dissolved in the last decades
of the twelfth century. However, the location of Erzincan on the earthquake-prone north Anatolian fault
line has amounted to the continual destruction of practically all of its medieval and post-medieval
architecture over the centuries36. Nevertheless, several surviving monuments in Kemah and Divriği
provide valuable information about the direction of architectural orientation in terms of artistic resources
and epigraphic stylization employed by the Mengujekids towards the end of the twelfth century.
The Kemah monuments consist of two brick structures built next to each other, on an elevated
location on the bank of the Karasu, a major tributary of the Euphrates. Although undated, both structures
are usually thought to have been built in the last decade of the twelfth century. Of these, the one known
as the Tomb of Mengücek Gazi (or Sultan Melik) is an octagonal brick chamber with a conical dome built
above a crypt which has a central column37. The exterior walls are articulated by recessed arched panels
and engaged columns on the corners. The entrance is decorated with geometric patterns executed in
unglazed brickwork set into mortar along with four glazed ceramic bowls38. An inscription above the
doorway in Kufic script contains a familiar verse from the Qur’an (21:35): “Every soul shall taste death.”
Another inscription in Kufic, now much damaged and mostly unverifiable, on an exterior wall was read
in the early twentieth century as the signature of an architect named ‘Umar ibn Ibrāhīm al-Ṭabarī39. His
nisba (name of relation), al-Ṭabarī, if read correctly, indicates that he (or his family) came from
Tabaristan, the region south of the Caspian Sea in Iran. This tomb chamber belongs to a rich tradition of

36 For a list of the most serious historical earthquakes in Turkey, see the information provided by the Kandilli
Observatory of Boğaziçi University (http://www.koeri.boun.edu.tr/sismo/Depremler/ thistoric.htm). Listed on this site
are the following earthquakes impacting the Erzincan region: 1268, Erzincan-Erzurum (15, 000 deaths); 1458,
Erzincan-Erzurum (32,000 deaths); 1584, Erzincan-Erzurum (15,000 deaths); 1939, Erzincan (nearly 33,000 dead and
over 116,000 buildings damaged/destroyed). To these may be added mentions of other earthquakes in medieval
sources; see Turan 2001, 96. It seems that at least one madrasa built by Fakhr al-Dīn Bahrāmshāh survived (or was
rebuilt and retained its original name) into the seventeenth century (Sakaoğlu 2004, 91). With the 1939 earthquake,
whatever remained of medieval and Ottoman Erzincan was levelled and the new city was built to the north of the
railway.
37 Ünal 1967, 151-61, Meinecke 1976, vol. 1, 12-6, vol. 2. 192-5, Önkal 1996, 46-53, and Sakaoğlu 2004, 51-4, 94-5.
Although popularly attributed to Mangūjak Ghāzī, this tomb cannot have been built at the time of his death around
1118; see further below footnote 39.
38 On these features, see Meinecke 1976, vol, 1, 12-6, vol. 2. 192-5, Bakırer 1981, 45-6, 183-6, and Demiriz 1972-73,
180.
39 This reading was first offered by Kemali 1932, 241-3; see also Ünal 1967, 158 and Önkal 1996, 51. A second
damaged exterior inscription in brickwork which has only been partially read contains the name of an individual with
the title shaykh al-mashāyikh (“grand shaykh”) but it is unclear whether this identifies the builder/patron or the
(original?) occupant of the tomb. Thus, the question of by/for whom it was originally built at the end of the twelfth
century remains unanswered although the possibility exists that it may have been built as a dynastic tomb and perhaps
even marks the site where Mangūjak Ghāzī was buried. Two inscriptions (one in Persian, the other in Arabic) painted
on the interior which appear to be later than the time of construction extol the merits of Mangūjak Ghāzī and provide
the geneaology of the Mengujekid dynasty up to Saljūqshāh (Kemali 1932, 241-2, Ünal 1967, 159-60, Önkal, 1996,
51-2, and Sakaoğlu 2004, 51-4, 94-5). The latter is depicted in the inscription as a son of Bahrāmshāh and “the pride
of the houses of Saljūq and Mangūjak” (mafkhar āl-i Saljūq wa Mangūjak) which suggests that Saljūqshāh’s mother
was probably the daughter of Qılıch Arslān II whom Bahrāmshāh married. As the date of these interior inscriptions is
not known and have unfortunately been repainted in a recent case of “restoration”, it is not possible to relate them
securely to the original patron/occupant of the tomb. Nevertheless, these inscriptions cannot be dismissed altogether
as the titles of Saljūqshāh given here repeat those found in an endowment deed for a rest home he established in
Kemah dated Ramaḍān 587/October 1191, on which see Yinanç 1970-74 and Sakaoğlu 2004, 94-5 and footnote 43
below. Amplifying the mystery of this tomb further is the mummified body preserved in the crypt, the subject of
extensive local folklore on which see Sakaoğlu 2004, 51-3.
 The Mosque-Hospital Complex in Divriği: 179

brick construction and decoration that extended from Transoxiana to Azerbaijan (including Tabaristan)
between the tenth and the early thirteenth centuries. In particular, the geometric brickwork of the entrance
side, the articulation of the other seven sides with blind niches, and the peculiar structure of the crypt with
a central octagonal column supporting the vault find their closest counterparts in late twelfth-century tomb
chambers from Azerbaijan. The Tomb of Mu’mina Khatun (1186) in Nakhchivan and the Gunbad-i
Qabud (1196-7) in Maragha represent the epitome of this architectural tradition especially in terms of their
geometric brickwork highlighted with a measured but effective use of glazed bricks. Both buildings also
feature a crypt with a central supporting column40. In comparison with these examples, the Kemah tomb,
which is decorated only with unglazed bricks, represents a pragmatic rendering of a sophisticated
decorative practice undertaken by an émigré architect in the 1190s. The embedded glazed bowls, for
example, could be interpreted as a resourceful counterpart to glazed bricks or tiles which were perhaps
not as readily available in Kemah in the late twelfth century as they were in Azerbaijan41.
It is probable that the two-room building standing next to this tomb, known popularly as the Tomb
of Behramşah or the Zaviye of Sultan Melik, is also the work of the same architect42. Despite the fact that
there are three cenotaphs in the first room which is accessed by the single entrance, it is doubtful that this
building was originally intended to be a tomb. The entrance has brickwork decoration similar to that of
the Tomb of Mengücek Gazi, including the peculiar use of an embedded glazed bowl, making it likely
that the two structures were built around the same time and that this two-room structure was originally
related to the tomb chamber in some capacity, probably as a lodge (zāwiya) of some sort43.
Although the current state of knowledge hinders a less speculative interpretation about their dating
and functional relationship, these two buildings in Kemah nonetheless provide an idea about Mengujekid
architecture in Erzincan during the reign of Bahrāmshāh at least up to the turn of the thirteenth century.
Accordingly, it may be assumed that at least some (if not most) of the public buildings constructed in the
last decades of the twelfth century in the main Mengujekid capital were built in a style and technique
derived from contemporary architectural practices in greater Iran—in particular Azerbaijan—and adapted
to local circumstances. This aspect of Mengujekid architecture at the end of the twelfth century provides
a sense of the cultural-political context which saw the introduction of especially brick architecture and
decoration from the eastern Islamic world into medieval Anatolia where they ultimately encountered both
local and regional traditions of stone architecture, giving frequently mixed but often intriguing results.
Such a mixed result can be seen in the Kale (“Citadel”) Mosque of Divriği built in 1180-1 by the
local Mengujekid ruler Sayf al-Dīn Shāhanshāh (r. 1171? - 1196) who was Bahrāmshāh’s cousin and the
future ruler Aḥmadshāh’s grandfather44. This stone mosque which was, until the disastrous “restoration”

40 An earlier tomb chamber in Maragha, the Gunbad-i Surkh dated 1147-8, demonstrates the development of this
architectural style (without, however, this particular crypt construction) from at least the middle of the twelfth century.
On the tombs in Maragha, see Godard 1936. On the Tomb of Mu’mina Khatun in Nakhchivan, see Yazar 2007, 90-
104.
41 A single glazed bowl is embedded in the geometric brickwork decoration above the entrance of the Gunbad-i Surkh
in Maragha (1147-8) which is one of the earliest buildings in Azerbaijan to employ, in a very limited manner, glazed
tiles on the exterior; Pickett 1997.
42 Ünal 1967, 161-4, Meinecke 1976, vol. 1, 15-6, vol. 2. 192-5, Önkal 1996, 343-7, and Sakaoğlu 2004, 51-4, 94-5.
43 It may be that this building was related to a house for the destitute (dār al-ḍā’if) established by Saljūqshāh, the
endowment of which, dated 1191, survives in a published copy and seems to be partially quoted in one of the two
painted inscriptions in the interior of the tomb; see footnote 39 above and Yinanç 1970-74, Ünal 1967, 163-4, and
Sakaoğlu 2004, 54, 94-5. An inscription in brickwork above the entrance, now almost entirely destroyed, was
tentatively read as (including?) the name of Saljūqshāh ibn Bahrāmshāh; see Kemali 1932, 243, Ünal 1967, 163, and
Önkal 1996, 346).
44 Berchem and Edhem 1917, 56-62, Meinecke 1976, 13-5, 110-2, Kuban 2002, 108-10, and Sakaoğlu 2004, 217-26.
Sixteenth-century and later Ottoman tax and court registers attribute this mosque to Sulaymān, the father of
Shāhanshāh, which would mean that the initial construction might be dated to the decade or two before the 1180; see
Sakaoğlu 2004, 230-6.
180 OYA PANCAROĞLU

of 2007-8 undertaken by the Directorate of Pious Endowments, one of the oldest to survive unaltered from
late twelfth-century Anatolia, was furnished with a portal built and decorated in a combination of stone,
brick, and color glazed tiles arranged into geometric patterns. Signed by the master (ustādh) Ḥasan ibn
Pīrūz of Maragha, the portal shows the resourceful early adaptation of a portal form developed in the
tradition of brick architecture in greater Iran—extending from Transoxiana in the east to Azerbaijan in the
west—to the Anatolian tradition of stone architecture45. In other words, the design of this portal was
derived primarily from the decorative idiom of contemporary Islamic brick buildings in greater Iran and
complemented by the craft of stone architecture and its attendant technical and aesthetic dispositions. This
encounter between two distinct yet mutually responsive traditions is also evident in the interior of the
mosque. The mihrab niche of the mosque, carved out of a single stone block, may be described as a
simulation of the muqarnas form (the three-dimensional composition of niche-like elements) without the
proper articulation of its components which lack the customary concavity of the individual niche units
(Bakırer 1976, 71, 128-9). Apparently first developed in Iraq in the late eleventh century, the early
examples of muqarnas were created from plaster applied to a wooden framework46. Fine examples of
muqarnas applications are found in the late twelfth-century tombs of Nakhchivan and Maragha where this
three-dimensional decorative mode was skillfully adapted to brick and mortar. The full rendition of
muqarnas in stone took place in Syria and Anatolia following a phase of adaptation in the second half of
the twelfth century47. The mihrab of the Kale Mosque in Divriği with its “flat” articulation thus represents
a particular moment in the early Anatolian encounter of muqarnas with stone. Like the numerous modes
of geometric surface decoration in low relief, muqarnas and its various applications spread in the period
of the breakdown of the Great Seljuk sultanate in the twelfth century and constituted a significant aspect
of the visual culture of the emerging successor states48. To these two may be added the application of
colored tiles or bricks in exterior decoration which was becoming increasingly more prevalent in
especially Iranian architecture (as masterfully exemplified by the Tomb of Mu’mina Khatun in
Nakhchivan)49. The nascent presence of these three elements in Mengujekid monuments of the late
twelfth century associated with builders with connections to greater Iran attests to the pioneering and
experimental aspect of architecture in both branches of the Mengujekid dynasty at the end of the twelfth
century and gives a sense of the “translation” of newly imported eastern (or Islamic) architectural
practices to eastern and central Anatolia and their encounter with locally available materials and artistic
practices.

45 Meinecke points out in particular the resemblence in design and decoration of the Divriği portal to that of the nearly
contemporary Maghak-i ‘Attari Mosque in Bukhara, dated 1178-9, as well as the correspondence between the interior
plans of the two mosques. He suggests that architects of Azerbaijan must have been familiar with architecture of the
eastern Iranian world (Khurasan and Transoxiana), elements of which they then transmitted into Anatolia (Meinecke
1976, 15, n. 40). For a plan and discussion of the Maghak-i ‘Attari, see Gangler et al., 2004, 120, 126, 135. The
transmission of various aspects of eastern Iranian culture into Anatolia is appreciable but the mechanism of such a
transmission (and the possible role of Azerbaijan) remains to be investigated.
46 Tabbaa 2001, 103-24. The initial trigger for muqarnas may have been the tripartite division of the squinch as it
developed in brick architecture in Central Asia towards the end of the tenth century and implemented in Iran and Iraq
through the eleventh and twelfth centuries.
47 For the development of muqarnas in Anatolia, see Ödekan 1977 and Ödekan 2002.
48 On the development of geometry in the late- and post-Seljuk world of the twelfth and thirteenth centuries, see (in
addition to Tabbaa, 2001) Necipoğlu 1999, 97-109. Both authors credit Baghdad as the main engine behind the
development and dissemination of geometric decoration (notably the interlacing strapwork designs and muqarnas).
Though the cultural force exerted by Baghdad on the medieval Islamic world need not be doubted (including the role
it played in the greater elaboration of geometric modes), a future investigation of the particular artistic and cultural
relationships between regions such as Khurasan, Central Asia, Transcaucasia, Azerbaijan, and Anatolia is likely to
reveal modes of transmission (not only for the geometric mode but also for other artistic, cultural and social patterns)
that are fairly independent of Baghdad as a center or even a catalyst.
49 For the development of glazed bricks and tiles on the exterior of buildings in the greater Iranian world and Anatolia,
see Pickett 1997 and Meinecke 1976.
 The Mosque-Hospital Complex in Divriği: 181

A second monument from the reign of Shāhanshāh shows that just a decade after the construction
of the Kale Mosque, royal building activity in Divriği began to employ stone as the exclusive material of
construction and decoration. The so-called Sitte Melik Tomb in Divriği, built in 1193-4 by Shāhanshāh
just two years prior to his death in 1195-6, is especially noteworthy for the high quality of its stone
construction and decoration as well as its extensive inscriptions (fig. 11)50. The low-relief geometric
carving within the rectangular frame of the portal and the small but properly articulated muqarnas hood
above the entrance provide evidence for craftsmen who could execute the most current design elements
of Islamic architecture now entirely in stone51. These elements are also displayed on the exterior cornice
of the Sitte Melik comprising a course each of muqarnas and low-relief geometric decoration, crowned
by a long inscriptional band. Low-relief geometric decoration is also used to frame two elongated niches
set into the walls adjacent to either side of the entrance. The elongated niches themselves, however, point
in the direction of Georgian or Armenian architecture52. The Sitte Melik thus represents, among the
surviving architecture of the Mengujekids, a new phase of synthesis and qualifies Divriği as a significant
nexus in a diffuse regional network of design and praxis53.
Such a synthesis was probably deliberately sought by Shāhanshāh whose sequence of titles in the
extensive cornice inscription of the Sitte Melik as well as above the entrance of the Kale Mosque and on
its minbar speak volumes on his multi-directional mode of political self-identification54. Accordingly,
Shāhanshāh styled himself with an extensive combination of Turkic, Persian and Islamic terms of power
and grandeur and defined his locus of supremacy as Rūm, Shām and Arman. In addition making a
statement about territorial ambition, the naming of these three regions in the inscription delineates the
particular intersection of political and ethno-cultural space in Shāhanshāh’s vision that also informs the
building of the Sitte Melik itself: located in Rūm but partaking of the architecture and decoration of both
Shām (i.e. Islamic forms such as muqarnas) and Arman (i.e. forms derived from Armenian and Georgian
architecture of Transcaucasia and eastern Anatolia)55. As a contemporary of Bahrāmshāh in the late
twelfth century, Shāhanshāh clearly did not lag too far behind his politically more senior counterpart in
Erzincan in terms of the construction of his image based on the intersection of different cultural spheres.
In this regard, the Sitte Melik may also be said to anticipate the Mosque and Hospital Complex which
represents yet another phase.

50 Berchem and Edhem 1917, 63-9, Önkal 1996, 37-42, and Sakaoğlu 2004, 389-97. There are two dates given on the
tomb: 1) 590/1193-4 at the end of the cornice inscription and 2) 592/1195-6 above the entrance. It is assumed that the
earlier date refers to the year of construction while Shāhanshāh was still living and the latter to the year of
Shāhanshāh’s death.
51 Although Ögel 1987, 6-7 finds the muqarnas hood here to be “hesitant,” it does represent a much more confident
attempt than the one in the Kale Mosque.
52 Such elongated niches are to be seen typically on the exterior eastern façade (one on either side of the apse) of
Georgian and Armenian churches from the tenth century. The pairing of the niches at the Tomb of Sitte Melik, on
either side of the entrance, further recalls this symmetrical placement. Other examples of such a placement on
Anatolian tomb chambers include the Mama Hatun in Tercan (Erzurum) datable to the early thirteenth century.
53 The Tomb of Sitte Melik is one of two tomb chambers built just around the same period in Divriği. The second one,
known as the Tomb of Kamereddin, bears the date of the 20th of Sha‘bān 592 (July 19, 1196) and identifies it as
belonging to a high-level official named Qamar al-Dīn with the title of ḥājib (royal chamberlain); see Berchem and
Edhem 1917, 62-3, Önkal 1196, 42-6 and Sakaoğlu 2004, 398-402. This tomb has the same high-quality stone
masonry as the Sitte Melik but with an entirely different and plainer approach to decoration which suggests that the
decorative elements of the Sitte Melik were reserved exclusively for the royal image of the ruler himself.
54 For the cornice inscriptions of the Sitte Melik, see Berchem and Edhem 1917, 64-7, Önkal 1996, 40-1, and Sakaoğlu
2004, 392-3. For the portal inscriptions of the Kale Mosque, see Berchem and Edhem 1917, 57-8 and Sakaoğlu 2004,
228. Of the minbar of the Kale Mosque, only a single inscriptional fragment remains with titles of Shāhanshāh; see
Sakaoğlu 2004, 229-30.
55 It is worth noting that Shāhanshāh seems to have been the only Mengujekid ruler of Divriği who minted copper coins
in his name. His numismatic titles are low-key in comparison with his titles on architecture; Sakaoğlu 2004, 136-41.
This probably has to do with the limited epigraphic space of coinage. Of the eight types of copper coinage identified,
the supremacy of the Seljuk sultans was acknowledged only on two of them.
182 OYA PANCAROĞLU

Divriği: Architectural Venture in the Period of Seljuk Ascendancy

Shāhanshāh’s ambitious architectural-epigraphic statement in the Sitte Melik should be seen in the
specific political climate of the 1190s, a decade which saw a number of important events in the political
configuration of Anatolia and the neighboring regions (Cahen 1988, 55-65). The passage of the Third
Crusades under the leadership of Frederick Barbarossa in the first half of 1190 amounted to a major
challenge for the Seljuks who had recently become a politically fragmented dynasty after sultan Qılıch
Arslān II divided the territories among his numerous sons sometime before 1188. As the Seljuks princes
clashed with each other in a bid to gain sovereignty, a state of virtual civil war threatened the Seljuk state
until 1197-8 when Rukn al-Dīn Sulaymānshāh finally ascended the throne in Konya, sending Ghiyāth al-
Dīn Kaykhusraw into exile. Further east, the same decade also saw the definitive downfall of the Great
Seljuk empire in Iran, Iraq and Syria in 1194, triggering a subsequent scramble for power among its
successor states, notably those in Azerbaijan. From the vantage point of Divriği in 1194 when the Sitte
Melik was built, Shāhanshāh could not have failed to see this near-dissolution of political authorities in
the region as an opportunity for local Mengujekid self-aggrandizement. The Tomb of Sitte Melik could
be seen as a result of this perception of opportunity and an embodiment of the multi-referential system of
self-identification cultivated by the Divriği Mengujekids in their particular geopolitical and cultural
position.
This method of self-identification sensitive to the conjunction of geohistorical circumstances
impacting the Mengujekid realms may also be detected in the Mosque-Hospital Complex of Divriği, built
some three decades after the Tomb of Sitte Melik. As in the case of the Sitte Melik, the events recorded
around the time of the building of this monument were largely external to the Mengujekids of Divriği. At
the turn of the thirteenth century, the Saltukid dynasty based in Erzurum had been already terminated by
the Seljuk sultan Rukn al-Dīn Sulaymānshāh. The early decades of the thirteenth century saw the second
rise of the Seljuks under the sultans ‘Izz al-Dīn Kaykāwūs (r.1211-19) and ‘Alā al-Dīn Kayqubād (r. 1219-
37) who consolidated and significantly expanded the power of the dynasty now with strongholds both on
the Black Sea and the Mediterranean (Cahen 1988, 66-89). Until his death in 1225, Fakhr al-Dīn
Bahrāmshāh carried out a consistent policy of alliance with these sultans and maintained a general attitude
of deference to their rising star. However, the Mengujekid territories of Erzincan, Kemah and
Şebinkarahisar (Koloneia/Kughūnya) were decisively annexed by ‘Alā al-Dīn Kayqubād in 1228—the
same year the Divriği complex was completed—following Dāwūdshāh’s unsuccessful attempt to forge a
new set of alliances.
The coincidence of the political termination of the Mengujekids of Erzincan-Kemah-Şebinkarahisar
and the construction of the grand Mosque-Hospital Complex in Divriği is undoubtedly significant. The
ascent of the Seljuks and its inevitably detrimental impact on the fortunes of neighboring polities could
not have been lost on Aḥmadshāh. This is evident in the formulation of the main foundation inscription
of the mosque where Aḥmadshāh clearly avoided any overt term of self-aggrandizement or self-praise
(see Appendix). While the intention of the patron may have been simply to communicate a sense of
modest piety not inappropriate for a place of communal worship, it is also probable that this epigraphic
reticence served a purpose of inter-dynastic political discretion in the period coinciding with the
annexation of Erzincan-Kemah-Şebinkarahisar by the Seljuks. Indeed, the element of political discretion
in the formulation of the foundation inscription is revealed by the second inscription of this portal, placed
and written rather awkwardly—almost certainly as an addendum—above the arch of the entrance, to
acknowledge the final dominance of ‘Alā al-Dīn Kayqubād over the Mengujekids as signaled by his
termination of the senior branch of the dynasty (see fig. 12 and Appendix)56. A seeming equilibrium

56 It must be said that this additional inscription cannot be dated with certainty. It may have been placed there anytime
between 1228 and 1237 (the year in which ‘Alā al-Dīn Kayqubād died).
 The Mosque-Hospital Complex in Divriği: 183

between Muslim piety and political discretion is further reflected by two small inscriptions, hardly visible
at first glance, placed to the right and left of the entrance (on the capitals of the two engaged columns)
which combine a formulaic benediction on an unnamed ruler (malik) with the beginning of the well-
known Throne Verse from the Qur’an on the oneness and power of God (see Appendix). Finally, the
inscription over the entrance of the east portal which was probably reserved for the access of the royal
party consists of a slight variation on a Qur’anic verse which asserts that dominion (al-mulk) belongs to
God alone (see Appendix)57.
If Aḥmadshāh’s inscriptions on the mosque make a case for the tenuous nature of power balances
impacting the Mengujekid dynasty in the late 1220s, Tūrān Malik must have been even more acutely and
personally aware of the developments which undoubtedly cast a gloomy shadow on the prestige once
enjoyed by her own senior branch of the dynastic family. There is an allusion to that prestige in the
inscription of her hospital (see Appendix) dated just a few years after the death of her father in 1225 and
in the same year when her two brothers—Dāwūdshāh of Erzincan and Muẓaffar al-Dīn Muḥammad of
Şebinkarahisar—were removed from power and sent into exile in Seljuk territory. Interestingly, Tūrān
Malik’s foundation inscription was not supplemented with any acknowledgment of the suzerainty of ‘Alā
al-Dīn Kayqubād, which may perhaps have to do with the fact that a hospital, unlike a mosque, is not
directly linked to the idea of sovereignty by means of the khuṭba (requisite Friday sermon given in the
mosque) in which the holder of political authority is declaimed. Despite the real and psychological
pressure exerted on them by the Seljuks, both patrons clearly had significant resources at their disposal in
Divriği to undertake such a grand project. The investment of these resources in the form of a charitable
architectural project with a legal endowment could well have been conceived to circumvent the imminent
possibility of the loss of their property and possessions in the face of uncertainty posed by the Seljuk
threat.
In the years immediately following the construction of the Mosque-Hospital Complex, Aḥmadshāh
must have had to maintain the low political profile he first expressed on the mosque portal inscription.
Indeed, after 1228, the Mengujekids of Divriği existed as a tiny pocket of dynastic difference, divested of
any claim to real power, surrounded on all sides by formally Seljuk domains, and, as before, entirely
invisible to the medieval writers of history58. The early 1230s saw the zenith of Seljuk political and
military ascendance in Anatolia, marked especially by the decisive Seljuk-Ayyubid victory against the
troublesome and persistent Khwārazmshāh threat in eastern Anatolia. Although it is not possible to chart
the course of the next decade for the Mengujekids of Divriği in any degree of detail, the fact that they
ultimately weathered the Seljuk storm of the late 1220s and early 1230s is made evident by three more
inscriptions of Aḥmadshāh in Divriği: two from the citadel, dated 634 (1236-7) and 640 (1242-3),59 and
a third one from the minbar of the mosque dated 638 (1240-41) (see Appendix). In these, the name of
Aḥmadshāh is proclaimed with a series of titles and honorifics which were conspicuously missing from
the portal dated 1228-9 and, furthermore, there is no mention of Seljuk supremacy. This is explained by
the fact that ‘Alā al-Dīn Kayqubād had been succeeded in 1237 by his son Ghiyāth al-Dīn Kaykhusraw
II (r. 1237-46) soon after which Seljuk central authority was seriously undermined during the course of
the violent socio-religious Bābā’ī uprising which was suppressed only after much difficulty in 1240 (Ocak
1980). These events almost certainly afforded the Mengujekids of Divriği a reprieve from the pressures
of Seljuk territorial expansionism but the actual termination of the dynasty, much like its inception, is not
recorded in the sources. The fact that the extant redaction of the endowment deed of the mosque is dated

57 As mentioned in the beginning of the article, the west portal of the mosque is assumed to be a later (i.e. post-1228-9)
construction and therefore the titles of Aḥmadshāh mentioned there are not taken into consideration in this assessment.
58 Historically, the mountainous topography of Divriği and its formidable citadel gave it an advantage as a shelter for
political difference. In the ninth century, for example, Tephrike/Divriği became the base of the Paulicians who asserted
their independence from the Byzantines; see Foss 1991.
59 Berchem and Edhem 1917, 88-9 and Sakaoğlu 2004, 204-5.
184 OYA PANCAROĞLU

just two days after the defeat of the Seljuk army by the Mongols at the Battle of Kösedağ near Sivas
suggests that a sense of urgency may have gripped the patrons in the immediate aftermath of this major
event. Given the incomplete aspect of parts of the complex’s decoration, it is possible that the drawing up
of the endowment deed had been postponed for some time60. Remarkably, it appears that the dynasty made
it well past the Mongol invasion of Anatolia in 1243: the latest Mengujekid inscription from citadel of
Divriği belongs to Aḥmadshāh’s son and successor, al-Ṣāliḥ, and is dated to the first of Sha‘bān 650
(October 7, 1252)61.

Two Roads to Divriği: Artists from Ahlat and Tbilisi

Thus, much had changed around Divriği between 1228 and 1240 which allowed Aḥmadshāh to
maintain his local power until sometime in the mid to late 1240s or even the early 1250s and pass it to his
son. In the third, fourth and fifth decades of the thirteenth century when Aḥmadshāh ruled in Divriği, the
Mengujekids continued to have access to a vibrant network of cultural and artistic production. The
geographic orientation of this network is indicated by the signatures of two artists in the Mosque-Hospital
Complex62. The first signature is that of Khurramshāh ibn Mughīth al-Khilāṭī, found on the interior
masonry of both the mosque and the hospital63. The nisba of the artist, al-Khilāṭī, indicates Ahlat, located
on the northwestern shore of Lake Van, as the place of his personal or family origin, while his patronymic
(Mughīth) and personal name (Khurramshāh) suggest that he came from a Muslim family with a Persian
linguistic or cultural background. The second signature is inscribed on the 1240-1 minbar of the mosque,
identifying its artist as Aḥmad ibn Ibrāhīm al-Tiflīsī and pointing to the Georgian city of Tbilisi for the
master’s place of origin64. Although no further information exists for either of these craftsmen, their
places of origin merit consideration within the sphere of political and cultural relations which
encompassed the Mengujekids in the late twelfth and early thirteenth centuries.
Tbilisi, the historic capital of eastern Georgia, came under Arab-Muslim control as early as the
eighth century and remained as an outpost of the Abbasid Caliphate, surrounded by Armenian and
Georgian states, until it was retaken by the Georgians under King David IV in the twelfth century (1122)65.
Soon after the accession of the Georgian queen Rusudan to power in 1223, the city was occupied by the
Khwarazmshahs in 1226 and taken by the Mongols in 1236. The latter two events were undoubtedly
disruptive for social and cultural continuity in the city and it is, therefore, not entirely surprising to find
an artist like Aḥmad ibn Ibrāhīm seeking opportunities for employment in a haven such as Divriği around

60 Of course, this assumes that the date on the extant redaction of the endowment deed of the mosque is historically
accurate (see footnote 6 above).
61 Berchem and Edhem 1917, 89-90 and Sakaoğlu 2004, 205.
62 A third signature which is on the east portal of the mosque is unfortunately too damaged to be read beyond the first
name of the artist, Aḥmad (see Appendix); Berchem and Edhem 1917, 79 and Sakaoğlu 2004, 314.
63 The signature ‘amal-i (“the work of”) Khurramshāh ibn Mughīth al-Khilāṭī can be clearly read on the keystone of the
western arch of the dome in front of the mihrab in the mosque. The signature in the hospital, located high on the eastern
wall of the main vaulted hall opposite the entrance, has been read as ‘amal-i Khūrshāh (or Khūrshād) Akhlāṭī which
has caused some to assume that two different artists were involved (Sakaoğlu 2004, 380). In fact, the orthographic
differences between the two names (Khurramshāh and Khūrshād/ Khūrshāh) in the Arabic script are slight enough (at
least with regard to the basic letter forms involved) to suggest that the latter signature contains a writing mistake and
that both signatures denote a single artist. Another issue with the signature (which is by no means unique to Divriği)
has to do with the fact that it does not convey a precise notion about the artist’s function in the project: was he a
builder/architect, stonemason, or a supervisor? The signature formula of ‘amal-i (“the work of”) denotes first and
foremost the role of the master artist but does not rule out specific skills such as stone carving.
64 Berchem and Edhem 1917, 81-2, Oral 1962, 43-9, and Sakaoğlu 2004, 273—5, 316-23. It has been suggested that
this Aḥmad may be the same artist as the one who signed the east portal (see footnote 72); Gabriel 1934, 188 and Ögel
1987, 25. Given that both works feature geometric decoration, this is within the realm of possibility but cannot be
proven as the remainder of the east portal artist’s name has defied attempts to reconstruct it.
65 Thomson 1991 and Minorsky [Bosworth] 1986.
 The Mosque-Hospital Complex in Divriği: 185

this time. Given the apparent absence of art historical record concerning Islamic art and architecture
produced in medieval Tbilisi, however, it is not possible to assess the specific contribution of this artist’s
background to the work he produced for the mosque in Divriği. Nevertheless, Tbilisi was specifically
remarked by the ninth-century historian al-Balādhurī for the extensive wooden (pine) architecture in the
built environment of the early medieval city (Minorsky [Bosworth] 2000). The minbar of Divriği, with
its complex surface geometry of interlacing star patterns and extensive pious inscriptions, is indeed
among the finest of its type known from the medieval Islamic world. Tbilisi’s contribution to Islamic
culture and learning is further evinced by the work of the renowned polymath, Ḥubaysh ibn Ibrāhīm al-
Tiflīsī (d. ca. 1203-4) who attached himself to the court of sultan Qılıch Arslān II and made his career in
Anatolia during the first rise of the Seljuks (Yazıcı 2004). Given that Ḥubaysh al-Tiflīsī’s books are
among the earliest erudite works known to have been written in the Seljuk realms of Anatolia, it may be
assumed that he acquired his scholarly training in Tbilisi.
More can be said about Ahlat (medieval Khilāṭ), the city located on the northwestern shore of Lake
Van, from which the architect Khurramshāh hailed66. Ahlat was already a prosperous commercial and
cultural center in the eleventh century when it was part of the Marwanid dynasty (990-1085) of Diyār
Bakr, the region which included, in addition to Ahlat, Āmid (modern Diyarbakır), Mayyāfāriqīn (modern
Silvan), and Nusaybin. Nāṣir-i Khusraw, the Persian Isma‘ili poet-philosopher, visited the city in 1046
and reported on its predominantly cash economy as well as its notable trilingual culture comprising
Arabic, Persian, and Armenian (Nasir-i Khusraw 1986). The prevalence of these three languages in the
eleventh-century reveals Ahlat’s location at a point of political, social, and cultural intersection. These
intersecting identities were also acknowledged by the next rulers of Ahlat, a Turkmen dynasty descended
from a mamlūk (slave) emir of the Great Seljuks named Sukmān al-Quṭbī. Known officially on their
coinage and in the contemporary sources by the dynastic name of the Shāh-i Arman (“King of
Armenians”), this Turkmen state ruled over Ahlat and the neighboring region between 1100 and 1207.
During this time, the Georgians posed the greatest threat to their claim on this region which induced the
Shāh-i Arman to join a coalition of Muslim forces from eastern Anatolia and Azerbaijan and take part in
military expeditions against Georgia. Ultimately, in 1207, it was the Ayyubids of Syria who took Ahlat
from the Shāh-i Arman but the city continued to be the target of Georgian expeditions through the 1210s
and of the Khwarazmshahs in the 1220s.
Medieval sources record a devastating earthquake in the region of Ahlat in 674 (1275-6) which
destroyed most of the buildings in the city (Sümer 1990, 57). The standing medieval architecture of Ahlat
is mainly from the Mongol (Ilkhanid) period and dated after this earthquake. Nevertheless, the existence
of a vibrant environment of artistic production in Ahlat can be discerned from the hundreds of
monumental tombstones that survive in its main medieval Muslim cemetery dated predominantly
between the twelfth and fourteenth centuries (Karamağaralı 1972) (Fig. 13). The outstanding stone
carving on these tombstones with their brilliant geometric and stylized vegetal designs reveal the intensity
and quality of artistic activity in medieval Ahlat. While the twelfth-century tombstones consist largely of
grave markers in the form of cenotaphs with a stepped form, there was, during the last two decades of the
same century, an apparent sudden change towards the employment of monumental stelae for marking the
head and foot of the grave with a rectilinear or cylindrical marker lying in between. The earliest of such
stelae (from the 1180s and 1190s) are decorated with stylized vegetal patterns which gave way, in the
early thirteenth century, to examples with very high quality workmanship and sophisticated designs
consisting most notably of geometric ornament based on interlacing star patterns. From the second half
of the thirteenth century, many Ahlat tomb stelae were also capped with an imposing cornice-like feature
which has an inward curve frequently decorated with muqarnas.

66 On the medieval history of Ahlat, see Minorsky and Taeschner 1960, Sümer 1990, 47-56, 65-84, and Turan 2001,
101-42.
186 OYA PANCAROĞLU

In terms of their rectangular shape, imposing height, fine carving, and especially distinctive cornice
at the top, thirteenth-century Ahlat tomb stelae visually evoke the Armenian commemorative stele known
as khatchk’ar (“stone cross”) (Fig. 14). Khatchk’ars, produced throughout eastern Anatolia and
Armenian Transcaucasia starting as early as the ninth century but peaking especially between the twelfth
and the fourteenth centuries, typically feature a large and elaborate foliated “living cross” above a round
medallion symbolizing Golgotha (both often contained in an arched niche) and usually framed by a dense
matrix of stylized vegetal and geometric decoration (Donabédian 2007). From the end of the twelfth
century, khatchk’ars came to be distinguished by an inward curving cornice, a peculiar feature which may
have been introduced to protect the decoration below from weathering. Although very few khatchk’ars
have been preserved in modern Turkey, a handful of surviving examples from near Ahlat and Erzincan
give an idea about the refinement and distribution of this art form in eastern Anatolia which seems to have
inspired the makers of the Ahlat Muslim tomb stelae to synthesize the overall form of the Armenian
khatchk’ar with the ornamental and epigraphic disposition of the Muslim funerary marker.67
From the dates on the Ahlat stelae, it is possible to determine a significant interruption in the
production of such stelae during the second quarter of the thirteenth century (Karamağaralı 1972, 44-5).
This interruption may be attributed primarily to the siege of the city by the Khwarazmshahs under the
leadership of Khwarazmshāh Jalāl al-Dīn Mangubartī, first in the late autumn-winter of 1226 and again
in 1229-30 (Sümer 1990, 55-6). The second siege lasted for a relentless eight months, causing a severe
famine that nearly annihilated the population between August 1229 and April 1230. Further devastation
was brought on by a particularly destructive three-day plunder by the soldiers of the invading army once
they entered the already battered city. The disruption documented by the tomb stelae attest to the
likelihood of the departure of at least some of Ahlat’s craftsmen already prior to the final blow of 1229-
30 (possibly around the time of the first Khwarazmshah attack in 1226) in search of more secure centers
of patronage. In all likelihood, Khurramshāh numbered among those craftsmen who sought their fortunes
away from the continuous onslaughts of the 1210s and 1220s.
Most of the Ahlat tomb stelae bear artists’ signatures which indicate the prestige associated with
their production and the importance given to identifying the makers (and probably their workshops)
(Karamağaralı 1972, 82-103). Analysis of these signatures through the thirteenth century also reveal an
emphasis placed on master (ustādh) and apprentice (ghulam, shāgird) relationships and lineage. This
particular emphasis evokes the principles of the futuwwa organization—medieval urban brotherhoods
promoting chivalry and mystic spirituality based on a system of initiation with distinct stages of maturity
through which the new recruits would be gradually guided.68 Most members (known as ākhīs) of the
futuwwa organizations in Anatolia cities belonged to professional backgrounds in commerce and craft
who continued to practice their trade while regularly applying themselves to the ritual and communal
precepts of the organization. Ahlat, with its vibrant commercial activity, is known to have been an
important and early gateway for the westward transmission of the grassroots futuwwa movement which
appears to have emerged in eastern Iran and eventually spread into western Iran, Azerbaijan and Anatolia.
The futuwwa movement was established in Anatolia in the thirteenth century and thrived through the
fourteenth and fifteenth centuries, creating a network of social cohesion through the provision of
hospitality and assistance to the displaced, especially during times of political uncertainty. The fact that
two of the several futuwwa guidebooks composed in Anatolia were penned by artists—a gilder (zarkūb)
and a painter (naqqāsh)—reflects the prominence of craftsmen among the ranks of the futuwwa.69
Although the signatures on the Ahlat tomb stelae only hint at (and do not confirm) the association of these

67 Two of the most monumental khatchk’ars, measuring around six meters in height and dated 1191 and 1194, are
preserved in the vicinity of the monastery of Aprank, located midway between Erzincan and Erzurum
(http://www.virtualani.org/aprank/index.htm).
68 Cahen 1988, 315-20, Ocak 1996, and Gölpınarlı 1949-50.
69 These two authors are Aḥmad ibn Ilyās al-Naqqāsh al-Khartburtī and Najm-i Zarkūb; Gölpınarlı 1949-50, 13, 14.
 The Mosque-Hospital Complex in Divriği: 187

artists with futuwwa organizations, the re-emergence of their craft after the Khwarazmshah disruption
could well be attributed to the workings of just such an organization providing the network and support
needed for the process of renewal. Khurramshāh’s journey of relocation to Divriği could also have been
facilitated by the Anatolian futuwwa, which, by the end of the thirteenth century, had made such great
impact on Muslim society that it was also adopted (with some adjustments) by the Armenian population
of Erzincan (Goshgarian 2007).

Artists in Motion, Designs in Transition

Although the Ahlat tomb stelae show that the events of the 1220s had a negative effect (albeit
temporarily) on the art of funerary stone carving which had just reached a certain high point in its
development since the late twelfth century, they remain silent on the question of their direct visual
relevance for the stone carving on the Mosque-Hospital Complex of Divriği. Certainly, the high quality
of the carving on the Ahlat stelae match the level of the work in Divriği, but the low-relief and
predominantly geometric decoration of the former cannot explain the high-relief and stylized vegetal
motifs that distinguish the latter. In the absence of architecture predating the Mongol period in Ahlat, the
relevance of Ahlat for Divriği can be perceived from the probable workings (rather than the actual
products) of artistic activity and synthesis both before and after the Khwarazmshah devastation. In other
words, it is the processes of visual design and synthesis rather than any extant work of art or architecture
from Ahlat which shed some light on the means by which the seemingly unusual decoration on the Divriği
portals came to be.
Among the Ahlat signatures are two names which yield important clues concerning a particular
aspect of artistic activity. The first clue is furnished by an artist named Qāsim ibn Ustādh ‘Alī who
signed two tomb stelae at the end of the fourteenth and the beginning of the fifteenth centuries
(Karamağaralı 1972, 102, 240-2). The same artist also signed the Tomb of Erzen Hatun in Ahlat dated
1396-7. Although not contemporary with the Mosque-Hospital Complex of Divriği, both the tomb
stelae and the tomb chamber signed by the artist are very much within their respective traditions and
his crossover between two types of work probably reflects an earlier practice of artistic continuum.
Thus, it seems that the art of the stone carvers of tomb stelae could also extend to architecture where
they could be identified as the primary artist. The second clue emerges from the signature of an artist
named Aḥmad al-Muzayyin who was active between the 1240s and the 1260s (Karamağaralı 1972,
91-2, 130-1, 146-53). The professional designation of muzayyin may be translated as “decorator”
which, in the context of the tomb stelae, refers specifically to the decoration executed in stone carving.
However, given that this professional designation is unique to this artist among the Ahlat signatures
and, furthermore, that his decoration represents a series of innovations among the stelae of the middle
of the thirteenth century, it is most likely that Aḥmad al-Muzayyin’s work was not limited to the
carving of tomb stelae but included similar design work in other media. This is all the more probable
since the primary designs on Aḥmad al-Muzayyin’s tomb stelae consist of interlacing star patterns, a
mode of geometric design which is most notable for its contemporary application on a wide variety
of media from stone and wood to metal and paper, manifesting itself on architecture, furniture, objects
and manuscripts. As his works represent the revival of the art of tomb stelae after the hiatus that
corresponds to the period of the Khwarazmshah and Mongol invasions, Aḥmad al-Muzayyin’s role in
the rejuvenation of stone carving probably derived its force from his skills as designer. Although not
numerous, these particular Ahlat signatures point towards a long-term artistic environment in which
design constituted the main platform of activity and the possibilities of its application were not
restricted by any particular medium or craft. It may be surmised that the design process in places like
Ahlat turned on the pivotal role played by the development of patterns on paper which could be scaled
up or down or otherwise adjusted to match the craft and work at hand.
188 OYA PANCAROĞLU

The artists who were responsible for the refined art of funerary stelae in Ahlat were thus not only
versed in the craft of stone carving but also fully in command of an autonomous design process. This
autonomy could be seen as a function of the prestige enjoyed by artists working within a network of
patronage reinforced by the futuwwa organizations. In circumstances of mobility, whether voluntary or
involuntary, artists who were trained in the autonomy of the design process must have had a clear
advantage of flexibility and innovation. Seen from this perspective, the work of Khurramshāh in Divriği
begins to come into focus as the product of a process of design which was not contingent upon established
medium-specific practice. In other words, Khurramshāh either created or procured designs on paper
which he then adapted for architectural application. Similar to Qāsim ibn Ustādh ‘Alī and Aḥmad al-
Muzayyin (and doubtless many other artists of Ahlat), Khurramshāh must have been a designer who could
adjust and apply his design work to different scales or media. This observation is supported by the
evidence of artistic activity in Ahlat which forms Khurramshāh’s background. It is further strengthened
by the finding of a number of incised designs, or graffiti, consisting of concentric and interlocked circles
on the masonry of the Mosque-Hospital Complex70. It has been already convincingly shown that at least
one of these graffiti provided the reduced-scale geometric blueprint from which both the elevation of the
north portal of the mosque and its central geometric motif of the hexagon could be generated to
proportion. These graffiti were undoubtedly transferred to stone from paper as on-site drawings
mediating the design process from start to finish.
It has already been recognized that the unique decorative language of the Divriği Mosque-Hospital
Complex must be the product of a synthesis of vocabulary derived from a multiplicity of sources (Kuban
1997, 195-6). It is this synthesis which has been described as a “miracle.” Identifying the precise sources
that provided the visual ingredients of this artistic synthesis is itself a feat that would qualify as a
“miracle” by any standard of art historical methodology71. The challenge of such a task lies not in
discovering ever more comparative visual material but in ascertaining their relevance and cohesion when
tested against the totality of the monument. Inevitably, any attempt to unravel the synthesis in this sense
exposes not only the shortcomings of the comparative method but also runs against the insurmountable
problem of “silences” in the visual record as a result of natural and human destruction over the centuries.
Certainly, the comparative method does yield notable correspondences in seemingly discrete sets of
artistic vocabulary without, however, providing a complete glossary for the entire composition in
question.
The aim of this essay was to consider the development of other syntheses which can be discerned
in the historical process leading to the construction of the Mosque-Hospital Complex of Divriği. The
political and cultural history of the Mengujekids reveals the workings of a nexus of long-distance relations
from which the image of both the dynasty and the buildings it constructed was derived. Located on the
western cusp of a large region in which the geohistorical conjunction of Anatolia, Iran, and Transcaucasia
had a long and rich history, Divriği, by virtue of the political circumstances of the late twelfth and early

70 Bakırer 1999. Further discoveries of grafitti (mostly of circle-based geometric designs) on the masonry of both the
hospital and the mosque have lent strength to this remarkable insight into the workings of the design process at Divriği;
see Bakırer 2001. For another instance of the use of on-site incised designs (in this case for the construction of
muqarnas vaulting in an early thirteenth-century Armenian church assembly hall), see Ghazarian and Ousterhout
2001.
71 A frequently cited comparanda for the high-relief vegetal decoration of Divriği is the stucco decoration of the tomb
known as the Gunbad-i ‘Alawiyyan in Hamadan, western Iran. The building bears no date and opinion has been
divided on whether it should be dated to the late Seljuk period (late twelfth century) or to the Ilkhanid period
(fourteenth century). The earlier date has been claimed by Shani 1996 but this study has a number methodological
problems and examples of such stucco are more numerous from the fourteenth century. In any case, Gunbad-i
‘Alawiyyan’s similarity to Divriği in terms of motif and composition is only partial and any genetic link between the
two (regardless of the date of the former) must derive from the dissemination of such designs on paper rather than a
direct transfer between the practices of stucco and stone carving.
 The Mosque-Hospital Complex in Divriği: 189

thirteenth centuries, was afforded the right conditions for artistic synthesis. Recognizing the role of this
geohistorical conjunction in the history of the Mengujekids is a necessary step in gaining a vision,
however restricted it may be, into the dynamism of the process of design and production represented most
informatively by the mobility of the artists. The cultural encounters of this geohistorical conjunction often
resulted in overlap or synthesis and were built on an extensive foundation of artistic relations which, in
turn, were typically made possible by the political vision for the survival of dynastic entities. Among the
most notable examples of such artistic relations is the interchange between Armenian and Islamic visual
cultures as potently exemplified by the tenth-century Church of the Holy Cross at Aghtamar on Lake Van,
the doors of the Church of the Holy Apostles at Muş (to the west of Ahlat) dated 1134, the khatchk’ar
form of the Ahlat tomb stelae, and the numerous early thirteenth-century examples of the adaptation
muqarnas vaulting in the architecture of the gavit, the assembly hall appended to Armenian churches72.
Given the particular intensity of social mobility in this region in the early thirteenth century, it is most
likely that builders such as Khurramshāh found themselves in just such a dynamic network of design and
craftsmanship that transcended local idioms and encouraged the merging of elements from a variety of
media and across cultures. Although the actual mechanisms of the formulation, transmission, and
application of designs in the medieval period are only partially discernible today, the conditions for cross-
pollination could not have been more favorable. Built at a particular historical junction when the ebb and
flow of political forces around and beyond Divriği had gathered pace, the Mosque-Hospital Complex
testifies to the transient creation of a hub of artistic ingenuity in the last bastion of Mengujekid identity.
Much like the quiet that defines the eye of a storm, this monument appears to be embedded in a cloud of
silence yet has much to say about the circumstances surrounding its conception.

Doç. Dr. Oya PANCAROĞLU

Boğaziçi University

Appendix:
Inscriptions with References to Rulership
from the Mosque-Hospital Complex of Divriği

North portal of the Mosque:

1. Main foundation inscription

This congregational mosque was ordered to be built for the sake of God Almighty by the servant
in need of God’s mercy, Aḥmadshāh son of Sulaymānshāh. May God perpetuate his sovereignty. On
the date of the year sixhundred and twenty-six.

72 On Aghtamar which has long been recognized for its inclusion of Islamic iconography of rulership, see Jones 2004.
The Muş doors (now preserved in the Museum of Armenian History, Erivan) feature a central field of Islamic
interlacing star patterns and a frame of running animals and equestrian figures embedded in a vine scroll; see Durand
et al. 2007, 151-2. One of the artists who signed it, Lukas, was qualified as a “painter” which once again points to the
role of designs on paper in the transmission and synthesis process. On the use and adaptation of muqarnas in
Armenian architecture, see Ghazarian and Ousterhout, 2001.
190 OYA PANCAROĞLU

2. Secondary added inscription

In the days of the reign of the great sultan ‘Alā al-Dunyā wa al-Dīn Kayqubād son of
Kaykhusraw, partner of the Commander of the Faithful.

3. Inscriptions on the right- and left-hand column capitals

May God continue the rule of the just king and [give] to him happiness and pleasure.
God, there is no god but He, the Living, the Eternal. [Beginning of Qur’an 2:255, Throne Verse (āyat
al-kursī)]

West portal of the Mosque: Foundation inscription

The first to build this blessed congregational mosque, longing for God Almighty’s benevolence
(was) the weak servant in need of God’s mercy, Aḥmadshāh son of Sulaymānshāh son of Shāhanshāh,
helper of the Commander of the Faithful, may God perpetuate his sovereignty and increase his power. In
one of the months of the year sixhundred and twenty-six.

East portal of the Mosque

Sovereignty belongs to God, the Unique, the Victorious. The work of Aḥmad ….

Minbar of the Mosque

The learned and just king, Ḥusām al-Dunyā wa al-Dīn Aḥmadshāh son of Sulaymānshāh son of
Shāhanshāh, the helper of the Commander of the Faithful, may God perpetuate his sovereignty, ordered
the setting up of this blessed minbar for the sake of God Almighty, on the date of the year sixhundred and
thirty-eight.

Portal of the Hospital

 The Mosque-Hospital Complex in Divriği: 191

The just queen, in need of God Almighty’s pardon, Tūrān Malik, daughter of the fortunate king
Fakhr al-Dīn Bahrāmshāh, ordered the building of this blessed house of healing, longing for God
Almighty’s benevolence. May God accept it, Amen. In one of the months of sixhundred and twenty-six

Özet
Divriği Ulu Cami ve Darüşşifası: Bir İlişkiler ve Geçişler Tarihi

Anadolu Ortaçağ mimarlık tarihinde “eşsiz” sıfatı ile en sık tanımlanan yapı kuşkusuz 1228-1229
tarihli Divriği Ulu Cami ve Darüşşifası’dır. Yirminci yüzyılın ilk yıllarından beri yerli ve yabancı pek çok
araştırmacının çalışmalarına konu olan bu yapının sanatsal ayrıcalıkları, aynı yüzyılın kapanış
yıllarında—gizem derecesi daha da arttıralarak—en son bir “mucize” bağlamında düğümlenmiştir.
Özellikle taçkapılarında yer alan ve benzerine başka yerde rastlanmayan taş tezyinatı ile ünlenen Divriği
Ulu Cami ve Darüşşifası’nın tarihsel arka planı ise anahatlarının ötesinde irdelenmemiştir. Hatta, bir çok
kez “Selçuklu” kültür çatısı altında ele alınan bu yapının esas banisi olan Mengücekli hanedanının tarihi
üzerinde fazla durulmamış veya bu tarih yorum kapsamının dışında tutulmuştur.
Bu makalenin başlıca amacı, Divriği Ulu Cami ve Darüşşifası’nı Mengücekliler’in onikinci ve
onüçüncü yüzyıl tarihi ışığında tekrar gözden geçirmektir. Bugüne gelebilen bilgiler az da olsa,
Mengücekliler’in değişken koşullar içinde gelişen siyaset ve kültür tarihi hakkında anlamlı veriler
içermektedir. Onikinci yüzyılın ikinci yarısından itibaren Orta Anadolu’dan Güney Kafkaslar ve
Azerbaycan’a uzanan devletler silsilesi içinde oluşan sanat, edebiyat ve bilim faaliyetlerinde önemli bir
yere sahip olan Mengücekliler, Büyük Selçuklu devletinin parçalanmasını izleyen süreçte meydana gelen
bölgesel çalkantılardan ve Anadolu Selçuklu devletinin onüçüncü yüzyılın başlarında yürüttüğü
genişleme politikasından önemli kayıplarla çıkmıştır. Bu kayıpların en başında başkent konumunda olan
Erzincan ile beraber Kemah ve Şebinkarahisar (Köğonya) şehirlerinin Selçuklu sultanı Alaeddin
Keykubad tarafından 1228 yılında ilhak edilişi gelir. Bu olumsuz gelişmelerden sıyrılarak kurtulan
Mengücekliler’in Divriği kolunun ise daha onikinci yüzyılın sonundan itibaren—anlaşıldığı kadarıyla
ikincil siyasi konumun getirdiği prestij ve kimlik arayışı sayesinde—mimarlık sahnesinde yenilikçi veya
bireşimci denebilecek bir çizgi izlemiş olduğu görülmektedir.
Divriği Mengüceklileri’nin 1180-1240 yılları arasında yaptırdıkları eserlere imza atan Meraga,
Ahlat ve Tiflis menşeli sanatçıların, geniş bir coğrafyada etkin olan geç Selçuklu çağı kültür ağı içerisinde
yer aldıklarını ve hamilerinin izledikleri devlet ve kültür siyasetinin görsel ifadesini bu ağın sayesinde
gerçekleştirdiklerini saptamak mümkündür. Divriği Ulu Cami ve Darüşşifası en azından onuncu
yüzyıldan beri etkileşim içinde olan İslam ve Ermeni sanatlarının belli öğelerini özellikle tasarım
aşamasında yeni bir kaynaşma platformuna taşıyarak, yerel Mengücekli hanedanının hem gezgin
sanatçılara sağladığı kültürel ortamı hem de kendisi için belirlemekte olduğu kimlik yapısını ortaya
koymaktadır. Tasarımın uygulamaya geçişinde—ki Divriği’yi “mucivezi” kılan tam da bu geçişdir—
oynadığı rol ile Ahlat’lı sanatçının kendi muhitinde birbirleriyle kesişen sanat eksenlerinin de payını
vurguladığını saptamak önemlidir. Divriği Ulu Cami ve Darüşşifası Anadolu’nun Ortaçağında süregelen
çok merkezli ilişkilerin siyasi ve sanatsal yönlerini yansıtırken, tarihsel bağlamın da bu tür geçişleri
desteklediğini söylemek mümkündür.
192 OYA PANCAROĞLU

Bibliography
Aflaki, S. 2002, Feats of the Knowers of God (Manāqeb al-‘ārefīn), J. O’Kane (trans.), Leiden.
Anonymous. 1999, Tārikh-i Āl-i Saljūq dar Ānāṭūlī, N. Jalali (ed.), Tehran.
Arel, A. 2009, ‘Eski eserlere yeni pazarlar’, Radikal 2 (21/06/2009).
Arıkan, Z. 1991, ‘Divriği Kazası’nın İlk Sayımı’, Osmanlı Araştırmaları 11, 49-71.
Artuk, İ. and Artuk, C. 1970, İstanbul Arkeoloji Müzeleri Teşhirdeki İslâmî Sikkeler Kataloğu, vol. 1,
İstanbul.
Bachmann, W. 1913, Kirchen und Moscheen in Armenien und Kurdistan, Leipzig.
Bakırer, Ö. 1981, Selçuklu Öncesi ve Selçuklu Dönemi Anadolu Mimarisinde Tuğla Kullanımı, 2 vols.,
Ankara.
Bakırer, Ö. 1999, ‘The Story of Three Graffiti’, Muqarnas 16, 42-69.
Bakırer, Ö. 1976, Onüç ve Ondördüncü Yüzyıllarda Anadolu Mihrabları, Ankara.
Bakırer, Ö. 2001, ‘Divriği Ulu Camisi ve Darüşşifası’nda Yapıdan Gelen İzlerin Düşündürdükleri’, in: O.
Eravşar et al. (eds.) I. Uluslararası Selçuklu Kültür ve Medeniyeti Kongresi: Bildiriler, vol. 1, Konya,
73-83.
Balgalmış, A. 1994, ‘Divriği’, Türkiye Diyanet Vakfı İslam Ansiklopedisi, İstanbul, vol. 9, 1994, 452-4.
Berchem, M. and Edhem, H. 1917, Matériaux pour un Corpus Inscriptionum Arabicarum. Troisième partie,
Asie Mineure. Tome premier, Siwas, Diwrigi, Cairo.
Brookshaw, D. P. 2005, ‘Odes of a Poet-Princess: The Ghazals of Jahān-Malik Khātūn’, Iran 43, 173-95.
Bryer, A. 1985, The Byzantine Monuments and Topography of the Pontos, 2 vols., Washington, D.C.
Cahen, C. 1988, La Turquie pré-ottomane, İstanbul.
Chelkowski, P. 1995, ‘Niẓāmī Gandjawī’, Encyclopaedia of Islam, 2nd ed. Leiden, vol. 8, 76-80.
Crowe, Y. 1972, ‘The East Window of the Great Mosque in Divriği’, Art and Archaeology Research Papers
2, 105-13.
De Blois, F. 1997, Persian Literature: A Bio-Bibliographical Survey. Volume V: Poetry of the Pre-Mongol
Period, London.
Demiriz, Y. 1972-73, ‘Mimari Süslemede Renk Unsuru Olarak Kullanılan Keramik Çanaklar’, Sanat Tarihi
Yıllığı 5, 175-208.
Donabédian, P. 2007, ‘Le khatchkar’, in: J. Durand et al. (eds.) Armenia sacra: Mémoire chrétienne des
Arméniens (IVe-XVIIIe siècle), Paris, 153-161.
Durand, J. et al. 2007, Armenia sacra: Mémoire chrétienne des Arméniens (IVe-XVIIIe siècle), Paris.
Foss, C. 1991, ‘Tephrike’, Oxford Dictionary of Byzantium, Oxford, 2025.
Gabriel, A. 1934, Monuments turcs d’Anatolie, Vol. 2, Amasya-Tokat-Sivas, Paris.
Gangler, A. et al. 2004, Bukhara: The Eastern Dome of Islam, Stuttgart.
Ghazarian, A. and Ousterhout, R. 2001, ‘A Muqarnas Drawing from Thirteenth-Century Armenia and the
Use of Architectural Drawings during the Middle Ages’, Muqarnas 18, 141-54.
Godard, A. 1936, ‘Notes complémentaires sur les tombeaux de Maragha’, Athar-e Iran 1, 125-60.
Goshgarian, R. 2007, Beyond the Social and the Spiritual: Redefining the Urban Confraterneties of Late
Medieval Anatolia, (Ph. D. thesis) Cambridge, Mass.
Gölpınarlı, A. 1949-50, ‘İslâm ve Türk İllerinde Fütüvvet Teşkilâtı ve Kaynakları’, İktisat Fakültesi
Mecmuası 11, 3-354.
Gülsoy, E. 1995, ‘XVI. Asrın İlk Yarısında Divriği Kazâsı Vakıfları’, İlmî Araştırmalar 1, 107-30.
Gülsoy, E. and Taştemir, M. 2007, 1530 Tarihli Malatya, Behisni, Gerger, Kâhta, Hısn-ı Mansur, Divriği ve
Darende Kazâları Vakıf ve Mülk Defteri, Ankara.
Ibn Bibi. 1956, El-Evāmirü’l-‘Alā’iyye fī’l-Umūri’l-‘Alā’iyye (ed. A. S. Erzi), Ankara.
Ibn Bibi. 1996, El-Evamirü’l-ala’iye fi’l-umuri’l-ala’iye: Selçuk name (trans. M. Öztürk), 2 vols., Ankara.
Jones, L. 2004, ‘Abbasid Suzerainty in the Medieval Caucasus: Appropriation and Adaptation of
Iconography and Ideology’, Gesta 43/2, 143-150.
 The Mosque-Hospital Complex in Divriği: 193

Karamağaralı, B. 1972, Ahlat Mezartaşları, Ankara.

Kaya, M. 1988, ‘Abdüllatif el-Bağdâdî’, Türkiye Diyanet Vakfı İslam Ansiklopedisi, İstanbul, vol. 1, 254-5.
Kaya, S. 2006, I. Gıyâseddin Keyhüsrev ve II. Süleymanşah Dönemi Selçuklu Tarihi (1192-1211), Ankara.
Kayaoğlu, İ. 1978, ‘Vakfiyenin Değerlendirilmesi’, in: Y. Önge et al. (eds.) Divriği Ulu Camii ve
Darüşşifası, Ankara, 10-1.
Kemali, A. 1932, Erzincan: Tarihî, Coğrafî, İçtimaî, Etnoğrafî, İdarî, İhsaî Tetkikat Tecrübesi, İstanbul.
Kuban, D. 1997, Divriği Mucizesi: Selçuklular Çağında İslam Bezeme Sanatı Üzerine Bir Deneme, İstanbul.
Kuban, D. 2002, Selçuklu Çağında Anadolu Sanatı, İstanbul.
Kuniholm, P. I. 2000, ‘Dendochronologically Dated Ottoman Monuments’, in : U. Baram and L. Carroll
(eds.) A Historical Archaeology of the Ottoman Empire: Breaking New Ground, Hingham, Mass.,
93-136.
Le Strange, G. 1905, The Lands of the Eastern Caliphate, London.
Lewis, F. 2000, Rumi: Past and Present, East and West, Oxford.
Meinecke, M. 1976, Fayencedekorationen seldschukischer Sakralbauten in Kleinasien, 2 vols., Tübingen.
Melikoff, I. 1965, ‘Dānishmendids’, Encyclopaedia of Islam, 2nd ed. Leiden, vol. 2, 110-1.
Minorsky, V. 1945, ‘Khāqānī and Andronicus Comnenus’, Bulletin of the School of Oriental and African
Studies 11, 550-78.
Minorsky, V. [Bosworth, C. E.] 2000, ‘Tiflīs’, Encyclopaedia of Islam, 2nd ed. Leiden, vol. 10, 479.
Minorsky, V. [Bosworth, C. E.] 1986, ‘al-Kurdj’, Encyclopaedia of Islam, 2nd ed. Leiden, vol. 5, 486-97.
Minorsky, V and Taeschner, F. 1960, ‘Akhlāṭ’, Encyclopaedia of Islam, 2nd ed. Leiden, vol. 1, 329-30.
Mordtmann, J. H. [X. de Planhol] 1978, ‘Ḳarā Ḥiṣār: Şebin Kara Hisar’, Encyclopaedia of Islam, 2nd ed.
Leiden, vol. 4, 578-9.
Nasir-i Khusraw. 1986, Nāṣer-e Khosraw’s Book of Travels (Safarnāma) (trans. W.M. Thackston Jr.),
Albany.
Necipoğlu, G. 1999, The Topkapı Scroll: Geometry and Ornament in Islamic Architecture, Santa Monica.
Nizami Ganjavi. 1374/1995, Kulliyāt-i Khamsa-yi Ḥakīm Niẓāmī Ganjawī (ed. S. Nu’mani), 3rd printing,
Tehran.
Nizami Ganjavi. 1945, The Treasury of Mysteries, London.
Ocak, A. Y. 1980, Baba’iler İsyanı, İstanbul.
Ocak, A. Y. 1996, ‘Fütüvvet: Tarih’, Türkiye Diyanet Vakfı İslam Ansiklopedisi, İstanbul, vol. 13, 261-63.
Oral, Z. 1962, ‘Anadolu’da San’at Değeri Olan Ahşap Minberler, Kitabeleri ve Tarihçeleri’, Vakıflar Dergisi
5, 23-77.
Ödekan, A. 1977. Osmanlı Öncesi Anadolu Türk Mimarisinde Mukarnaslı Portal Örtüleri, İstanbul.
Ödekan, A. 2002, ‘Anadolu Selçuklu Çağında Mukarnas Bezeme’, in: D. Kuban (ed.) Selçuklu Çağında
Anadolu Sanatı, İstanbul, 329-35.
Ögel, S. 1987, Anadolu Selçukluları’nın Taş Tezyinatı, Ankara.
Önge, Y. et al., 1978, Divriği Ulu Camii ve Darüşşifası, Ankara.
Önkal, H. 1996, Anadolu Selçuklu Türbeleri, Ankara.
Özbek, Y. 2004, ‘Divriği Hamamları’, Erciyes Üniversitesi Sosyal Bilimler Enstitüsü Dergisi 16, 191-228.
Pancaroğlu, O. forthcoming/2010, ‘Convivial Congregations: Devotion and Hospitality in the Buildings of
Medieval Anatolia’, in: R. Kana’an and Z. Hirji (eds.) Places of Worship in Muslim Societies, New
York.
Peacock, A. C. S. 2006, ‘Georgia and the Anatolia Turks in the 12th and 13th Centuries’, Anatolian Studies
56, 127-46.
Pickett, D. 1997, Early Persian Tilework: The Medieval Flowering Kāshī, London.
Sakaoğlu, N. 2004, Türk Anadolu’da Mengücekoğulları, İstanbul.
Shani, R. 1996, A Monumental Manifestation of the Shī‘ite Faith in Late Twelfth-Century Iran: The Case
of the Gunbad-i ‘Alawiyān, Hamadān, Oxford.
194 OYA PANCAROĞLU

Shukurov, R. 2001, ‘Turkoman and Byzantine Self-Identity. Some reflections on the Logic of Title-Making
in Twelfth- and Thirteenth-Century Anatolia’, in: A. Eastmond (ed.) Eastern Approaches to Byzantium,
Aldershot, 259-76.
Shukurov, R. 2005, “Trebizond and the Seljuks (1204-1299)”, Mésogeios 25-26, 71-136.
Stern, S. M. 1960, ‘Abd al-Laṭīf al-Baghdādī’, Encyclopaedia of Islam, 2nd ed. Leiden, vol. 1, 74.
Sümer, F. 1990, Selçuklular Devrinde Doğu Anadolu’da Türk Beylikleri, Ankara.
Tabbaa, Y. 2001, The Transformation of Islamic Art during the Sunni Revival, Seattle.
Thomson, R. 1991, ‘Georgia’, Oxford Dictionary of Byzantium, Oxford, 840-1.
Toorawa, S. 2004, ‘Travel in the Medieval Islamic World: The Importance of Patronage as Illustrated by
‘Abd al-Latif al-Baghdadi (d. 629/1231) (and other littérateurs)’, in: R. Allen (ed.) Eastward Bound:
Travel and Travellers, 1050-1550, Manchester, 53-70.
Turan, O. 2001, Doğu Anadolu Türk Devletleri Tarihi, İstanbul.
Turan, O. 2002, Selçuklular Zamanında Türkiye, İstanbul.
Ünal, H. R. 1967, ‘Monuments salguqides de Kemah (Anatolie orientale)’, Revue des études islamiques 35,
149-72.
Yaqut. 1977, Mu‘jam al-Buldān, Beirut.
Yavuz, A. T. 1978, ‘Divriği Ulu Camisi Hünkâr Mahfeli Tonozu’, in: Y. Önge et al. (eds.) Divriği Ulu Camii
ve Darüşşifası, Ankara, 137-54.
Yazar, T. 2007, Nahcivan’da Türk Mimarisi, Ankara.
Yazıcı, T. 2004, ‘Ḥobayš b. Ebrāhim b. Moḥammad Teflisi’, Encyclopaedia Iranica, vol. 12, London.
Yinanç, R. 1970-74, ‘Mengüceklere Ait Bir Vakfiye Sureti’, Tarih Araştırmaları 8-12, 17-21.
 The Mosque-Hospital Complex in Divriği: 195

1. Mosque-Hospital Complex of Divriği, view from above

2. Mosque-Hospital Complex of Divriği, plan (from Cahen 1934, fig. 112)

196 OYA PANCAROĞLU

3. Mosque, north portal 4. Hospital, portal

5. Mosque, east portal 6. Mosque, north portal, left side

 The Mosque-Hospital Complex in Divriği: 197

7. Mosque, north portal, right side, detail 8. Mosque, north portal, right side, detail

9. Hospital, portal, left side, detail of human 10. Hospital, portal, right side, detail of human
head above a vegetal roundel head above a vegetal roundel
198 OYA PANCAROĞLU

12. Hospital, north portal, detail of upper inscription

11. Tomb of Sitte Melik

13. Tomb stelae, Ahlat (from Bachmann 1913, Taf. 45)

14. Khatchk’ars in Armenian cemetery, Ahlat (from Bachmann 1913, Taf. 45)