1930122JuneAgaOgluFa..

THE FATIH MOSQUE AT
CONSTANTINOPLE
BY MEHMET AGA-OGLU
OUR years ago I pointedout in an articleentitledDie GestaltderalterMohammedijein Konstantinopelund ihr Baumeister,'that the first imperial mosque
erectedat Constantinoplehad a shape other than that of the present Fatih. Its
shape was not the outcome of direct Byzantine influence as had hithertobeen
believedbut was rather an organic continuationof purely Turkisharchitectural
that
had arisen in Anatolia. This fact is now confirmedby furtherdocumentary
thought
evidencewhich I have been fortunatein finding,and whichinducesme to returnoncemore
to this questionso importantfor the historicaldevelopmentof Turkisharchitecture. But
before I pass on to the new historicaldocumentsit is necessaryto repeat once more what
was said in the above-mentionedarticle.
Ten years after the conquest of Constantinople(1453) the conqueror,Sultan Mehmet
II, ordereda "worthy" mosque to be erected in place of the famous but already "dilapidated"2 Churchof the Apostles. In some such words the Turkish sources report on the
constructionof the Fatih mosque. The contemporaneoushistorianssuch as Torsun Bey
and Kritobulos give us detailed informationabout the architecturalactivity of Sultan
Mehmet II and mention also the constructionof the Fatih mosque. But all these reports
are so generalthat it is exceedinglydifficultto form a definiteidea as to the shape of the
Fatih.A Frequent earthquakes damaged the mosque repeatedly so that it had to be
repairedin I509 under the reign of Sultan Beyazid II, the successorof the conqueror.4
A second fatal earthquake took place on May Iith, 1179 A. H., which destroyedthe
large central cupola entirely and damaged the whole building so severely that a reconstructionon the old foundationswas impossible.
ContemporaneousTurkish sources give us the best and most detailed information
concerning this destruction. One of these sources was already used by Djelal Essad.5
It is the valuable book about the ConstantinopleMosques Hadiqatiil-Djewami(Garden
of the Mosques)written in 1768 (1182 A. H.) by Hafis Husseyn Effendiben Hadshi Ismail
Aiwan-Serayi.6He reports: "The severe earthquakewhich took place in our times on
ThursdayMay IIth, 1179 [1765A. D.] on the third day of KurbanBairamone hour after
I. In Belvedere, 1926, no. 46, pp. 83-94.
2. Ewliya Chelebi, Seyahat-Name, Istanbul,
A. H., I, p. 138.
5. Constantinople, de Byzance 4 Stamboul, Paris, 19o9,
p. 215.
6. Printed in Istanbul, 1864 (1281 A. H.), in 2 volumes.
1314
3. Torsun Bey, Tarihi Abulfatik, publ. by Toe,
Istanbul, 1330 A. H., p. 63; Kritobulos,Tariki Sultan
Mehmedi Sani, publ. by Toe, Istanbul, 1328 A. H.,
I do not understandwhy Fr. Babingerin his malicious
article directed against me (OLZ, no. 7, July, 1927) calls this
pp. 128,129.
4. C. Gurlittsays so in his BaukunstKonstantinopels,
Berlin,1912, p. 59. However,I cannotsupporthis state-
mentbydocumentary
evidence.
179
very importantbook"a secondarysource." On the other
which
hand,in Wasif'sMehasin-ti-asarvehagaig-il-akhbar,
accordingto Babinger"gives all desirabledetails about
the destructionof the conqueror'smosque"I did not find
a singlewordabout the Fatih mosque.
180
THE ART BULLETIN
sunrisedestroyedthe large cupola of the mentioned[Fatih]mosque. The remainingwalls
of the buildingwere torn down to the groundand the reconstructionwas begunanew."'
Another report which agrees with the first we find in a manuscriptfrom the private
libraryof Dr. Halil Edhem Bey at Constantinople.8This manuscript,as yet unknownto
art students, is entitled: History of Sultan Selim Djami, erectedby Mimar Sinan at
Adrianople. It contains a number of notes by Mustafa Daye-Zade, a state official in
EuropeanTurkey whose father was the chief architectof the new Fatih mosque. On leaf
seven of the originaltext we read the followingnote about the destructionof the mosque:
"After the complete destruction of the holy mosque of the conqueror,the late Sultan
Mehmet Ghazi Khan, this (mosque)was built anew and completed."
Without taking into considerationthe other contemporaneoussources which mention
the earthquake, these two reports suffice to prove the destruction of the mosque and
especiallythe demolitiondown to the groundof the remainingwalls, for the purposeof a
reconstruction.9
After this collection of facts, let us turn to the question of the original shape of the
destroyedFatih mosque, the question which chiefly interests us.
What was the shape of the old mosque? The Turkish and occidentalsourcesthat give
an answerto this questionhave unfortunatelybeen disregardedas yet. Of all the chroniclers
of the fifteenth and later centuries only Torsun Bey describes the mosque. But his
descriptionis so general that it is impossible to get even an approximateidea of the
structure. To be sure he says the mosquewas built "in resemblanceof the Hagia Sophia."
Yet it is not to be doubtedthat the writerdoes not referto the ground-planof the building
but rather to the splendorof the edifice as a whole, for, as we shall see at once, the old
mosquehad nothing in commonwith the shape of the Hagia Sophia.
On the other hand, we have a detailed descriptionof the old Fatih by the famouswriter
and traveler Ewliya Chelebiben Mehmet Zilli Dervish, the TurkishPausanias,who saw
the mosquebeforeits destruction-he died shortly before 1679 (1o9o A. H.). I shall give
here the translation of the important passage relating to the question before us: "By
stone stairs which are on either side of the mosqueone enters the interior. The height of
the mosquefrom the groundto the roof amountsto 60.9 yards and from the groundto the
floorof the interior2.8 yards. The cupola built in fifteen sections rests on four supports;
on the Mihrab-sidethere is a semi-cupola. On the left and right sides (of the interior)
stand two beautiful porphyrycolumns."'•
This importantdescription,which shows us the originalshape of the Fatih mosque, is
supplementedby an equallyimportant note in the above-mentionedHadiqatiil-Djewami.
There we read: "The cupola was erected on four pillars instead of two elephant feet and
two porphyry columns as before, and the porphyry columns were buried outside. The
interiorwas [by this change]greatly enlarged."'n
Fromthese statementswe may safely concludethat the old Fatihmosquewas not covered
by one main cupolaand foursemi-cupolason the foursides as the presentedificeis (Fig. 4);
7. Vol.I, p. 9.
8. To His ExcellencyDr. Halil Edhem Bey I express
my best thanks for the copy of the manuscripthe most
obliginglyput at my disposal.
9. Cf. also H. Saladin, Manuel d'art musulman,I:
Architecture,
Paris, 1907, p. 506.; Djelal Essad,Constanti-
nople ...,
p. 215.; C. Gurlitt, Die Baukunst Kon-
stantinopdes,pp. 58 f.; Diez-Glilck, Alt-Konstantinopel,
Munich, 1920, p. 21.
io. Seyahat-Name, Istanbul, 1314 A. H., I, p. 138.
I1. Vol. I., p. 1o; cf. also Djelal Essad, Constantinople
S. ., p. 215; and C. Gurlitt, op. cit., pp. 58 f.
0.1
RR
W" ;::
...
..
K*~-p
-4MT
xirz
...
.i
?ilmi*
FIG. I-Constantinople: Old Fatih Mosque
Draming by Melchior Lorichs. 1559
liWli
......
~,
'
Am
...
.....
......
I"ul
FIG. 2-Constantinople: Old Fatih Mosque
Drawing by Melchior Lorichs. 1570
FIG. 3-Constantinople: Old Fatih
of a Water Conduit.
FIG. 4-Constantinople: New Fatihl
Mosque. Ground Plan. 1771
FIG. 5-Constantinople: Old Fatih
Mosque. Ground Plan. 1463-1471
F1G.6-Constantinople: Old Fatih Mosque. Detail
of General View of City. 1578
FIG. 8-Diwrigi: Medresse
Ground Plan. 1228
FIG.
7-Erzerum:
Chifte-Minareli Medresse.
Ground Plan. iio6-iii6
FIG. 9-Konya: Sircheli Medresse
Ground Plan. 1242
THE FATIH MOSQUE AT CONSTANTINOPLE
183
but that it had one main cupola and one semi-cupolaresting on two elephant feet (round
pillars) and two porphyry-columns.
Besides these reports we have some well-preservedand long well-knowndrawingsby
occidental artists which show us fairly clearly the old mosque in its original shape.
These are the drawingsby the two Germanartists MelchiorLorichsand WilhelmDilich,
publishedby E. Oberhummerand C. Gurlitt.12
The artist Melchior Lorichs,"'who stayed at Constantinoplevery probably between
I557 and I561 depicts on his well-knownview of the city the Fatih mosque with the
followingannotation: "Esky SuldanMohammet,derdie StadtConstantinopel
Eingenommen
vonnden ChristenMoschea."
This view of the city, dated 1559, was doubtlessly drawn from nature duringhis stay
at the Golden Horn. It shows the mosque in its originalshape (Fig. I). It towers high
above the schematically drawn buildings with cupolas (probably the eight medresses
belongingto it). We see quite distinctly the rectangularbody of the mosquewith its two
minaretsand the cupolas of the court arcadeson the right side. On the high walls of the
body of the mosque we observe also three side cupolas which are not mentioned in the
descriptionof Ewliya Chelebi. To the right of the minaretsis the usual portico with five
cupolasso characteristicof Ottoman-Turkishmosques. The largecentralcupolarests on a
polygonal drum-like substructure pierced by windows and provided partly with wall
pillars, partly with buttresses. Below this polygonal drum-likesubstructurethere is a
semi-cupolaon the kiblahside (left) which is drawntoo small in the picture. But we must
not forgetthat this illustrationis an enlargedsectionof table XIII of the drawingspublished
by E. Oberhummer,and that this total view of the city was taken from the Galata side,
from the north. Consequently, the kiblah side with the semi-cupola is not distinctly
visible, as it looks to the southeast.14
There is no question but that Melchior Lorichs' picture which supplementsEwliya
Chelebi's descriptionis reliable. We can prove it by comparingthe other still existing
mosques with their representationon the same picture. This is of so much greater importance as we have still anotherspecial view of the mosque,also by MelchiorLorich,in
which several architecturalparts are lacking,especiallyall cupolasexcept the largecentral
one, without which a late Turkish mosque is inconceivable(Fig. 2). The strongly Europeanizedsurroundingsof the mosque,however, prove that this woodcut was not made
during Melchior Lorichs'visit to Constantinople,but later in Europe,and the date I570
confirmsthis view.
But deservingof notice is the engravingdone by WilhelmDilich16with the annotation:
"MahmmetisM" (see the tailpiece of this article). I shall not enter into the details of
12. E. Oberhummer, Konstantinopel unter Suleiman
d. Gr., Munich, 1902, pl. 13.; C. Gurlitt, Zur Topographie
Konstantinopels im i6. Jahrh., in OrientalischenArchiv, II,
PP. 55f., figs. I1, 12.
13. The artist was born at Flensburg in I527 and is
known as a painter and engraver. He took several
journeys,also to Constantinople.In 1582 we find him at
Copenhagenas courtpainterof FrederickII. He probably
died after i59o. His pictureof Constantinopleis kept at
the universitylibraryat Leyden. Concerningthe artist cf.
Neues allgemeines Kiinstlerlexikon, Bearbeitet von G. K.
Nagler, Linz a. D., 1907, IX, pp. 27f.
14. Cf. the photo of the city panoramataken from
Galata tower in Diez-Gliick,Alt-Konstantinopel,
Munich,
1922.
I5. Wilhelm Dilich, also named Dilich-Schiifferor
Scheffer was born at Wabern in Hesse in 1571-72. He was
a gifted architect, engraver,and chronicler,and died in
April, 165o, at Dresden. The engravingis from his book
publishedby himselfin Casselin I6o6entitled"Einernewe
Beschreibung Wahrhaften abriss der jetzigen Stadt Con-
stantinopel."cf. Thieme-Becker,Kiinstlerlexikon,vol. IX,
p. 288 f.
184
THE ART BULLETIN
this engravingas the first glance at it convinces us that this work is nothing but a "fantastically adorned"1 copy of MelchiorLorichs'picture.
Let us now try to reconstructthe originalgroundplan of the old Fatih mosquewith the
help of this literary and pictorial informationat our disposal. Accordingto the abovementionedstatements in Turkish sourcesand judging by MelchiorLorichs'drawing,the
mosque must have had originally the following ground plan: a rectangularspace was
surroundedby high walls. Its centralpart, as Ewliya Chelebiexpresslystates, was vaulted
over by a cupola and a semi-cupolaand the two side parts were coveredby three smaller
cupolaseach, as is shownin MelchiorLorichs'drawing(Fig. i). The entirevaulting of the
"
mosqueas well as the outerwalls rested on four free supports,on two elephant feet" and
two porphyrycolumns,as is reported in Hadiqatil-Djewami. The two "elephant feet"
(round pillars) such as are used also in several later mosques, stood on the kiblah side
supportingthe cupola,the connectingsemi-cupola,as well as the south and northwalls and
two small cupolassituated right and left of the semi-cupola. The two porphyrycolumns
supportedthe other four side cupolas. The unequalsupportsin the interiorof the mosque
were consequentlynot placed in the cupola square as C. Gurlitt believes," but the two
round pillars and the front wall of the body of the mosque formed the central cupola
square. The connectionbetween the middle and side spaces was effectedby wide arches.
In this way a united space was created. In front of the mosquelay the porchso characteristic of all Ottoman-Turkishmosques. Here it was covered by five cupolas. On either
side rose the minarets with one gallery each for the muezzins (Fig. 5). In front of the
mosqueitself extendedthe rectangularcourt surroundedby arcadescoveredwith cupolas.
In the middle of this court was the ablution fountain, which had a roof resting on eight
columns.18
This, then, was the originalgroundplan of the old Fatih mosque,which as yet has been
unknownto art students. We shall see later that this groundplan became of fundamental
importancefor the developmentof Ottoman-Turkisharchitecture. These statemen'tsof
mine concerningthe reconstructionof the old Fatih mosque have found the appreciative
approvalof some art students"9while othersunfortunatelyhave disregardedthem.
Now I shall present new pictorial evidence which confirmsthe facts stated by me in a
convincingand final mannerfour years ago.
Fig. 3 shows the mosque of the conquerorwith the following Turkish annotation:
"The sublimemosque of the fatherof VictorsSultanMehmetKhan Ghazi." This picture
is taken from a plan of the water conduits kept in the library of Koprtilti-ZadeMehmet
Pasha in Constantinople. The plan is a scrolla few meterslong showingthe waterconduits
of Constantinoplelaid duringthe vezirate of Kiprtilti-ZadeMehmetPasha. This plan is
16. C. Gurlitt, Zur TopographieKonstantinopelsim
z6. Jahrh.,in Orientalischen
Archiv,II, p. 4.
17. C. Gurlitt,Die Bauk. Konstant.,p. 59.
x8. EwliyaChelebi,op. cit., I, p. i39.
19. Cf. E. Diez, Die KunstderislamischenV6lker,new
d'unecompange
ed., 1928,pp. 1o4f.; Al. Gabriel,Les&tapes
dans les deux Irak d'apr s un manuscritturc du XVI
sidcle,in RevueSyria, IX, 1928, p. 340.
Prof. Dr. C. Gurlitt, the past master of Turkish art
study, wrote to me in a letter soon after my article had
appeared:"Many thanks for your workabout the Fatih
mosque. I agree absolutely with your views, having corrected my own held so far. I am glad that it is a Turk
who gave us this information, thus rectifying our opinions."
Prof. Dr. Sarre expressed his approval still more strongly
in a letter of June 25th, 1926, saying: "It is very gratifying
that you have shed light on the history of one of the most
important edifices of Constantinople, at the same time
elucidating the development of Turkish architecture in
general, thus delivering us from preconceived opinions."
In a similar strain were the letters of Prof. Dr. K. Woermann, Prof. Dr. K. Wulzinger, and others.
THE FATIH MOSQUE AT CONSTANTINOPLE
185
dated 1673 (Io83 A. H.) and was consequently drawn almost a hundredyears before the
destructionof the Fatih mosque by the earthquake. It shows the importantarchitectural
monumentsof the city, among them the Fatih, amidst the schematicallydrawnhouses. In
the foregroundof the picturewe see a rectangularsquarewith fourtrees. On two sides of it
are the still existing medresse buildings, and on a third the mosque with its court and
two tUirbehs.The external appearanceof the Fatih correspondsexactly to the picture
we formedof it three years ago. A cubic structurewith entrancegate and eight windows
supports the cupolas, a large central cupola, on the kiblah side a semi-cupolaand on the
front side the three smallerside cupolas.
The three smaller side cupolas on the other side of the building we must supplement
accordingto the groundplan (Fig. 5). In front of the mosque are the two minarets and
the court with fourcypresstrees as they weredescribedin detail by Ewliya Chelebi.
Another, though rather inexact, drawing which nevertheless shows on the whole the
shape of the old Fatih we find in the famousHiiner-Nameof ShahnamechiLogmanEffendi.
This illustrated book was written in 1578. It is kept in the Library of the Topkapu
Serayi Museum under no. 1431. Among the illustrationsof this work is the generalview
of Constantinoplein which one can see distinctly the city walls, the Hagia Sophia, the
mosque of Sultan Beyazid II, and the Suleymaniye. Severalother historicmonumentsas
Yedi-Kule, Topkapu Serayi, the Hippodromwith its obelisks, and the serpent column
are easily recognizable. Among these monuments is also the Fatih mosque in its old
shape. In our enlargeddetail of the part of the generalview in whichthis mosqueappears
(Fig. 6) one recognizesthe central cupola, the semi-cupolaon the south side (drawnperspectively wrong) and also the small side cupolas, of which one, however, is missing.
Though some details are not given in this illustration-its size makes full detail impossible
anyway--in its chief parts the mosque is depicted true to nature, thus reaffirmingour
opinion.
After showing these two further pictorial evidences I should like to raise again the
question as to the position of the old mosque in the developmentof Ottoman-Turkish
architecture,and I propose to treat this question more in detail than was done in the
article in the Belvedere.
First of all, let us ask: Does the old Fatih mosque with its shape (Formerscheinung)
fit into the courseof developmentof Ottoman-Turkisharchitectureor not?
The mosque in its original old shape, as is seen from its reconstructedground plan,
representsa pronouncedbroadspace buildingwhichwas inclosedby walls on all four sides.
Its roofingconsistedof severalcupolasof differentsizes and one semi-cupola.
The first glance at the ground plan reveals to us the fact that we have in this building a
space construction which marks an important step in Ottoman-Turkish architecture
towards space unification. The chief elements of which the broad space of this mosque
is formed are the cupola squares. It is the prevailing space form in all Ottoman-Turkish
mosques. It is also to he found in Seljuk-Turkish architecture. In the older mosques
these cupola squares appear in a smaller scale independently, as in the Mahmud Chelebi
Djami in Isnik,20but also in co~rdination and subordination to other space units, as in
several mosques at Brussa and Adrianople. The second space form which makes its
20.
Groundplan in C. Gurlitt'sDie islamischenBautenin Isnik, in Orient.Archiv.,III, fig. 27.
186
THE ART BULLETIN
appearancein the old Fatih mosque is the semi-square,the roofingof which consists of a
semi-cupola. This is a structural element which was unknown in Turkish architecture
before the constructionof this mosque. It was first employedin this building.
How did the combinedbroadspace of the mosquearise out of these fundamentalunits?
We first find on the two narrowsides of the entirespace,whichwas moreor less undivided,
a co6rdinationof equal square space units. These were not separatedfrom one another
by walls as in the earliermosques at Brussa but were united by arch openings. In the
center of these rows consisting each of three squarespaces there is a larger square space
and beside it at the back a semi-squarespace which are connectedby arch openingswith
each other as well as with the lateral space units. This arrangementof the central space
unit is also found in the earlier mosques; with a difference,however. While there the
center consists of equal square units, here the second space was diminishedby one-half
and was coveredwith a semi-cupola. This meant an importantstep towardscentralizing
the entire space. In spite of the tendency towardsuniting the individualspace units, we
recognizehere, however, distinctly the adding spirit of the earlier buildings as also the
aim to create a broad space suitable for the so-calledpeople'smosque (the Ulu-Djami at
Brussa).
Consequently,we have here in the old mosquea new space arrangement,which shows a
new model and a logical continuationand further development of the space formation
developed in Brussa. But in spite of the unifying tendency, it does not yet achieve an
organicfusionof the space units and is still bound to the needs of kultussuppositions.
Another formal element which is characteristicof the old mosque buildings is also
stronglyapparentin this building. It is the orientation. The mosquestands perpendicular
to its longitudinalaxis. This is clearlyindicatedby its main entranceand its broadcentral
aisle.
In spite of its adherenceto the earlier architecturaltraditions, the old Fatih mosque
shows a new artistic feature, that is to say it passes beyond the types of ground plans
prevailing until then. This severance from the old traditions doubtless proved to be
decisivefor the furtherdevelopment. It involves a skillful varying and developingof the
existingform, an attempt whichhad been mademuch earlier,thoughin a differentfashion,
in the Uetsh-Sherefelimosque at Adrianople.21 Hence, from the standpoint of artistic
appearancealso the old mosque was only a modificationof the old traditionalform. It
was not yet perfectionbut only a transition.
We still have to considerthe position of the old Fatih mosquein the historicaldevelopment of the Ottoman-Turkisharchitecture,and it is of importanceto do this more in
detail. We have first to answer the question: what is the origin and developmentof the
Turkish mosque building in general and then define the position of our architectural
monumentin this courseof development.
The mosques of Ottoman-Turkisharchitecturewere not developedfrom the Christian
Byzantine churches as has been commonly believed,22but from the Islamic-Persian
medressebuilding. This historic fact will cease to sound paradoxicalwhen we considerit
21. Cf. groundplan in C. Gurlitt,Die BautenAdrianoArchiv.,I, igio-II, p. 51, fig. 8.
pels, in Orientalisckes
22.
It may suffice to mention of older works Schnaase,
Geschichte der bildenden Krinste, III, p. 471; of more
modern works, Wilde, Brussa, Berlin, 19o9, p. 1o.
THE FATIH MOSQUE AT CONSTANTINOPLE
187
more closely. It may be assumed that the origin of the Islamic-Persianmedressestyle is
bound to the historic expansionof the Turkishpeople. This is best provedby the fact that
those Islamic countries which remaineduntouched by the Turkish invasiondo not know
this type of constructionat all.23
The medressesin Persia, CentralAsia, and Asia Minor have all a definite architectural
shape, with the exception of a few variations. This shape is an open rectangularcourt
with vaulted liwans. These are open towards the court and lie in the middle of the four
court fagades. Between them are rows of dwellingcells.24
As was already stated, this medressebuilding which was first developedunder Turkish
rulein the eleventhcenturyin Persiaand Mesopotamia25
spreadwith the historicexpansion
of the Turks over the western Islamic countries.
The Chifte-MinareliMedressein Erzerum,dating from the time of the Seljuk Sultan
Melik-Shah (iio6-iii6), is the oldest medresse in Asia Minor known to us26 (Fig. 7).
It consists of a rectangularcourt aroundwhich the two-storieddwellingcells are grouped.
Between them in the axial cross rise the four liwans. The one lying on the narrowfront
side serves as entrance. We see in this constructionalready a differencein shape when
comparedwith the Persian medresses.
This is of great importancefor our question. The two liwans on the narrowsides, that
is to say, on the front and back sides of the building,do not lie in the middleof a numberof
dwelling cells as in the Persian medresses (cf. medressein Khargird),but they are each
flankedhere by one or two largerrooms. (On the groundplan of this medressethe rooms
to the left and right of the back liwan are not shown.) This resulted in a diminutionof
the court. Here for the first time we meet a new element unknownin Persian medresses.
It is the arcade gallery running aroundthe court in front of the dwellingcells. This is a
characteristicfeature of all medressesin Anatolia.27 In this oldest medressewe observe
already a tendency towards diminutionof the open central court. This tendency is still
more distinctly noticeable if we consider more closely another medresse of the second
half of the thirteenth century. The medresse in Diwrigi, built by the Djami Kebir in
1228, has again an inner rectangularcourt surroundedby arcadedhalls with axiallyplaced
liwansbetween which the few dwellingcells are arranged (Fig. 8). But throughthe reduction of the numberof dwellingcells the court space is hereso diminishedthat it has become
almost squarein shape.
Through the omissionof the lateral liwans in later buildings,as we see it in the Sircheli
Medresse in Konya built in I242, the court was still more contractedso as to become
yet more nearly square (Fig. 9). There remained nothing else but to make this open
court still smallerand to cover it with a cupola. This was really done in two cases known
23. The Egyptian and Magrebinian medresses are
shaped quite differently from those we speak of here.
Cf. K. A. C. Cresswell, The Origin of the Cruciform Plan
of Cairene Madrasas, Cairo, 1922; G. Margais, Manuel
d'art musulman, L'architecture Tunesie, Algerie, Maroc,
ff.
Espagne, Sicile, Paris 1926-27, II, pp. 5oo00
24. For instancethe medresseat Khargird. Cf. Diez,
Die chorasanischeBaudenkmidler,fig. 32.
25. E. Diez, Persien. Islamische Baukunst in Churasan,
Hagen, i. W., 1923, pp. 61 ff.
26. W. Bachmann, Kirchen und Moscheen in Armenien
und Kurdistan, Leipzig, I913, PP. 74 f.
27. The Mustansarije, which was built in I233 under
the AbbasideKhalifMustansir-billahin Bagdadand often
restored,belongs as far as its shape is concernedto the
group of Anatolianmedresses. Not only its ground plan
but also its aspect bears the characterof Seljuk architecture. Sarre-Herzfeld, Archaeolog. Reise im Euphrat- u.
Tigrisgebiet, Berlin, 1911-20, II, pp. 16o f., figs. 198 and
x99; cf. also Cresswell,op. cit., p. 35, fig. 9.
I88
THE ART BULLETIN
to us, in the Kara-Tai Medresse built in I251 (Fig. io) and in the Inche-Minareli28 of
the same time (Fig. 18). They both still exist in Konya.
The curtailmentof the entranceand of the lateral liwans,the extraordinarydiminution
of the inner court, and at last the coveringof the latter with a cupola gives a type which
is usually called the cupola medresse.
The originallybroad open court had in this way been changed into an interior space
which became of enormousimportanceto the furtherspace development. To be sure, in
these buildingsthe individualareas were still completely differentiatedand kept in their
usual arrangement. But by creating the central large space coveredby a cupola,Turkish
architecturalspirit had very nearly reached the goal towards which it strove. This was
not the end, however,it was ratherthe real beginningof space development.
This development was caused by climatic influences. From the dry plains of central
Asia the Turks had emigratedto Anatolia, where the climate is wet and rainy for nine
months of the year. We need not point out that these climatic conditionshad a decisive
influence. The very first mosquesbuilt by the Turks in Anatolia show how the architects
had to take the climate into account.
When the Turks accepted the religionof Mohammed,they also took over the Arabian
mosque--at least functionally. But this type of mosque, which was kept unchangedin
nearly all Islamic countries, as many still extant examples show, was changed here in
Anatolia so much that it can hardly be recognized. This transformationwas no doubt
partly due to climatic influences.29Close examination of such a mosque shows clearly
that the broad inner court of the Arabianmosques was here diminishedand contracted
to the utmost. It has ofrly the size of a pillar square in the middle of the mosque which
remains uncovered as a last remembranceof the former open court. Under this open
central part lies the fountain.30This same feature we find also in the Ottoman-Turkish
mosques, as for instance in the Ulu-Djami at Brussa.31 There is no doubt that the raw
climate of the country had partly brought about the roofingof the medressecourt with
cupolas.
The dissolution of the Seljuk empire and the reorganisationof the state under the
dynasty of Osman gave a fresh impulse to the further developmentof architecture. In
East and South Anatolia the peculiar Turkisharchitecturalspirit had come into contact
with Persian and Arabian, or rather with Syrian and Mesopotamian architecture. And
by the conquests in West Anatolia Turkish architecturewas partly influenced in its
outwardappearanceby the local traditionsof the conqueredByzantine territories. After
the conquest of Constantinople(1453) this transformationbecame still stronger. But it
is not permissibleto deny altogetherthe existenceof a Turkisharchitecturalspirit as has
been done with few exceptionsuntil this day. The Turkish architecturalspirit remained
in its essence the same as it had been before the reorganizationof the empireand before
the conquest of Constantinople. This fact must always be borne in mind in studying
Ottoman-Turkisharchitecture. There is a differenceno doubt between Seljuk-Turkish
28. See Fr. Sarre,Konia, fig. 25.
zu Milas, reprint
29. K. Wulzinger,Die Pirus Moschee
der Tecknischen
from the Festschiftzur Hundertjakrfeier
zu Karlsruhe,1925, P. 4.
Hockschule
30. As examplethe Eshref Rum Djami in Beyshehir.
Fr. Sarre, Reise in Kleinasien,Berlin, I896, p. 127 and
fig.; cf. also Halil Edhem,Kaisari Shahri,Istanbul,I334,
p. 63.
31. Wilde,Brussa, p. 35.
FIG. Io-Konya: Kara-Tai
Medresse. GroundPlan. 125I
FIG.
14-Constantinople: Murad
Pasha Mosque. GroundPlan
1466
FIG. i i-Brussa: Mosque of
Murad I. GroundPlan. 1365
FIG. 2-Brussa: Mosque of IldrimBeyazid. GroundPlan. 14,02
Atik-Ali
Ground
Plan
Pasha Mosque.
'497
FIG. 15-Constantinople:
i6-Constantinople: Sultan
Beyazid Mosque. Ground
Plan. 1507
FIG.
F
F
.....,?x:.-,
ijE'.
iiiiiiii~iiiIn
iiiiii~iiiiiIX
i~i!!MiN
.
.......
•iiiiiiiiiPiiiilipiiiiiiiiii•Jiiiiiiqk
...... ....
...
.
FIG. i8-Konya:
Inche-Minareli Medresse. XIII Century
FIG. i9-Constantinople:
Shah-Zade M
THE FATIH MOSQUE AT CONSTANTINOPLE
191
and Ottoman-Turkisharchitecture,yet this differencedoes not consist in fundamental
principlesbut rather in a few external architecturalfeatures. The essence of an architecture, which from its beginning shows an intense striving for enlargement and for
unificationof the interior,consists in the form of the groundplan and not in architectural
details or in the elementsof constructionwhich are only the means.
We have seen before how Seljuk-Turkisharchitecturecreated an interiorspace out of
an open inner court. This interiorspace was the basis for the developmentof OttomanTurkisharchitecture.
Long ago H. Gliick showedin his articleOestlicherKuppelbau,Renaissanceund St. Peter
that the Ottoman-Turkishmosques are an organicdevelopmentfrom the Seljuk-Turkish
medresse.32 In the mosqueof Sultan MuradI in Brussa33which was built in 1365 we have
a vivid exampleof this. The first glance at the groundplan of this mosque (Fig. i i) shows
us that we have in this building a further developmentof the Seljuk-Turkishmedresse
building. Let us look at the groundplan of the mosque. The center of the buildingis a
cupola-coveredsquare with a fountain in the middle, as the last vestige of the original
court. This central space lies always deeper than the surroundingspaces, which is very
characteristicof its originalshape. We furthermorestill distinguishthe four axial liwans
covered by barrel vaults. They are joined by arch openings with the central space.
Between these liwans are the rooms that are accessibleby doors. We have here a formal
shapingand arrangementof spaces closely relatedto the medressestyle.
The mosque of Sultan Ildrim-Beyazid(Fig. 12), which was completedin 1402, and the
Yashil Djami (Fig. 13), built 21 years later, in 1423, both in Brussa, represent a new phase
of this development,with the difference,however, that the liwan on the front side has
become an entrancehall, and that all areas are vaulted by cupolas.
The same type we find in the earlier mosques at Adrianopleand Constantinople,as,
for instance, in the old Fatih mosque (1463-1470) and in the Murad Pasha mosque, built
at the same time, 1466 (Fig. 14). But thereis one differencehere, for all four lateralunits
are separatedfrom the centralones by walls.
In this principalformof the mosquetherelay a possibilityforfurtherspace development.
The omission of the inner walls separating the mosque spaces was an important step
towardsenlargementandunificationof space. But by that the secondcentralcupolasquare
was diminishedby one half and was covered with a semi-cupola. This last change and
the introductionof a new Byzantine structuralelement was necessaryin order to obtain
a large centralspace. The Atik-AliPasha mosqueat Constantinople,built in 1497,shows
this further development most clearly (Fig. I5). Although this mosque was erected twentyfive years after the completion of the old Fatih, there is no doubt but that the same
architecturalthought was decisive in shaping its groundplan. It is not necessaryto look
for other models. The increaseof the usual two lateral cupola-coveredspaces to three
This questionwas dealt with a short time ago by
32.
K. Wulzingerin his article Die Pirus Moscheezu Milas,
Karlsruhe,1925, PP. 5-7. It was first dealt with by H.
Gliick (cf. p. i6o of the above-mentionedarticle). cf. also
U. H61scher,Entstehungund Entwicklungderosmanischen
Baukunst,in Zeitschriftfir Bauwesen,Berlin, 1919, pp.
354-396.
33. The date of construction of this mosque was
unknownuntil recently (cf. Wilde, Brussa, p. 13). Now
we are able to fix it, thanksto the publicationof an older
Turkishhistoricalbook Tarichi-al-Osman,
by Orudshben
Adil el-Kusas (Berlin, 1926). It contains besides other
valuable informationof art-historicalcharacterthe exact
date of the constructionof this mosque. Cf. also H.
Lewenklaw von Amelbeurn, Neuwer musulmanischer
HistoritiirkischerNation, Frankfurt,
I595, p. 140.
192
THE ART BULLETIN
lying on both sides of the central space, the diminutionof the kiblah space by one half,
and the omissionof the walls separatingthe mosque space led to the groundplan which
we have beforeus in the old Fatih mosque (Fig. 5).
This was nothingbut an organicand living evolutionof Turkisharchitectureas a whole.
Yet the Fatih mosque in its old shape preparedstill other possibilitiesfor space enlargement. The almost geometricrepetition of the parts developedin this mosque, that is to
say, the semi-cupolawith lateral smallercupolason the front side of the mosque,resulted
per se in a groundplan which the well-knownarchitect Khaireddinapplied with ease in
buildingthe Sultan Beyazid mosque at Constantinople,1501-1507 (Figs. 16, 20).
We do not dispute the fact that the shape of the Hagia Sophia may have given an
impulseto this last step whichwas, however,alreadyprepared. Consequently,the ground
plan shape of the Hagia Sophia was reachednot by mere imitation but by a specifically
Turkishdevelopment. We are thereforejustified in saying that the time of the conquest
of Constantinopledoes not signify this sudden revolutionin late Turkish architectureas
art historianshave hitherto believed. On the contrary,it was a movementof the Turkish
architecturalspirit that grew out of Anatolian conditions, attaining only graduallyits
acme in the new capital.
Only one more step was necessaryto create a vast unified edifice. This was the great
work of the architect Kodsha Sinan Aga. His Shah-Zade mosque at Constantinople,
(Figs. I7, I9), which was completed in 1548, is itself again a continuation of the idea of
centralizingthroughthe additionof two lateralsemi-cupolas. In this Sinanpassedbeyond
the schemeof the Hagia Sophia. This was the last and consistentexpressionof the striving
towardsspace enlargementand space unificationwhich was the aim of the Turkisharchitectural spirit from its very beginning,as we have shown in several examples.
Here in the Shah-Zademosque we have the first completely successful attempt to
create a unifiedspace, although from the point of view of constructionthis mosque does
not representperfection. This constructiveweaknessis most clearlyshownin the majority
of the buttresses. These are placed outside on the kiblah side, yet in the interioron the
front side, and are divided left and right by a wall, thus placing them partly outside,
partly inside. This constructive weakness is further shown in the dissimilarityof the
cupola-supportingarchesin the interiorof the mosque,for those which supportthe central
cupola are narrowerthan those supportingthe smallercornercupolas.
The Sultan Ahmedmosque at Constantinople,which was completedin 1616, that is to
say, sixty-eight years after the Shah-Zade,is a work of the architect Sedefkiar Mehmet
Aga. It may be considered a creation of the highest constructive and artistic value
(Figs. 22, 23). With its four semi-cupolas leaning aginst the central cupola it follows in
its space formation the plan of the Shah-Zade. With this mosque Turkish space constructionreachesits culminatingpoint.
Not until the epoch of the declineof Turkishpower,which coincideswith the declineof
Turkish architecture,did the new Fatih mosque, completedin 1771, receive its present
groundplan and elevation (Figs. 4, 21). It is nothing but a repetition of the Shah-Zade
mosqueerected by the great master Kocha Mimar Sinan Aga.
There is one more question to discuss: In how far have Byzantine influencesaffected
this courseof development?
ii4i.
FIG. 20-Constantinople:
ii~iiiiiiiiiiiii0
iiiiiiiii!iiiiiiiiA
Sultan Beyazid Mosque. r507
".7:i
OLi
FIG. 2 i-Constantinople:
New Fatih Mosque. 1771
MIS
PO
1,
........
.
.....
fle
xg,
..............
fwp:
Ims
..
Yv"
fzA
. . . . .......
Mb
42.,
X3
v ?00
fO?
.,.:a
A
sm
to
-
IN.
'k.
N
FIG. 22-Constantinople: Cupola of Sultan Ahmed Mosque. i616
...roll;
FIG. 23-Constantinople: Sultan Ahmed Mosque. 1616
THE FATIH MOSQUE AT CONSTANTINOPLE
195
The common assumption that the late Turkish mosque style is dependant on the
Byzantine church style, especially on the domed cross church34must be denied on the
strength of the above-mentionedfacts.
We must also deny the assumption that the Byzantine architecturalspirit manifests
itself in Turkisharchitectureafter the conquestof Constantinople.35There was no more
living Byzantine architecture during the rise of the Ottoman-Turkishspace construction. It is not to be supposed that the Hagia Sophia,which itself occupies a unique
position in Byzantine development, can be consideredthe direct model for the large
domedmosquesof the late time."*But it shall not be deniedthat the largespacedimension
of the Hagia Sophiamay have given a furtherimpulse to a developmentthat had already
begun. We have seen above that the Beyazid mosqueat Constantinoplewas in its ground
plan not a mere repetitionof the Hagia Sophiabut was alreadypreparedby the old Fatih
mosque, which itself was the outcomeof a long development.
Ottoman-Turkisharchitecture,to be sure, employs in the main the same structural
elementsas the Byzantine, as, for instance, columns,pendentives,semi-cupolas,but these
elements are only a means by which an original thought is realized. Although this
structural thought aiming at space enlargementand space unification may be found
among the Byzantines, it is not to be doubted either that the idea of space creationwas
born independentlyin Turkish architectureand went its own way. This is best proved
by the architecturalmonumentsthemselveswhichwe have been consideringherebriefly.
34. H. Wilde, Brussa, p. io; cf. K. Wulzinger Die
Pirus Moscheezu Milas, p. 9.
35. C. Gurlitt,Die BaukunstKonstantinopels,
p. 58.
"a~L1~b~OFT"
~(~~I~ W-r
~3EI~I~
36. Cf. K. Wulzinger in Altitrkische Keramik, by
Al. Raymund,Munich,1922,p. 16.
7M
da
r'a
-
Constantinople:Old Fatih Mosque--Engravingby WilhelmDilich
afterDrawingby MelchiorLorichs