The 'House of the Prophet' and the Concept of the Mosque
JEREMY ) OHNS
The earliest mosque for which there is archaeological evidence, and lor winch (lie dale is reasonably secure and almost universally accepted, is that built for al-IIajjaj ibn Yusuf at Wash in 8^/703 (fig. i ;. It consists of an enclosure approximately 100 m square, at the qibla end of which is a hypostylc hall. There is no trace of a imhrah niche, bul there are the thickened inundations of what Safar and Crcswell believed to be a monumental maqsura. The resl of the enclosure is occupied by the com 'lya.nl. bordere<l on three sides by a single colonnade.1
The hypostylc structure at Wasit was by no means (he only type ol mosque to be built in the Bos/yoos. Many uncertainties continue to surround the Alarwanid rebuilding' of the Aqsa. mosque in Jerusalem, but there can be liltlc doubi thai (he works initiated by 'Abel al-Malik and completed under al-Walkl resulted in a mosque that was strikingly diilercnt from Wasit. ;fk;.2). The prayer-hall was characterised by aides set perpendicular to the qibla wall. The central nave was almost twice the width of the flanking aisles, with a high gable roof, and culminated in a great dome over the three bays before the ffib la aisle. The nave was, elaborately decorated, and the dome, arid possibly the qibla \v<.CA. were covered with mosaics. The mosque may have had a mihrab niche. The Aqsa seems to have had no courtyard of its own, but was set on the axis of the Dome of the Rock, and thus integrated into the new complex of al I Jar am al-Sharif.1 Just a few years later (By 96/706-715), al-Walid's new mosque al Damascus represented another distinctly dilferenl type (fig.3). Again, the prayer hall was characterised by a broad central nave, covered with a mighty gable roof, and culminating in a dome in front of the qibla aisle: but, here, the three aisles were parallel, not perpendicular, to the qibla wall. This mosque, too, was richly decorated. According to Sauvaget, the mosque commissioned in 88/707 or 90/700, by al-Walid on the site of the Prophet's Mosque in Medina mug.4; represented an architectural compromise between the hypostylc mosques oflraq, such as Wasit. and the 'basilica)' mosques of Syria with their heavily accentuated
1. Satar 1945. Civs well ]<)&), vol. I 1, pp. i -\2 8. 'I-I'.U, PP44. '4- n<>. ■<<)-, ;wiili hil>lio«T;ipliy .
Cresu'i.'ll & Allan 1989, pp.40, 42 tv-i-j, Roussi'i 2. Ruby (<»iihcumiiiL;.
/for/rv/ .1 tf\tidL<.J?m^«hnuiraU''-rtrh I':i<t>i/y rdh: (I l>y [nv:iy|n \u:-. jCj [ >Nn)RI> M [ nih> P." 1:.! ART. VO!I'M!! IX,        I IW!. IJ\H >KI> I <;<.;.
[Ľ RIMY I Oil.N S
thk 'house or the prophet' and iiik cioxcept of thk mosouk
-hjilJ_i_I_i_.
Scale approximately i: 1600
Figure 1. I •■ I'. to])) Reconstruction ofllie plan ofllie mosque of al-Hajjaj ibn Yiisuf at YVasit 'idler Salar 194");
Figure 2. (Left, middle; Jerusalem, al-Aqsii: 1111 ■ Mumrmiil 1111 i.sque iallrr Hamillon 1049: Figure l>r (I .eh, below! I Jamasrus, i he mosque of al-Walid 'after C'reswell l()t.i()'l Figure 4. !'l op; Medina, the mosque of al-\Valid Figure 5, (Above) SanV, the mosque of al-Walki (afier Fiuslci 1978)
jeremy joii\St
axial naves, such as al-Aqsa and Damascus." The (Jreal. Mosque al S;ui';l\ also
attributed to al-Wahd. may refer directly to his mosque at Medina, but h also makes use of spolia and of architectural and constructional details inherited from the distinctive pre-Islamic traditions of South Arabia. (fi< 1
Other types of mosque are also attributed to al-Wahd. The mosque at VVnjar has an unusual plan: ihe prayer hall and lateral mvaqs, are two aisles deep, but the northern mvaq has only one aisle (kio.6). ' The palace al. Minya has a small mosque which, unusually, i.s longer than it is wide {vui.j;." The mosque atjabal Says was a small cubical chamber with an exceptionally deep projecting niche mihrab: the interior was divided into iwo unequal parls by two stilted semicircular arches thai carried the flat wooden roof (fig.8).7 None of the mosques at 'Anjar, Minya andja.bal Says had a courtyard.
Variety is also to be seen in two olhcr mosques for which there may be archaeological evidence, but from a date significantly earlier than Wash. In his forthcoming study of the Aqsa mosque, Julian Raby will argue that the predecessor of the Vlarvvanid building described above can be largely reconstructed on the basii of the excavations and survey conducted by Robert Hamilton in 1938 4 2 (no.9). Hamilton originally dated this building — his Aqsa I — to the reigns of Abel al-Malik and al-Walld but, before he died, he came: to agree with Raby that it must be earlier; Raby is now inclined to attribute it to Mif awiya. in the early 40s/6(ios. It was a substantial structure, built of stone, and dressed with marble. The covered prayer hall measured just over 50 m north-south, and at least 45 m east-west. The arcades, like those of its Marwanid successor, rati north-south, and were carried on reused marble columns and capitals. These, in turn, carried arches and a fenestrated wall which rose lo ceiling I leigh t. The central nave, perhaps for purely structural reasons, was slightly wider than the other arcades. North of the prayer hall are traces of a. portico, and, beyond that, of an open paved area with drains and cisterns, which may have been the courtyard of the mosque.'1
A very dillerenl mosque was apparently buill at Kiifa in 50/670 by Mr/awiya's governor Ziyad ibn Abihi.'* It. was summarily described by MuqaddasI, and visited in May 580/1184 by Ibnjubayr who gives the fullest description.1,1 In 1765, Niebuhr made a little plan of its ruins.11 Archaeological soundings by the Iraqi Department
'■5. Saiivugu 11.47.. id.
.. ("res-well & Allan i()8q, pp.u/
3.   ClX'-Wcll Il)'.H), , Vol,1.2. p.
(). On swell i<)(h). veil.1.2. pp.'jti'-; |. 7. (. rcswrll 1 ()()[>. veil.1.2. p. jy(>. It is inii'iiiiiiniJ thilL the diiiuiiMojis ol llns <iibu.il TntK(|ii<' iliillr mure than n in square and some (j in hi^li
rlosck approximate diose of die Ka'ha.
a. Raby ibrthcominR.
t). Taban itiji.). vol.1, p.2,].(;il>. line-(! ly
10. Muqaddasi ])\~~, pp. i itt-7. Ibnjubavr 1907. pp.240 -2.
11. Niebubr 177 p vol.n, pp.jbi 4. pi.42b.
THE *HOUSK Of THK PROIMIRT' AND Till: UOXCKl'T OF THI, MOSOl'K
Sc.ill- ;i|)|ii't).\imau-lv i :30c
of Antiquities scorn to indicate that XiyatPs foundation survives in situ beneath the present mosque and dictates its plan.'- From these sources. G re swell reconstructed Ziyad's mosque as an enclosure approximately too in square, at one end of which was a hvpostyle hall; the courtyard was bordered on two sides bv a double colonnade; again, there was no mihrab i'fig.io).1 ' All accounts stress die exceptional height 130 cubits or circa 15 m) of the si one columns of the prayer ball and nwaq%. 'The roof rested directly upon die columns, without the medium of arches, a le.ilure that Crcswell derived from the apadana or hall of columns of the Acliaemenian kings.11
In addition to these early archaeological buildings, the historians transmit descriptions full enough to permit hypothetical reconslruciions of at least five other early mosques: Basra 1 (?i4/?(J33), Kit fa 1 ; 17/638;, Fustat 1 (21/641 2:, Basra 11 (45/665), and Qavrawan 1 ^o/b/o).1 1 In addition, wc have brief reports of the foundation of many tens ofjawnmt' mosques during the first century of Islam.'"
1'J. Orrswt'll l(jlH), vol.i. I. p. |l> postscript':. 13. Crov.HI K569, vol.i. i. pp.^tj <1. ta-rauHKt Alliin i<)8(), pp.9 if'- Rous-ici ii|()2. pp.44. 13-, (i. nn.iiji) biMioiiTupliV'. i |. C rcswcll Kilii).\nl 1. 1, pp. p1) 7.
1,). Ocvccll itjfkj. vol.i. 1: Basra i. p.     knla I. pp.aj (i; 1'usiat i, i't <.& OivmucU 1940, vol.11.
]■>]).17j i)(>>: Basra 11. pp,| |  > Qa\r.mttii 1, ji.fi:. ili. It is asionishinc>'ihal ilu-sc liavt- ncvci .teen 'vhu'iiialifiillv cullcctcd; 1'insioi''- splendid
JKREMY JOHNS
I
I i I
i I
Finnic q. Jerusalem. al-Aqsa, the prc-N-lanvantrl ntoscjiir (after U.aby fonhcorning Scale approximately i; iooo
After Wasit in 84/703, archaeological mosques become relatively plentiful, and one can reconstruct die plans of"a dozen or imm\j/iwamif mosques built under Umayyad rule, from Cordoba in the west to Bhambor in the east 'figs 11 & 12}."
Considering these early mosques, archaeological and not, one is immediately struck by the great variety that they display in the plan and elevation of die prayer hall, in conslruclional techniques and materials, and in decoration. Clearly, each of the mosques at say Kula, Damascus, and Sarfa\ was influenced by different building- traditions which had their rools in the pre-conqucst cultures of respectively, Iraq, Syria, and South Arabia. At the same time, it is evident that they arc all variations on a single therne.
The ground plans of these earliest mosques all refer to a common model, to a menial template, to a concept of the mosque which was certainly already fixed at Wasit in 84/703, and may arguably be traced back to Jerusalem in the 4OS/660S and
catalogue* of the cm I y mosques ol'lran {Finster
1904-, which lists both sanding buildings and those which .ire known only from written sources demonstrates how useful this appi oach could be fin the smdy of the development ol'llie. mosque throughout the each Islamic world. 17. Aleppo: Creswell lyby, vol.r ), p.483. 'Amman: Norlheilgr 199V, pp.t>.'-<). figs 20-27. pis 8 10. Bhambor: Ashfaquc toGij. Cordoba i: Creswell icjio, vol.11, ppi'jii tit. Dar'a: Creswell 196c),
vol,i. [. pp.hT,o  i. Damascus: Creswell 19(19, vol.1. i, pp.iý;, <)(). Hama: Creswell 1969, \ol.l. 1, pp.17  2i- l.tarran: Creswell iqbo. vo.i. 2. pp.644 8. Jerusalem. Aqsa 11: Raby forthcoming, Medina.: Sauvagn 1947. Ram la: Creswell 1969. \ol.l. i. pp.482—5. Rusälh: Oilo-Dorn 1(557; Sack, forthcoming. San'a': Fluster 197ti; l'inster 1979: Fin Her 1982a; Finster 1982b; l'inster 198b; Serjeant & t.eweoek 1983, pp-H^ jo; Creswell & Allan iyf?ii, pp.)>■;-!!. Siisa: l'inster 1994. pp-^|9 .V--
t"h !•' "housf OF the PROP1ILI    and tit I-. cokcept Ol Tin: mosouk
Figure io. (Indwell's ixcoiihtnicuon of the plan oFKnfa 11 faftcr Kuban K|~ Stale ;-ippro\im;ilrly x: 1000
co Kula in die 50S/670S. The concept may be described as a walled enclosure, one end of which is occupied by a multi-aisled hall aligned upon a qibfo, and the rest by an open courtyard lined with porticos, in such a way thai die three elements iorm a unihed whole. What makes this concept distinctive is the arrangement of its principal components and their proportions. The components themselves, the open courtyard and the covered sanctuary and porticos are, of course, common to other religious building types, most obviously to the Graeco-Roman temple. Whereas the temple usually had a centralised plan, with die sanctuary at the centre of tlie tcmcnos, the plan of the mosque could be described as terminal, with the prayer hall at one end ol the enclosure. We can see this contrast most clearly in llie transformation of the l.cmenos at Damascus with its church on the central site of the original temple, into the mosque of al-Walid with the prayer hall occupying the southern end of the enclosure (no. 13}. As will be discussed in more detail below, the mosque was not unique in arranging a colonnaded courtyard and a covered sanctuary symmetrically
ľ'igure íl. Plans ofjnn-maľ mwques bílili under Unia\\.i<l nile
a. Maehna failer Crcswell lytiq:
b. San'ä' :aľier drswell & Allan igBcj:
c. .Süsa íaiicr dresu ell Ľ\r Allan nin();
d. Bharnböi infirr Ashlaqiu k|(K).
e. Rušala aller Ulbert iqqo)
1. 'Amman .alier Aorthctlge iQUib
g. Dar'a after ('resvell iq6u;
h. Harran suľier f'.resuell & Allan iqiki'i
i. ( orduba l '.after Crcsuell & Allan iqiit) Scale ajjproMinalelv I :~oo
jeremy johns
Figure ij. Damascus, L'mayyad mosque {after Slicrlin iQ/Q': (above", reconstruction of llie [)l:m of the Irmenos and church of Sljolm llic Baptist; (below; plan of the mosque as seen today
63
'ITIK *HOUSK OF THE PROPHKT' A NO TIIK CONCKPT OK TOE MOSOL L
about the central axis of tlie complex; indeed, in Late Antiquhy, tins arrangement may be found in a variety of types of religious building, including the church, the synagogue, and the temple. The proportions of the mosque, however, distinguish it from such antecedents in that its length find width arc typically equal, or nearly so; indeed, a mosque may even be wider than it is long.
The concept of the mosque dominates early Islamic religious architecture so completely that it dictates the plan of almost vw.ry jdmi* mosque built. The Mcecan haram, the Dome of the Rock, and the Aqsa mosque are the onlv exceptions: the first two have centralised plans; the Aqsa, after the const ruction of the Dome of the Rock, had no enclosed courtyard ol its own. ,; (Here 1 do not include the small, non-congregational mosques without courtyards, including "Anjar. Miuya andjabal Says, found in die Umayyad (jusurof fliladal-Shmn.)
That a standard concept of the mosque governed the form of the congregational mosques of Islam from at least as early as 84/703, and was rapidly exported throughout the conquests as far afield as Andalus and Sind, raises the question of the origin of that concept.
THE 'HOUSE OF THE PROPTIKt'
for most of this century, it has been almost universally accepted that the origins of the mosque lie in the house that the Prophet built at Medina utgs 14 & The Kncyclopaedia of hfaiii ('2nd edition; slates: Tieyond doubt, the, genesis of the mosque is to be sought in a single seminal building: the house of the Prophet, erected to Muhammad's own specifications in Madina in 1/622/'* Yet, the history of the development of other types of religious building, such as die church and the synagogue, raises doubts a prim 1 that the Prophet's House was the sole origin of the architectural type oi the rnosque.
it has been suggested thai the fact that the earliest Christian places of worship were ordinary houses may somehow reinlbrce the argument chat the Prophet's House was the origin of the mosque, but litis is not obviously so."1 For the lirsl two centuries, there was no distinctively Christian architecture, and Christian congregations met
Iri. Ziyad ibn Ablhi is nnlikeh itj havebitilt round mosques in Basra [baa Hillenbrand iq'.)i, p.(>!>'. I he text reads: 'every mosque in Basra, the laitnhii ui which is circular, wa-. among the buildings of '/A\i\d ikut'M mayjulm hi !-B<Mtii> kana! mltuhahi hu mttsladimttmja-wiia-/ia mw binS'i ^yildoi'i': Ihn al-l'aqih hj88, p. 176. Here, mhttf'n Miivlv means not the internal emu Ivard '.■y(i'i)! Imt cither the spare within which the mosque stood, or, more probably.
a sort nfplalli.'iTii or ihikknt: in the eoui[.y;ml ruin I'ronl ol'the entrance to the mo-;quc: rt. I .aw iXtv}, p.!031. eol.e   p.1052, col.a. it). Hillenbrand tijiig, p.tijd, eol.b, bin now sec Hillenbrand [0104, pp.39 [".. •20. I lillenbrand icjtfi'i. p.byq, rol.a: 'It is snrelv «j>riiji»-i 10 note that the earliest Christian places ot\\or-hip, the so-e;illeel tiluh. were also ordn tare houses'.
jr.REM V j 11 H \ -
in ordinary dwellings. The Christian meeting houses of the 3rd century, the so-called i/ikoi ekkte.sws or domus ecck.\iae. whether they were modi lied dwellings or structures built e\ novo, employed the vocabulary of domestic architecture. The domus ■■ ■ ■• /. of the first three centuries, however, had hide influence upon the architectural development oi the church from the 4th century onwards. Even before CoiistanUrie, monumental churches were built on (lie model of the basilica and, after the Peace of the Church, architects began increasingly to experiment with variations on the basilical form. The basilica, not the house, was the monumental prototype tor the church. The domus e.crlrsiae belonged to a specific period of early Christian history, before the community had developed monumental architecture, when domestic structures were used for practical reasons, and 110I in order to commemorate a parueular house. Only in the 4th century was a new and more appropriate monumental form, the basilica, adopted as the architectural model lor the church triumphant. Even then, no single basilica, but the whole class of basilical structures served as the model; and it was not until die late 4th or early 5th ceniury that a norm for the early Christian basilica began 10 emerge."1
'1 he origins of the synagogue are obscure, and plagued by the failure of scholars to distinguish clearly between the building and the institution, ['he earliest archaeological synagogue is that at Delos, which seems to be dated to the late 2nd or mid-ist ccntuiy bg, but very little is known archaco logically of the development of the synagogue until the 3rd ceniurv. Although more than too synagogues art: now attested in Palestine, most of these date from the 3rd to 7t.lt centuries. This wealth of new archaeological evidence has forced scholars 10 abandon the. chronological classification of the synagogue into early, transitional, and laic types."' All three types were current as early as the 3rd 4th centuries, indicating that, by this dale the synagogue was subject to diverse architectural influences, including Hellenistic and Roman public, buildings, and the emerging Christian basilica.. It seems likely thai such architectural diversity characterised the synagogue from its very origin. Although lite Temple was often commemorated in synagogue art during late antiquity, the architectural form of the synagogue is not modelled on die Temple. In short, the synagogue, like the church, evolved gradually over a long period of time, and was subject to architectural influences that varied in space and time. It did not emerge rapidly and attain maturity within a single generation; nor did it commemorate in its architecture a single prototypical building/ ;
Thus, at first sight, comparison with the origins of the church and of the synagogue
'-'1.   I he fundamental study remains Kiauthentic] ui'ii, pp.'J^  7(1. Id which While lQ()(i, liassiui, rs|>. \!\).)02 48. makes important revisions. 22. Seager u)i><).
j. I,c\ itie t;)9j is a usoh.il introduction and summary. Sec also: (Juunann 19H1, hcvuir h1íi1. C.'hiat tci'jJ, Levine I'lii}, Haohlili 1080. 1 le~lici'\-I-mian u)i).|., all nith exlen-iw bibliography.
THE 'HOUSE OF THE PROIUIEl ' AND THE CONCEPT OF TI1K MOSOUH
illustrates how improbable il is that the architec I ural type of the mosque should haw been generated by a Single seminal building", and suggests that the origins of the mosque might be soughl in the much more complicated and organic process of* die transformation of pre-existing architectural forms into the Islamic: mosque. At the same lime, however, litis comparison highlights two important respects in which the early history ofllie mosque was striking!) dilTerent from that of lite church and the synagogue. First, only Islam was. almost from its inception, the religion of a ruling elite capable of sponsoring its own architecture. Second, in contrast to the prolonged evolution of church and synagogue, both Islamic hislorical tradition and archaeology bear w ili.es- that, well within 100 years of the hiprt, Muslim leaders were building mosques according lo a common standard; otic which dictates die form of congregational mosques, throughout the Muslim world, lo this day.
What this suggests is that the crucial quesliou is whether the mosque, like the church and the synagogue, gradually evolved from pre-exisiing architectural forms, or whether, unlike them, it was deliberately 'created* by the Muslim elite. One process does nut necessarily exclude the other: the architectural type or types that became die mosque may have evoked gradually before Islam, and then, after the hijrft. have been deliberately adopied (rather than created ex ni/rito'. by the new Islamic elite. The current orthodoxy that die Prophet's House was the origin of the mosque begs the question whether the mosque evolved or was created. PaTore getting lo grips with that crux, therefore, it is necessary to scrutinise the basis of the orthodox position.
CAETAXI AND THE ' HOUSE OF THE PROl'IlET"'
Cactani devotes thirty pages to a discussion of'The. origin of the mosque and the foundation of die Muslim rite' in which he seeks to demonstrate that the religious institutions and practices of earl v Islam were less the products of an individual's conscious design than of broad, long-term hislorical processes.-' Caetani's premise was thai, the mosque could no I have sprung fully-formed from the Prophet's head on the very first day that he set foot in Medina but was rather the produel of an organic historical process.J' He sought to locate the evolution of the mosque in die decades after I he liijra, in what, he described as the 'slow transformation of a building intended for exclusively domestic use, into a meeting place for' die believers, and finally into a mosque destined lor worship'.-*''
24. (l.u'lLini vol.1, pp.j'.J'J (in.
25 ■ CaiTani 11)1)5, vol.), p. [;;2.
ji). CaoLani 11)113. vo'-'- P-4S-
Jh.Kr.AlY JOHNS
Muhammad's first thought,' lie writes, 'on coming to dwell in Medina ... was ir> build himself a dwelling, a dar. which in Arabia in those days consisted of an enclosure of humble rooms, almost huts, grouped irregularly and concentrically around an open courtyard dial was more or less spacious according 1o the wealth and the number of the family that occupied il. If the family was not very numerous, the rooms were all grouped together on one side and, because Arab private life required a private courtyard; closed on all sides for domestic purposes ... and lo keep domestic animals. ... the rest of the area was completely enclosed by a wall. I he open space ilius formed the meeting place for all the fauiilv, and the ensemble of these architectural elements was so compact, that a single door communicated with the outer world. Bit by bit. as the fauiilv grew, whether by new marriages, or by the increase of children ami grandchildren, onto the rooms alrcadv existing around the courtyard wall were built new rooms identical to the loi mer until, one day, with the constant growth ol the family, all the periphery ofllie courtyard was taken up with dwellings ... This system ol house-building, thai I have seen still in use in many of the poorest and most isolated \ illages of the Near East (Arabia, Syria, Mesopotamia, etc.;. was that adopted by the Prophet.... The purely domestic and utilitarian nai ure ol the building ... is demonstrated bv a number ol traditions which the biographers of the Prophet failed to see in their true light. All of them believed that the prime intention of the Prophet was to build a place of worship, and that the dwellings of the Prophet and his family were thus an addi lion, an ad hoc arrangement of secondary importance to simplify the process of construction. Rather, it was completely the contrary: in fact. Muhammad set. about building himself a house lor his own particular and private use. and only at a later stasje. through circumstances that none could have foreseen, and as the result of a process dial was completely unintentional, the spacious courtyard in the middle of the PropheCs House assumed a public character and eventually, after the death of Muhammad., even a sacred character, that of a true place of worship. The creator of the mosque was not the Prophet, but historical determination that is above every human will."'
Caetani follows Otis passage with a long discussion, based on the Sahih of Rukliari, of the activities and even Is said to have taken place in the Prophet's courtyard which, he asserts, demonstrate that "the original idea of the Prophet was lo build not a place of worship but a simple private house like that of every other inhabitant of Medina'.2'1
As we shall sec, Caeiani's idea, of what mighl have constituted ordinary domestic life in yih-century Medina was at best idiosyncratic, but it is his naive use of ahadltk that betrays the fundamental weakness of his argument.-'1 These traditions ceach correct behaviour and prohibit transgressions, or thev establish precedents for activities thai were common in mosques in mature Islam, or they authenticate the claims to companionship or excellence of individuals or groups. Thus, for example, the story about an Arab urinating in I he mosque does not demonstrate that lu: 'took it to be the simple courtyard of a private 1101116'; it is rather an entertaining and mstruc-l.ive teaching story: the Prophet stops his followers from interrupting the man in full
i-j. Caetani i<)0;"), vol.i. pp.4f;7 a different critique of Caeumi's argument,
•>\\. Caetani 190-,, vol.i, p.4 11. see Akkuuch 1935. pp-jgH tm-
thk "HOUSF OF thr PROPHF.t' an'i) thk concept Oť THE MOSQUE
flow, instructs them to wait until he has finished, and then quips 'You have been .si-in lo make things easy, not to make them difficult!1/"1 Again, the stories about, spitting in the mosque, far from demonstrating that it was just an ordinary house, teach that pollution of a mosque is wrong ('spilling in the mosque is a sin'), and should at Least be contained by spitting into one's robe or under one's fcet.:!1 Yet again, the Ethiopian spearmen who performed in the courtyard were not standard domestic fittings, but rather establish a přeceděni for similar displays in mosques in celebration of the 'id.Finally, the presence in the mosque of wounded alter the Battle of the Ditch docs not indicate that it was really a house, but instead sacralises the excellence of the lianu CJ haf far and the glorious marlyrdom of Sa'd ibu Mu'adh."
Such slories may or may not be historical: if they are. they demonstrate not that the Prophet's Mosque was really a house, but rather that it was used lor a wide variety of military, political, and social aclivilies, as well as for prayer.51 M\ subject here is the architectural form, not the institution of the mosque, but it is worth stressing that al-masjid al-jamir was never simply or, perhaps, in early Islam, even primarily a pi ace of wors hip
Caclani's a.rgumcnl is further undermined by the rather tetchy passages in die (htr'an that seek to secure a little domestic peace for the Prophet: 'O believers! Do not enter the houses of the Prophet {buyul al-nably) for a meal, unless leave is given to you and without waiting for the proper time (i.e. until ii is ready;. But, when you are invited, enter and. when you have had the meal, leave, without lingering for talk';"' and 'Surely those who call to you from behind tire apartments ial-liujurál), most of them have no sense. And. if 1 hey had patience until you would come out. to them, thai would be better for them'.57 In these passages, the Prophet's domestic apartments ibuvUt or hujurat) are explicitly treated as private space.
Nonetheless. Caetani described the Prophet's House, not his mosque, as a courtyard approximately 100 cubits (50 m) square, enclosed by a mud-brick wall 7 cubits high, pierced by three entrances. At first, Caetani related, there was no structure within the courtyard, but after the Muslims began to complain of the heal of the sun whilst they were at prayer, a portico or ;.'//.•'•.■ was buill against die northern wall of palm trunks supporting a roof of'woven palm branches and mud. In the southwest
■'jo. Bukhan i.Któ, vol.l. p.(17, lines 3 li,
31. Bukhúri 18G2. vol.l, p. 114. line p. 115. line 11. p. [ I |, line (>, p.iqfi, lilií' il\ p. jo",, line 14.
32. Bukhári líJhQ, vol.l, ]>.J;>. line i>, p..1].?, line <>. p.2") 1, line 12.
•53. Bukhan iiife, vul.i. p. 127, line 1 r. 34. Hillenbrand 1994, pp.40, 42. Lammcns (191 r, pp.30-4, 240-50; iqyG, passim, rsp. pp.tr,  (>) and Pcdcrsen (10,80,, p.ti.pi, cols a b; are strongly
influenced by (.aelani's argument hul, unlike him,
stress dial die mosques of Islam, like pre-lslamic
-ancLuaries :'in Islamic tradition1, were from die first
loci of military, political, and social activity, as well
as places of'prayer
35. (resurll ipíiq. vola. I, pp.43 4.
3Í5. Quř''(JHXXX.-Vj.
37. Ouc'iiM xiax.4- y
jeremy JOHNS
corner, a smaller portico or was built to shelter the poor and homeless Muslims. A (lor the change ol die (fib/a in '2/024, the structures within the courtyard were rearranged: the ^ulla was transferred to the south wall, the sujja to the northeast corner, and the principal entrance Írom the south 10 the north. Onto the exterior of the eastern hall ol 1 he enclosure were, built dwellings the I he Prophet's wives, two al lirsl. increasing to nine by the lime of his death; none was built onto the western half. Access to these dwellings was only possible from the courtyard.;"
Tn lour particulars. Caetani s reconstruction ni (he Prophet's House departs horn the ethnogra|iliic model upon which it, is purportedly based, hirst, the area of the courtyard is huge. 2,500 square metres, far greater than any purely domestic dwelling is likely to have been, and far greater than the cramped courtyards of the Arab houses thai Caetani himsell describes.l!! Second, instead of die single entrance of Cactani's dar, which typically would have been a bent-cntrancc to conceal the interior of the courtyard Írom the street, the Prophet's House has three (and later four; gates. Third, (lie dwellings oi the Prophet's wives open directly onto the courtyard, the public space of t he structure, exposing the houses to public gaze."' Fourth. Caetani placed the apartments of the Prophet's wives against the exienor oi'thc courtyard wall: this is architectural nonsense: the structures surrounding the eourlyard should be built against the inside, not the outside, of the enclosure wall.11 Caetani stitched together a highly selective reading oflslamic tradition and a distorted ethnographic model and, Frankenstein-like, gave life to an architectural monster.
("RES WELL AND THE 'HOUSE OF THE PROPHET1
Tf Caetani played Frankenstein, Creswell wasjames Whale, the man who made the picture of the monster. Most students of Islamic art and archaeology arc able to reproduce the plan ol the Prophet's I louse published by Creswell in Early Muslim ArdaU'cktre jfio. n]). Many are familiar with i he isometric projections and reconstruc-
ts. Caetani : ii}0.y. voti. Iip.377 <|. ;.;(). Samhudi ufjbi), p 107, lint's ()  10 Ov ]<][>ti. \ol.i. j).'j ;M. lines \-±   i |: repons a divergem tradition, attributed ui Kliäiip lbn/avd il>n J halm, that the eourKard origin,lib measured 7(1 In bo cubits aipproxtmaleh ;[-> 111 by ;-jn in). See. ( In-swell icibo. vol.1. 1.15.7. n.y Akkoudi 'nyij- piyjí!/, 301 i; ur»ucs tluit tlie Prophet enlarged the mosque alter die cajHure ol'KHa\bar in -/ď:>b' í), Hillenbrand ■ [(jc).|. p -',<>' I"1'1' ahcadv | sol n leel out that the size ol' [he eourlyard indicalcs dial i( was noi lliai ol an ordinar\ domestic clw c llin.Ľ.
40.  ľhe Iradilioni^Ls were apparently uneasy about
this LinseenlK laek ol'pi iv;tcv. and iqjiu'l dial L'mai .Salama built a imirFbriek wall in front other house '10 obstruct the g;i/es ol the jieojile'. but this soirv smv- prinrijjally as a \'eb]ele tor die Prophet"-condemnation oi buildinty. Ibn Sa'd 190]. \ol.i. 2. j). 11li, lines $ I!.
ii. CcoU'rev King kindly informs me dial although compounds with some structures built onto the outside oldie eourKard wall are not unknown in Arabian domestic architecture and mav be found, lor exai nple. a [ I la'il. he has i lever seen a compound re-eniblin^ the Prophet's 1 louse as reconstructed by Cicswcll.
74
THE 'HOUSE OF TFIF. PROl'HKi' AND TUE CO.NOEF'I Or THE MOSOl 1'.
tions of thai plan published, amongsl others, by Leacroft, Kuban, and Hillenbrand (FiG.rrya. c). Some even believe such reconstructions to be based upon archaeological evidence. Only Akkouch, or so it would seem, has critically examined Crcswcll s sources, arid produced his own reconstruction and plan (no. 16): his argument may be naive, but it is based upon a scholarly examination of the Arabic sources, and it is shameful that his telling criticisms ol CreswclPs reconstruction continue to be dismissed, in Sauvriget's words, as 'le point de vuc musulman traditionef. '"■
Creswell's account, of the Prophet's House is derived almost verbatim from that of Caetani, except in one important detail. (.ireswell places all nine of the houses of the Prophet's wives adjoining each oilier in a line on the exterior of the east wall ol the mosque. In a note to the text, he gives the sources for this reconstruction: 'L'mari's Masälik al-Ahsar: Samhüdfs iüudäsat al Wafa in (he Biiläq edition and as 'translated and epitomized by YVüslenfekP; DiyärbakiTs la'rikh al-Iihaims; and ( ael.ur.'s account of the Prophet's Mosque in the Annafi dell'Islam. '
Of the three Arabic sources, two may be swiftly dismissed. The brief and highly synthetic account of 'I.'man (d. 74?»/1^48") contains nothing of direct relevance to the location of the houses. :' Diyarbakri id. 982/1574) transmits only traditions to the effect that the houses were 'spread out' {muslaSra) to the north, ea.sl and south of the courtyard."'
Hie third. SainhiidT (d. 911/150(1], is the single most iniormativc source upon the hislory of the Prophet's Mosque. The full text of his history of Medina was destroyed in the lire of 886/1481, but not. before the author himself had made two abridgements: the Wafa' al-Wafa , in which traditions are presented systematically and critically discussed; and the Khulasat alAVafa', a much more synthetic epitome in which SamhOcll docs not. hesitate to advance his own interpretation of (he sources. The Khuläsa reports traditions that place the houses upon all sides of the mosque except the west. Kot a single tradition in the khuläsa places the houses against only the east wall of the enclosure. In the Khuläsa. Samhudf s own opinion seems to be that the houses were distributed on all three sides except the vvesi or, possibly, upon the north and east sides only.';; Only in the Wafa fd-Wafa. after reporting the confusing testimony of the scholars of the slra (alii al-siyar), viz. that the houses lay on all sides except the west, whilst their doors opened, directly into die mosque and to the
\:>. Akkoncli nyyy Smivagii [947. \>.y. U.-2. 43. Sic! Wü-viciifcld's German summary was iutually ufthe Wafa'at Wafa-I ■[. ( aeswell iq(k), veil.1. i, p.K, 11.1: 'L'maii 19a |. voll, p.i2(>, line 12   p. 127, line % Sanihudl iJifili. pp.nil) u. i'j(> !5. p.14,^, lines 27  30. Wüslenield 18(10, pp.fjo  I. W>  }J,/i'. Diyarhilkn 1!.)}(;,, vnl.i.
p.-jt-io. line-3. Si  12, 10       Uaeiani :c)<rj. vol. I.
l>l'-:?77 1<S-
■\~y 't man uyi]. vnl.i. pp. vi[\ 7. si). Diväiliakri iM>j. voll, p ;;9<i, lines 19 57. Sauvagel i() 17, p| 1.J7 " 9. 48. Samhüdi in(iM. p.iad, line yo   p.:";, line
JEREMY JOHNS
-KuYfa
c
			
		1	a
	Bab JibriL--	1 <Z	c\l
—Bab HAuka		c:	&J
		1 ie 1	1
B
----Bull ALLkH
A
Bab i Jibi-il T
1S0 - 77 7m--
7>7
it
14. Reconstruction of die plan of'Muhammad's House [alter Crcswell 19(H):
(a; Before change of qibl(v, b. after change itfqibh. a. (he Prophcl s house: b. enlargement oi 'Umar; c. eidargcmenl ol ' L lliinan; [n '■ \\ iiKims of mudbnck, roofer! with palm branches and mud: [y. ;<)) nxims ol reeds and mud. roofed willi palm bran*'lies and mud ;(Ircswcll never published die section 1 .•
J
K
CAB £N-HI\
H
AL-K15LA
T—
ii
lw
ll
u
ll
1 £
£L-HUZAKfAF At $!&LI
AR- RAHABA
Bid AR-f)AHMi
(i. Reconstruction ol LI if plan ol ibc Prophet's mosque and his dwellings (alter Akkouchc 19^,")';. I louses of: a. 'A'isha; u. Sawcla; <:. 1-atima; o. Hafsa; ic Umm Salarna :'?!; 1 al-'Abba.s; 1,. Abu Bakr; ir. 'Abd Allah ibn 'Umar. (Jalcs: :a! Bab 'A'isha: :b, Bab 'All: !r i Khawkhat Abi Bakr. :'i': Prophet's place of praver \nm.mH/i\\ \i \ lu.s mmbtir
TTTF. 'HOUSE Ol- THK I'KOl'HET' AND THE (CONCEPT OF THE MOSQUE
if,. Isometric recon,strur1.ions oft.hr Projil ifl.s bouse: Pop lo boitom: after Hillenbrand 1994; Kuban 1974; Leaeroft & Leacrolt 1(176}
[h.RKMY [OIIN'S
vvcsi — docs Samhiid! cite a.l-Khaub ibn Hamla/s comment that this could mean dial the houses all lay to the east; a comment not included in Wiistenfeld's summary of the Waja'.'" Crcswell believed that it was the Khulasa that Wustenfeld had 'translated and epitomized', while his German summary was in fit:! of the Waja alone. Clearly. Crcswell confused the two works, or was unaware of the Waja': in either case, be cannot have known of al-Khattb ibn llamla's comment. In short, as Akkouch so rightly concluded, 'Pes logemenls ifelaienl pas tons paralleled a la mosquec et sur un cote';llJ none of the Arabic, sources cited by Crcswell as his authority lor placing all the houses on the east wall of the enclosure supports such a reconstruction. Akkouch's own reconstruction (ViG.ib) may or may not be historically accurate, but al least it is based upon die sources. If not from the Arabic sources, whence did Crcswell derive the idea ? ''
The only other authority cited by Crcswell is the account of the Prophet's Mosque given by Caela.ni. The crucial passage in the Annak is at first sight ambiguous and could easily be misread as saying that all the houses lay against the cast wall of the mosque.'" Crcswell made precisely this misreading; yet. although be never slates il explicitly, he did have one further reason to place the houses all on the cast side of the enclosure.
Crcswell quotes at length from the T'ibaqat of Ibn Sa'd, who had from al-Waqidi, an eyewitness account oi the mosque immediately be I ore it was destroyed on the order ol TJmar ibn 'Abd al-'A/i/. in 88/707. 'Abd Allah ibn Yazid al-Hudhali reports '1 counted nine houses with their chambers i'addadlu iis'ata abyatin bi-hujari ha}'.''' Ibn Sa'd, apparently following his source, assumed that these corresponded to one house lor each of the Prophets nine wives, and that these were all the houses. Crcswell shared this assumption. Neither takes account of those houses built to the north and south of the mosque, nor those incorporated into the mosque by TJmar and 'Utlimaii.
In /•.••'./. Muslim Architecture, immediately above the quotation from Ibn Sa'd, Crcswell published his plans of the Prophet's House (fig. 14). In one (fig. 14b), he shows the building after the change of the qibla, and also indicates the enlargements of the courtyard carried out under TJmar ibn al-Khattab and TJthman. According to ihe plan. TJmar extended the mosque 30 cubits to the north, 20 cubits to the west.
.[.<). Samhfidi mjoU, yol.i. p.,')"-). lines id _n.
30. Akkouch 1933, p.31.)).
31. Civs well did not read Arabic and. wlicn published tr.inslations. epitomes or secondary accounts were not available to him, he relied upon lus Aral) students to read Arabic sources tor him:
1 lamilion iqqi, p. i3-'.
~,2. C .aela.ni iCjof,, vol.I, ]>-3;<i: 'All oi these ...
7°'
dwelling's ... arose on ihe eastern side ij.r. hall ; ol the mosque ... none was built on the western part' (' I"lil.l.i qucMi ... appui lainciiLi ... sorscro -ul lato orientale della mosehea ... noil ne cosirui nessuna dalla parte oe< identale':.
33. Ibn Sa'd qo.[, \ol.L •'>.p.it!o, line -27 p.jpn, line I.
Tilt 'HOUSE O). THE PKOPIIF.T* AX 1J THE CONCEPT O I" 1'IIE JIOSOCK
and jo cubits to the south. Had die houses of the Prophet's wives been built against any of these walls, l.hey would have been destroyed in the enlargement. Since 'Abd Allah ibn Ya/id saw all the houses in fKi/707, vlicv can only have stood against die east wall of the courtvard. This, as much as his misreading of Caelani's aeeonni. is why Creswell reconsl ructed the plan of the Prophet's House with the dwellings all built along its eastern side.
CreswclPs reliance upon the testimony of 'Abel Allah ibn Ya/.!cl. that he saw all the houses of the Prophet's wives as late as 88/707, focuses alienlion upon his reconstruction of the history of the mosque in die interval since the death of the Prophet. In most respects, CreswclPs account of the enlargements ol the mosque under 'Uinar and 'Uthman adheres eloselv to the testimony of the sources:'' the plan of the mosque was not. altered by Abu Bakr: " it was not until 17/638 (hat 'Umar demolished die Prophet's Mosque, extended it on three sides to make an enclosure 140 cubits north-south by 120 cubits east-west, and rebuilt it using brick, stone, and timber, as well as palm trunks and branches:"' 'l.'tlunan had 'f'mar's mosque demolished, enlarged to the north and west to give an enclosure 160 cubits iioilli-south by 150 cubits east-west, arid rebuilt it in dressed stone, rooled with teak;" neither 'Umar nor 'I. thman had the mosque extended towards the east. Bui Creswell docs not discuss 1 he traditions, including those reported by Ibn Rust a and quoted below, that claim that die qib I a wall was moved first by 'Umar to a line later occupied by the southern columns ol'al-Mahdf's maq.mut, and then by Tubman one bay blither south to its prc-moderu position. Nor did he take account of those traditions which specifically contradict his account of the location of at least some of the houses of I ho Prophet's wives.
Ibn Rusta report's (wo traditions to the effect that it was 'Umar ibn al-Khattab who incorporated into the mosque the house of his daughter Ilai'sa. the Prophet's fourth wife:
It was reported from 'Abd Allah ibn 'I Jm;ir ibn Hal's; he said "Urrtar ibn al-Khat(ab ... extended the wall of the qibla to (he columns which arc where the maqsiiin is today'. Then 'Uthman ibn Allan enlarged it until it reached the present wall. He sairi '[ heard mv father say 1'When there whs need of the house of I lafsa". she said "How am f to make mv way to die mosque?" He fl "marj said to her :'We shall sjivo you a house more spacious than yours and make you a belter way than yours". And he gave her the compound oUUbayd Allah ibn 'Umar, and it was a mirbad.''"'
It was reported Ifom 'Abel al-Rahnian ibn Sa'd 011 the authority of his masters that 'Umar ibn al-Khattab extended the wall of the qibla 10 the itiat/sum: Ten 'Uthman ibn 'Affan extended it to its present position. And he brought into |thc mosquej the t>reater
Crrswell Kibcj. vdl.l. 1, pi'.JJ 8. 35. Samlnlxlr iJibH, p.i-^i, line       p.inline i; Samliiidi i<)oi], vol.i, p.341, liries I (>, 9 rn. ')<>. Samhiitli Ji'b.'l, pp.i <p   j: Sambudi 1908,
■,..l.i,|)p-34i y2.
57. Samhneli i8(J8, [")pa;51 7; Samhudi njob. \(tl.t. |"JJJ.;|")4 be
jH. hni-lhe nih'had, see pp.iii  <>y. In Ion.
79
p.RKMY JOHNS
part of the compound of'Abbas ibn'Abd al-Miitlalib, that to the south, north ami west.: and also he brought into |lho mosque] the houses of TTalsa hint 'I "mar which lav to (he south. And the mosque remained in that state until it was enlarged by al-Walid ibn Ahdal Malik.
Both traditions locate I lafsa's house some distance south of the mosque.' 1 'hey occur in
a passage cited by Creswell, they explicitly contradict his insistence on placing all the
houses against the eastern wall old he enclosure, and yet he docs not discuss them.'"
Samhudi makes the following comment upon the traditions concerning T'mar
ibn al-Khattab's enlargement of the mosque:
... some ol lite chambers of"the wives of the Prophet were on the north side [of the mosque] ... 'L mar did not. bring any of" them in [to the mosque] hut, on contrary, al-AValid brought iheni in and thus 'Idmar left unchanged what was on the north side of [the mosque] standing just as it was, and extended the mosque right up 10 them.""
In other words, Samhudi believed thai, some of the houses stood a considerable clis-lance north of the original enclosure so that even after 'Ulnar had exlended the mosque 30 cubits to the north they still lay outside its walls. Again, Creswell cites the pages of the Khulasa in which this passage occurs, but does not discuss it."1 Such reports, however, are clearly in conflict with his principal source for tfie reconstruction of the Prophet's House: 7Vbd Allah ibn Yazicfs claim that he saw- all\ht houses of the Prophet's wives in a. line along the eastern Hank of the mosque as late as 88/707.
To conclude this discussion of the 'House of the Prophet', the following three points make clear- that the Arabic sources cited by Creswell as authority for his reconstruction of what he calls Muhammad's house explicitly contradict that reconstruction:
1. Creswell placed all the houses against, the east, wall of the courtyard, but most traditions agree that the apartments of the Prophet's wives lay on three sides of the mosque, except the west;
2. Not all the houses were built onto the perimeter wall of the mosque: the house oPHafsa lay well to the south, other houses la)- at least 15 in to the north, and Saliva's house lay a considerable distance from the mosque;''2
3. Tt follows that, according to the sources cited by Cactani and Creswell, the Prophet's Mosque and the dwellings of bis wives did not all belong to a single ensemble, as they believed, but rather comprised at least four, and possibly more, distinct and separate structural units, what the sources call 'his mosque and his dwellings' [masjidu-hu wa-niasakmu-hu}.''"'
59. thn Rusl.i        p.6", lines 3 ib.
do. Samhudi 1908, vol.i} p.350, lines ro 13:
see. SairihfklT 18H8, p. 133, lines 12 13,
hi. The passage does 110L occur in Wfisleiiteld's
epitome.
(J2. Samlindi 1908. vol. 1, p.3^6. line itc end; ef. Bakhan 186V. vol.11, pp.505 7. (J3. car. Ibn Hisham 1858, p.338.
80
tt1f, 'house of t1if prophet ant) the concept of the mosoijk
MIRE AD, MUSALLA AND MASJID
The demolition of the Caetani-Creswell reconstruction of the Prophet's House in the previous section raises [.he possibility that it may have been the Prophet's Mu.sqae that was the prototype of the mosque in Islam. According to Islamic, tradition, the structure that eventually became the mosque of the Prophet was, at first, nothing but an open enclosure with an unbroken wall in die direction ol the qibla (north) and with gates through the other three walls. Could this simple structure have been the origin of the concept of the mosque?
Comparable hypaethral mosques are widely known throughout the Islamic world, and seem to date from the ist century until today. It is generally assumed that they arc simplified versions of grander structures, and this is surely true of modern examples, Ancient hypaethral mosques, by their very nature, are difficult to date without excavation, and I know of none fully and competently excavated."1 Reports ot some of the first mosques to be built in the conquests, such as the earliest mosque at Basra, describe tbern as simple open enclosures, directly comparable to the mosque of the Prophet.10 An interesting group of hypaethral mosques in the Negcv has recently been published by Avni, who suggests that they can be dated to the 7th and 8th centuries ad because they are built on or near settlements of that date (fig. 17).A similar mosque in the Wad! Shlra in the Jordanian Hisma was described byjobling, who dated it upon circumstantial epigraphic evidence to 107/7^5-26."7 There is at least a possibility worthy of further investigation that the openness of these apparently early structures was in some way intrinsic to their design. (It is no less possible, however, that such mosques arc relatively recent structures in which, for decoration alone, stones bearing prc-Islamic and early Islamic inscriptions are reused. This is certainly the case, for example, with the recent 'tomb-mosques' of the Ahl al-Jabal in the Harra east of thejabal al-Duruz.;
There was little to distinguish the earliest stage of the Prophet's Mosque from the nmbadof Sahl and Suhayl that preceded it upon the same site, and that Ibn Zurara is said to have tised as a masjrdbefore the hijm."" This formal similarity between mirbad and masjidseems to be more than mere coincidence, for several traditions report that, belore his mosque was built, the Prophet would pray in marabid.*'" There is disagreement as to exactly what purpose marabid served: most traditions suggest that they were used as pens for camels, sheep and other livestock. The mirbad of Sahl and
6.]. For a lively discussion of I lie problem, sec Helms I'jyo. pp.73
65. Baladhurl 1863, pp.346-7; sec also Creswell iq(k). vol.]. 1, p.22. ()('). Avni 1994.
67. _[oblins> K)8q. 1 lie inscription is now published: Hoyland 1997I), pp.<|/ n>o.
68. 1'or the story in detail, sec pp.103 1-, below.
69. Bukhari [86:-!, vol.i, pp.70, 119; val.3, p.48. For others, sec Wensinck 1936, voJ.11, p.211, s.v. marabid.
81
J F. R f M Y | (• 11 X •
Suliayl is also described as au area lor drying dates, and this explanation is also given oľtlic mirbad ai Qttbä'. later a mosque and a rival in Tirstness' to the Alasjidal-.Nably.7" The dictionaries specify that this latter meaning is peculiar to die dialect of Medina, equivalent to the Yarn am mistah and xhc jarín of Najd. Pane cites Abu 'Ubayd al-Qasim as his authority that both mirbad and jarín are the Hijäzi equivalents for the andar of Syria and the baydar of Iraq.71 The primary meaning of anda>\ baydar and ■jarín is a threshing-Moor for wheat and other grains, which suggests a tentative link between the mirbad and the threshing-floor. Jarín. miAlah and mirbad all convey the idea of Hat. smooth surfaces and it is possible, of course, that mambidvterc particularly well-suited as places of prayer because, they had clean, level floors. :' In addition. Mi resiling Moors would often have been located on high places to catch the breeze, and high places were often sacred sites in ancient Semitic religion. Thus, it would be possible to advance what Wensinck calls a rationalistic explanation for the use of such open enclosures as places of prayer.' *
Such a rationalistic explanation cannot fully account for the rich associations of masjid dnd mirbad which may be observed at both Mecca and Medina. When Ishmael first arrived in Mecca., he built a dwelling on the spot later occupied by the Karba and 'built a circular hedge of doom palms around it and turned it into an enclosure for his sheep1.71 The Ka'ba itself, as built by his lather Abraham, is said by tradition to have been nothing more than an open enclosure surrounded by a dry-stone wall the height of a man.77' At Medina, as will be discussed in detail below, Muslim traditionms stressed that the mosque of the Prophet, like the Jewish Temple, was founded on a threshing-Moor/drying-Hoor, and emphasised the point by comparing the ?ulta of the mosque to the 'booth of Moses'. This booth would seem to associate the Ark of the Covenant (which was eventually placed in the Temple founded on the threshing-floor of Araunah thejebusite) with the Sukkôt booths built (on the anniversary of the consecration of the Temple; to celebrate the Jewish harvest festival. There arc other indications that the mosque at Medina perpetuated the association of the site with the celebration of the fruits of this earth: the raada or 'meadow' of the Prophet which lay be I ween the lorn b and the rninbar, the 'garden ol f aliina' [husia?!, al-sayyida Fatima) in the courtyard of the mosque; and the terms for the courtyard [al-rahaba) and for the enclosure around the tomb of the Prophet [al-fugir), which both have the primáty meaning of a pen for animals or a. place for drving dates."
70. I .eeker 199.1., pp-79 80. 93 |.
7:. Lane 1883, p.ioio, s.v. r.itrbmhm.
72. Sir also 111 Lpigraplnc South Aiabiaii wuitls
from the root SI.L, perhaps cognate v villi S1A 11
: — Arabic .S'/. II', whence uni.yi!//7<. probably
connoting 'to cover with flat stones or plaster':
u 2
líeesion ii|ti'.>, p.i.|«, s.v. SI,I.. 73. Wensinck 1917.
74.. Ibn Khaklun 187,8, vol. 11. p.21 b. lines 12 14 7-,. .Vniqi [tí-jíi, p. 10b, lines ir>  n.'jabari 1879, vol.i. p.] 130.
71». Al-ralfabm S/imhOcli 1908. vol.J. ]>.■.;■■;!{: el. t.ane
lilt 'HOUSE Of THE PROPHET' AND THE CONCEPT OF THE MOSOIT
In his article on the mum/la. Wcnsinck suggested (hat such associations with the fruits of the earth might he the common thread linking these two open enclosures, masjid. and mirbad. to the third, musalla." The musalla is often a large, empty, square enclosure provided with multiple entrances and an unbroken qibla wall.7" All that distinguishes the musalla from tin: mirbad and from the typical hypacthral mosque is its much greater size. Ai Medina, the Prophet is said to have used a musalla which lay southwest of the town for celebrating the two 'ah; on both occasions, the Muslims were preceded to the musalla by Rilal, bearing the 'anaza or spear, which was set in the ground as the Prophet's sutra. Some traditions identify the Prophet's musa/ld with the mosque of Cuba' which, like the Prophet's Mosque, is said to have been originally a mirbad. It. may be that traces of an ancient relationship between the mirbad and the musalla survived until recently in Morocco, where the extra-urban musalla was often a threshing-floor,7'' and where 'anaza referred to an external mihrab in the courtyard of a mosque or in a musalla}'"
Wcnsinck suggested that, in pre-Islamic times, several rites, including sacrifice,
i&'h,. p.iorji-1 nl.r p.i(i-)2. rol.a. A!-!i:.~tlr. Samhfidi 18G0, pa.pp last line; p.i-|5. lines 1 2: el". Lane itSti;-;. pajilli. s.v. hipviin. Impmluit.
77. Weudnek 1917.
78. Hillenbrand 1992, p.660, col.a.
79. 1 Vhiuc 190!', p.4(1-2.
Hn, Miles 1994, p.4<ij, col.I), citing an inscription ol 3^4/' Tji) from i lie 'unurji in the coiirKard ol'lhc mosque ofal-Quwiyiii. i'ez {= Repertoire 1931, vnl.VUI, no. 50911.
|1.k;.m\ JOHN"5
had been performed in an open enclosure, known variously as mirbad. muspd and musalla, and sought the association between such rites and the enclosures themselves in the special connection of the latter, in their guise of marabid, with the fertile earth.'11 As it stands, and with particular reference to pre-islamic Arabia, VVcnsinck's hypothesis is supported by too little evidence to be more than intriguing, bul three points are worth stressing.
First, the fruits oi this earth and of paradise figure largely in the (htr'an, and are explicitly' associated with the Alasjid al-llaram. The just will be rewarded with heavenly fruit."'' God alone causes the fruits of this earth to grow."'1 They arc amongst the 'signs' \ayai] to those who believe, or have knowledge or wisdom."1 The Icrlilily ol the dead earth is a sign oi the resurrection to come.1" Abraham prays that God may grant 'fruits' to the people of His Baytk"'
Second, hypaethral temples were a feature of ancient south Arabian religious architecture,'57 and, as we have just seen, Islamic tradition suggests that open agricultural enclosures had a special religious significance in pre- and early Islamic times. Islamic tradition may also preserve fragmentary details of pre-Islamic hypacthral cults, for example, when the Azd and Gbassan went on pilgrimage to the idol of Manat near Qiidayd, 'they would not go tinder the shelter of a roof until they had completed it1."" Other traditions associate open shrines with sacrifices of crops and cattle,"'' and with prayers for ran i, saldi al-uLsqa'."" The Prophet sacrificed in the musalla at the 'idal-kabir, and, when he led the salat d~istisqa\ he did so not in the covered mosque but in the open musalla or upon the summit of Abu Ojibays."1 In 17 18/638 9, when 'Umar ibn al-Khal.tab led the Medinans in saldi, al-istisqa\ he conducted the congregation out of the Prophet's Mosque and into the countryside.'''-' The idea that rogations for rain had to be made tinder die open sky survived into medieval arid modern times, and the tombs of saints who bad theprerogahve of istuqa have an opening in their domes."1
Third, the poets celebrated the power of successive Umayyad caliphs to bring rain and to restore fertility to the earth. That this is more than empty panegyric is demonstrated by reports that the caliph hirnselfled the salatal htisqd'.''4 This suggests
81. \\ cnsinrk 1917.
i'<2. (hir'tm Tl.23, XITL3J, XXXVI.57. XXXVT1. [2. XXXVIILjI, XLIII.73, XI.IV.j5. tn.22, i.v.32, LV.34.
i.v.lifi, iv■ i.;>o, lvi/jy, 1.Kxvti.
83. Qur'an XTV/52, xxni.iq, XT.1.47, i.v.11-12,
LXXX.24-3I.
. (htr'iin vi.ijo, xin.;-}. xvi.i 1, \v1.fi7, xvi.l.ii). xxxv.27 28.
R5. (hr'dn xxxvi.33 ;vi, v! 1 ■ =)" 58-
Mb, (,ho 'an n, 126'.
o ,
87. Srhniidl it)f!7, p-79-
f.i'.l. Azraqi '858, p.73; d'. bahd K|bJ!, pp.r J3 (>.
89. Ibn I Iishain 1858, p.53; cf. Qar'aii vi.137.
90. Ibn al-Kalbi J<)y<(., p.tt. t'or istirqii' sec: I'abd 1974!): Gold/ilicr 190b. pp.308 i2.
91. Bnkliilri i8()2, vol.i. pp.23b, alii; Ibn Sa'd 1004, vola. 1. p..sp Cf". i'alid 197.|.b, p.'jyo, cols a b; Wcnsinck 1917, pp.7 ().
();>. Taban iy-(), vol.). PPCJ575 b 93. I'alid 1971b. p.270. eol.b.
THE 'lIOUSF. OF THE PROPHET' AND THE CONCEPT OF HIE IIOSOUE
why Umayyad patrons were so ready lo take over from pre-Islamic Christian and Jewish culture the celebration in art of the fruits of this earih. The llieme dominated the mosaic pavements of churches and synagogues throughout Palestine and Transjordan in the jlh 71I1 centuries, but reappears in the Umayyad mosaics of Oasr al-llallabat. Oastal, Ousayr 'Amra and Khirbat al-Mafjar, and in the floor-fresco ofCJaia from Qasr al-fiayr al-Sharqi." In the audience hall at Khirbat al-Mafjar, pride ol place is given to an enigmatic image of a sprouting fruit and a knife, which may be linked to both the fruit and knife common in church pavements and to the ethrog and lulab ubiquitous in synagogue art; the latter, of course, belong to the celebrations of Snkkotr"'
This excursion from masjuho mirbad to musalla'lias raised the possibility that there may be a formal connection between ancient hypacthral enclosures associated with the celebration of the fertile earth and the Islamic mosque, but the link is at best tenuous and, in the absence of better evidence, it would be most unwise to place too much weight upon it. There are, in fact, good reasons to believe that, at the time of the Prophet, the mosque must have been a more complex structure than a simple hypacthral enclosure aligned upon a qihla. These reasons will be explored in the next two sections; for the moment, it seems probable that if the hypacthral enclosure really was one of the formal ancestors of the Islamic mosque, then it was a distant one, and must be located in a lime long before the rise of Islam.
PRIMITIVE ISLAM AND THE EVOLUTION OF THE MOSOUE
Crcswelfs conviction that the courtyard of the "House ofthe Prophet' was the origin of the mosque led him to envisage its early development as an evolutionary process driven by the functional needs ofthe early Muslims. The starting point was the empty courtyard, first, the sun beating down on the faithful at prayer created the need for shade, the roofed zulia that was the germ from which the prayer hall grew. Next, at the other end ofthe courtyard, the lack of shelter for the homeless mukajuila gave rise to the sufja which, in lime, became the nwaq ofthe mosque. Then, paving was introduced lest the actions ofthe faithful, clapping the dust off their hands after prayer, be mistaken for an intrinsic part ofthe new liturgy. And so on, until he had drawn a crude cartoon-strip of tent-dwelling savages discovering for themselves the most basic structural elements: walls, roof, paving, etcetera. This was what Creswell called 'Primitive Islam', the absence of all architectural knowledge' amongst the
94. Crone [Sr. Hind- iptifi. pp.ft 9, ^3; Rin^eren
!)",. Cbrislian ;md Umavyn.H pavements: I'ircirilln 1993. Hallabai: Bisheh 1993. Khirbat al-Maljar:
Hamilton 1959, pl.xerxa. Qasr al-Hayr: Selilunibeiy,er lollfi. pi. j-y See also Rally below, p. 197.
96. Johns forthcoming a.
JEREMY J' Ml \s
iB. Reconstruction ot Kaila i and the Iht at-lmjim all: i Hillenbrand 199 |;
bedouin Arabs before they came into contact with real architecture in the lands dial they were to conquer from the Byzantines and I ho Persians, fundamental to Creswell's notion of'Primitive Islam1 was his insistence, against all the evidence, even after it had been drawn to his attention, that Arabia at the rise oi Islam did not possess 'anything worthy of the name ol architecture'.'"
Barbara Finster and Geoffrey King have thoroughly criticised Creswell's gross underestimation of the importance of prc-Islamic Arabia to the development of early islamic architecture.''" Here, I wish merely to point out that if the townsmen of (he Hijaz were not tent-dwelling bedouin, and if the Hijaz was rtol an architectural desert, then both Creswell's crude caricature of'Primitive Islam', and his cartoon-strip reconstruction of how lite Arabs discovered architecture, from the tent to the Dome of the Rock in I wo short generations, become highly suspect.
To lake one familiar example: the widespread and persistent architectural ensemble of the mosque with the dor al-imarn a gains! its qihhi wall is first attested by the historians al Kufa in 17/638 ;fig.i8;. According to what should clearly be regarded as a folktale, the dimensions and shape of the mosque at Kftla were determined by an archer shooting arrows towards the four cardinal points, and the dar al-imam\\"is subsequenlly built against the qiblawdM because its treasury bad been robbed and ii was hoped that, the proximity of the mosque would thereafter protect it.'1'' Creswell employed this folktale to historicise his cartoon-strip account of
r.)~, Crrswrll li)lk), vol.]. [, pp.lo 11. cjfi. t'inslri 197   kiiii; 1991
99. Baladliun iB(r.j. p,y-(). '1 aban 1O79. vola. pp.j|!.ii 9, |)|).i'49i -i.
86
T1IF "TIOnSF OF tiif. propiif.t' A nt) THK CONCl-'P I' oj" thk mosoce
|- lijui i' IC). 15; t^'bdad. 1 .; issuer's reel inslri lotion ol ll le roui id cil\ of al-MansLir (after Crcswcll & Allai i i(|o< |!
"Primitive Islam', according 10 which "two trivial facts viz. the marking out of the mosque by arrow-casts and a burglary' are supposed to have produced by purest chance the architectural group that located the throne room of the caliph or his representative on the ipbla of the jdnvi' mosque.1""
Once again, the very sources cited by Crcswell in support of his argument in fact contradict it. Thus, for example, Baladliuri reports that Ziyad ibn Abihi moved the dar al-imam at Basra from the Dalma' to abut the tphlu wall ol the mosque because it was unseemly that the imam should have to pick his way through the congregation on his way to the mihrab."" No 'trivial fact' at basra, then, but an expressive political gesture. Exactly the same gesture dial can be seen, but on an imperial even cos-mological ~ settle, in al-Mansilr's Baghdad (fig. 19) where the caliph's palace was located at the very centre of the Round City, with the qubbal al-khadra directly behind the mihrab ol      jurnt' on the line of the qibla.
What is more, there are hints in Islamic tradition I hat this architectural ensemble may have had pre-lslamic antecedents. At Mecca, the Dar al-Nadwa, the palace built for Qusayy ibn Kilab and 'the only place where Ouraysh could settle their alfairs", lay on the north side of the max/id of the Ka'ba: that is, on the side of the pre-lsla.mie qibla.'"" At Medina, at least some of the domestic apartments of the Prophet were attached to his mosque, and several traditions locate the apartments of Aisha, in which the Prophet himscll is said to have dwelt most ol the lime, against die qibla
100. C re-swell rotio,, vol. 1, 1. pp.24 6. 102. Ibn llisham \\\~}S. pp.tlo, fi;,       I'aivt h)">(.
101. Balfullnin li'ib^, p. ^47.
JKRKMY JOHNS
wall." " The association of the jami' and the dar al-imam'm early Islam thus seems to have continued an established pre-Islamic architectural tradition, and was anything but the chance result of 'trivial facts'.
It is just such 'trivial facts' as the sun on the backs or the dusl on the hands of the earliest Muslims that Creswell employed to reconstruct tlx; evolution of the mosque, bit bv bit, from the empty courtyard of the 'House- of the Prophet". And yet, not only was the starting point for the development of the mosque inTslam patendv a far more complex structure dian the empty hypacthral enclosure which Creswell imagined to be the height of architectural endeavour amongst the primitive bedouin of Arabia, but also Creswelfs deterministic evolutionary process driven by functional need, not human will, cannot adequately account for the rapid and universal establishment of the concept of the mosque within one or two generations of the hijra. Such reservations are upheld by what. is. in all probability, t he earliest evidence for the mosque in Islam,
mosques in i hi: qur'an
The origin of the mosque was not a question that Islamic tradition considered especially important or in I cresting. Islamic tradition never once suggests that the mosque - at masjid was a specifically Aluslim creation, nor that the Prophet was its creator."'1 Passages in the Qur'an were interpreted as references to pre-Islamic mosques, and Muslim scholars simply accepted thai the first mosque in Islam, whichever that was. belonged to a tradition of mosque-building that went, back to Abraham, if not to Adam.1"1 There is some discussion of which was lire first mosque in Islam; most authorities report, that the first Muslims prayed and even conducted the Friday prayer in masajid before the hijra, and that the Prophet himself founded masajid. including the mosque at Quba.\ during the hijra to Medina, before his she-camel led him to the site of the May id al\\abiy.UH' Thus, Islamic tradition is for once unanimous that the structure built for Muhammad at Medina in the year i 2/•} was a masji.d.
In part, of course, this unanimity can be attributed to the ambiguity of the term mas/id. In mature Islam, the word comes to mean both the distinctively Islamic architectural form of the mosque and, much more generally, any place of prayer,
103. Sec above, pp.70 Ho.
io.p Pcderscn iqilp,, p.()45, col,a.
ioy Pedcrwti 1989, p.04 p col.b   p.U <ol.li.
Although Abraham is jjrnerallv held to have built
the Ka'h.i. some authorities [ I'aban 1870, vol.t,
pp.171! 8: Va'fjulji 18'b;, vol.t, p.4; Tha'labi 190b,
p.l';.|'i prefer Adam, to whom is also attributed a
mosque in Ceylon 'llui Khaldun 1858. vol. 11, p.yjq. lines 6 8). See also W'msinrk 192.J, pp.389 90. m(>. e.g. Baladhuri 1863, p.2. lbti Hishatri 183H. pp.335 8. ]bn Sa'd 1904, vofl.l', p:2. liukhart iJ.lb^, vol.i, pp.132-5, l'( 01 •• n 1989, pp.(148 7. Leeker "i'H.PP-?!   1 l(l- es[). pp.t>3- 78 8o&nn.i3 1(1. Sec also below, pp. qi q\>.
88
tue 'house of the prophet' and the cokcef'i  oe the mosque
monumental or not, including structures that, did not conform to the type of die Islamic mosque. As late as the Sdi/i-ph century. Ibn Khaldun could still use masajid to refer to Persian firc-temples [buyilt al-ndr), to Greek temples ihaydkil), and to 'the houses of die Arabs ibuyutal-arab) in the IJijay. which the Prophet ordered destroyed on his raids'."" This ambiguity requires a short investigation into the meaning of masjid'm the time of the Prophet.
The etymology ottnasftdis far from straightforward. In the medieval Arabic dictionaries, mayidh explained as a noun ofpla.ee itnasdar mitniy). derived from the verb sajada. 'lie prostrated himself, and meaning 'a place of prostration'. It is highly probable, however, thai ma.sjid \\'A?, borrowed directly from Aramaic as an isolated noun, and so was not at first associated specifically with prostration. The word msgd' appeal's in the Elephantine Papyri as early as the ^tfi century uc in an oath sworn 'by the Temple {bin-sgil')'","™ and the verb sgd appears in the Aramaic version Book fj'A!ukai\ also from Elephantine, when Ahikar bows clown [sgdjtj] and makes obeisance before the Assyrian king: in the same text, Psarhaddotfs courtiers arc 'those who bow down' [sgdwfty).-m But only the noun msgd' (variants msgd, msgdv' and masgd') occurs in Xabataean inscriptions of (he ist century ad. where it has the specific and apparently exclusive meaning of a stele, statue, or altar dedicated to a god.' " Similarly, only the noun msgd appears as a loan-word in Epigraphic South Arabian, with the inferred meaning 'praying-place, oratory'.''' Very few of the twenty-eight occurrences in the Qur'an ofl.be word masjid, or of die plural masajid, reveal anything about the practices followed in mosques. .Nonetheless, it is striking that inayidis never linked with the verb sajada, 'he bowed down', nor with its parts. On the contrary, three principal verbs describe the activities performed in mosques: dhakaro, to commemorate -the name of God);1'- qdma, to stand (before God;, and do 'c. to call upon (Himi.":; This may, perhaps, reflect the independent entry o[ masjid into Arabic.
Turning from the noun masjid to the verb sajada and its parts, extremely few of its
107. Ilm Khiildful tojll, vol.11, p.U'/t), lines Ii 13.
108. Sachau 191 1, no.33. pap\-rus -51', pp. 11 8 ip. pi.32, f niniad      1.110.33, papyrus 32, p.jo:
(towlev 1923. no.44, pp. 1 17  >'k bin see Porten & Vardeni 1980, vol.11, pp. 146  7, \>~/-j' 'by H[erem] the [<2"ocl] ill/by the place ol proslradon'.
109. Cowley 10,23, I1--1-- lines 13 & 10 resperiively See also Hollij/er Srjougclini; 1993. vol.11,
PP-775 t'-ii o. Ca 111 mean 1930. vol J, PP43, l>«- vol.11, p. i id, Hoftifcrr &Jonjrclinn- f<j< r ~>, vol. 11, p. (1(13. (Znptu Inscnplmnmn Semilitanan 1889,part 2, vol.i: nn.ifii, pp.190 -3 {*- Cooke 1903,110.97. pp.240,-51);
110.17(1, pp.ij; iio.tBf,, pp.209  10 •, = Cooke 1(103, p.238;: 110.188, p.212 : = Cookc 1903, p.^^St: no.iqo, p.213 :'-Cooko 1903, p.238:: no.jifl, p.a^ti !   CiOokr 1903.110.92, pp.238  9 & JausMii iS; SaviiJiiae 1909, vol.1, no.39, pp.2t.14      +'/■ pl.XlJ'. Cooke 1903. no.101. pp.254 j.Jausscn& Savignnr 1909. vol.i, no.82. p.223,: ms$fiif.'f. in. Beesten 1982, p. 123. : 12. Qur'äit xxii.4.0: yntOthauß Ikl >w» Haiti. 113. Qui ''tin vn.29: aqfmti wiijiilm-bmi 'uulu kulh ma'-juhn irn-i'it-itti.
[14. for example, {hir'im iv.102. For uijiiti, sec the ariieles bv R.Tottoli cited in 'I ottoli 19(18, p. 9)3. 11.1.
89
JETU.MY JOHNS
sixty-four occurrences in ilic (hrr'aii refer explicitly to sujüd as pari of Muslim ritual. '' Indeed, in the vast number of cases, as Abdelkader Tayob has recently observed in a perceptive study, the .••.•,',/'.'■ tidy distinction between 'prostration of honour' itaknm) and 'prostration ol'worship' ['ihndd) cannot be sustained; there is what he calls a 'juxtaposition of power and submission iit prostration'.11' In about a third of the occurences, there is no clear distinction between .ntjiid lo God and prostration before 1 lis human agents. There are no less than twenty reierettces to the prostration of the angels to Adam, and I he theocratic foundation of the caliphate is laid when God announces to the angels. :T will create a khalifa on the earth*, and then orders them to 'prostrate themselves before Adam (tisjuffii h-Adamy.'"' The inseparability of sujüd to God and to His deputy on earth is thus the concomitant of the indivisibility al'iddm to God and to the Prophet. What are the implications of this for our understanding of the role of the Masjid al-Xably in the early Muslim community ?
I wentv ot the Ivveitlv-eight Quranic occurrences of the word ituisj'id refer Lo sped lie mosques but. intrigumgly, none refers to the Alasjid al-.Sabir itself. They include: fifteen to al-Masjidal-llaräm,-17 presumably the Meccan sanctuary.11,! including one in antithesis to ol-Masjid at-Aqsci, traditionally identified with what later becomes al-IIaram al-Slianj injenisalcin;1one historical reference to <i!-inasjit( which, from the context, clearly refers to ihe Jewish Temple; '' one reference to a maxjid hu\\\. over the Tomb ofthe Seven Sleepers;1-1 and the pair of the masjul al-dimr and the masjid •/;' ••'.■,••/,, v.. : * The: last reference is of particular interest to this discussion.
The two mosques appear in the latter part o\'Sural id 'laicba, which appears to be largely concerned with the raid on Tabuk in the late summer of q/bjo. The relevant passage reads:
And as lor those who chose a mosque out ol opposition and disbelief, in order to cause dissent among the believers, and as a place of watching (or of ambush) on behalf of those who earlier made war against (lot I and His messenger, thev will surely swear: 'Our intention was nothing but good'. God bears witness that indeed thev are liars. Xevcr stand there. / A mosque founded on reverential fear p:c:v" from the lirst dav is more worthy that you should stand in it ...
(tea- Uadhma iiakbadhiJ masfidan dnvum vai-kujrmi im tqfnqan baynn l-muminmn rm-mädan li-man iy.ira.be ilaha n-a-imiih-hn t/iin tjabki <«< do-yatdifu/ina in cimdnfi it In I faisna aw Hülm yashtiadu innti. hum la kaditnhunn /hi Inqumfi-hi altodtm ia-majidun. uskisa ala I Uimin mm awuvli ran win ahaqqu an laquma Ji-lti ...).1 -•'
it-y  favnb aj'cj, I'-gu ei fmu. all, <Ma ''An nio-;l-t sec also, vna i-i". x\ .l")-;-5'}, xvu.lh. xvm.ji). nn. i iti, xxxviii.t j        tat me & Haids icjfiti, pp.4 j,
07. (hi '™ 11.1 ].p 11.1 ](). 11.150. ir.uji. it.]()(>, 11.j 17, vmi.;j 1, IX.7. IN. 1 r 1. IX.2u, xxii.2-,, xi via.icy & xi va 11.^7.
:itS. Bui (hr'än u.ki',) raises the possibility thai ilic vdiolc town ol Mecca is intruded.
IK), (hi! '(!>! x vt 1.1.
i"o. (jur'imxvii.7.
121. (hi 'an xviu.21.
1*2j. (hr'äu ix.io- 10'f
1-2 }. (hu'tili IX. ii17 i).
9°
THE 'llOESE Of THE PROPHEI'   AXI) THK ( ().\C Kl'T OF TUK MOSOt'K
Without exegesis, the passage is obscure: some who pretended to be followers of die Prophet deliberately selected a mosque with the evil in ten lion of dividing the community, and to be a watching-place, or possibly an ambuscade, on behalf of old enemies of the Prophet. The exhortation never to stand in their mosque implies that this is exactly what some of the Muslims did.'"1' Instead, they arc encouraged to worship in 'a mosque founded on taqiva; a term which seems to be generic,1'1' although it is tempting to identity this with the M■:■■:■■! al-!\abiy,1-'
The incident of the maspd ul-dimr has been extensively discussed by Michael Leeker, who makes a strong case that it is an atuhenlic incident from the lite of the Prophet.1His careful analysis concentrates upon four main traditions which, although they contradict each other on many points of detail, arc nonetheless all composed around a. common core. I'lte masjid at dirari* idenlihed with the mosque built by members of the 'Amr ibn 'Awf a I Quba' on the southwestern outskirts of Medina. All accounts agree that it was intended for one of die leading Meclman opponents of the Prophet. Abu 'Amir Abel Amr ibn Sayli ibn al-\uman, of the Haml Dubay'a ibn Zavd, a subdivision of'Amr ibn 'Awf. Abu Amir is generally known as al-Rahib, usually translated, perhaps loosely, as "the Monk', but lie is also called al-Yahudt, 'theJewV-" He is identificcl as a honll. and is portrayed as defending the traditional hamfiya, the din Ibiuhvn, against the innovations of Muhammad;11"1 the Prophet called him 'al-l'asiq', 'the Sinful'.1 The sources cannot agree as to why the Band 'Awf built their mosque: their alleged motives range from a desire not to walk too far to pray,1 ;i lo pious envy of the builders of auolher mosque,1 " to their reluctance 10 pray in a mirhad where a donkey had once been tethered,1 lo an openly aggressive plot to build a secure fortress for Abu 'Amir, whence he could appeal to the Byzantine /jiiysin for troops to drive Muhammad and his followers from Medina.1:11 But what is absolutely clear in all accounts is the rivalry, not just between Abu 'Amir and the Prophet and between the Bantt 'Awf of Quba' and the Muslims, but also between their respective mosques. One account has Abi~i Amir
l_>.|. Indeed, manv aeeonnls insist thai the Prophet iuitialh approved ol the Inundation ol die inasinl al fhmr. see Leeker it)j),->. 1x75.
il?5. I'or the surprising roiielnsmii that (ai/n a here means iyliiijtl'. see the sourees disi ussed 1>\ I.eeker
i<)9,")' PP-(\)   1 & lm47 .")"■
Leeker [I997, pp.94  ion: argues loiribh. hut run altogether et>n\ iiieiu^h. lliat die nmsjirf a!-Ultra's was the mosque ot Quba'. See also I .eekcr 1997, pp.70 (lo. 127. Leeker 199;,, pp.74 i+tj. i-2tS'. I.eeker 1997. pp.iJ'i. r;;r 129. Ibn Hisham 1858, pp.411 12.
r-jo. \a eter 1 [(93. p.\2 9.
l;.;t. I J'ckcr 199J. pp.li/  91: roiiuiKiitar. ot Miujalil ibn Si;k» man.
<-y>.  I ,eeker 1997. pp.70   do: report alliibntrd to Sa'id ibn Jubayr.
1      I.eeker '■<)<)',■ pp.80       report atl.ribe,le<l lo rUr\va ibn al-Xubavr. As has aheach been observed, the mnstjue at (,)nba' \v;e loinuled on 11n■ sile ol a iiiirh'i/i; <re also Leeker 199s. ppaji-j |. 134. Leeker 1997, pp.bfi  '''<> report aiuibntedto 'Abd Allah ibn 'Lmar ibn al-KJiatlab, d 79/(19^ vi' to 'Abd Allah ibn 'Abbas, d.'.ia/tioj  8: <S: pp.147 b. n.239 aonimenlarv ol"Abri ab|abbrir:.
91
JKRF.MY JOHNS
build himself a mosque presumably but not inevitably the may id al-duar specifically as a stage lor his propaganda against die Prophet.1"'' The incident of the mayid al-difar, as expanded in some of the earliest Si,,; literature, thus portrays the Prophet and Abu 'Amir as rival leaders of opposing communities, each with its own masjid, which is not just a gathering place for the community, but also a symbol of its autonomy, and of the authority and legitimacy of its leader. The destruction of the masjid al-dirat w-as a defeat both for Abu 'Amir, and for his followers who stood in it; a victory both for Muhammad, and for the Muslims who stood in ;thc mosque founded on reverential fcar\
The compilers of the Sim thus had a clear view of what the inseparability of sujud to Clod and to His Prophet, meant for the masjid. Elsewhere, it is reported that the first followers of the Prophet were known as 'the People of the Mosque'. When the poet 'Abbas ibn Mirdas was sent by his father to worship Daman, the tribal idol of the Bam Sulavnt. a voice; from within it said: 'Tell all the tribes of Sulaym that Damari has perished, but the People of die Mosque live on [qui li 1 qaod'i/i mm sulaynrhi kiiJl.a ha a.ivdd daman wa-'asha ahlu l-maxjidi).'VM' But the identification of the Muslims as the a/)/nl-mmji.dk originally Qur'anic. for the Muslims are there referred to as the people iahli-hi} of the Masjid a!-Hamm,,v'"
We must also look briefly at references in the Qur'dv to masajid in general. Most significant, here, are those passages which suggest that the early Muslims had several mosques: one, which regulates the conduct of Muslims 'in retreat in the mosques' during the fast, surely refers to physical structures.1;;; Several passages suggest that the Muslims laid claim to more pre-hlainic masajtd than the Masjid al-Hardtir. other gods but the One xi were worshipped in masajid:'"''' pnlvtheists serve in the masajid of God, but should not do so;1'" and Muslims wishing to commemorate God's name in maxq/id. m\ght be impeded from doing so.111 The last, of these passages would seem, from the context, lo refer to Christians ancljews. and raises the question, also relevant to the mosque built for Abu 'Amir 'the Monk', of whether the masajid belonging to
Kx>- '•"'"'•k't l()()"> PP-t-p") '■>■ n />:j9. i;>!>.        1 lrham lít^íi. p.'ji-j. '1 Ik iv is no indication in the Lcxl which mosque was intended, the M'isjiii'fíi-,\nhn. the Mti.yi'lal-UctrSm, or -,01110 oilier mosque
''37- '™ It--)/. Sec also a description ol'llie liana tlanaslt as the ail! a! w.yw/í/of the mosque of Quba' cited by Lccker 1 1 .,. p.inj. r-;ii. (%'. 'an 11.187: n a-ln mkitshtm-hunna nit mfam 'akfjimt; j; I maw/Mi. See also Qur'ati vii.sK) 31, ix.17  li!. 1.xxii.1íj.
r-;<r (.htr'iiH [.xxii. tIS: iva anna ! itmsujidii L-iiiiin ja-ltf tafiiimu'n Slain uiiaiiau," 1.11 c ■ mosques belong to (iocl
alone,    do not invoke am with find". 1 ]u. Qitr'ail ix.17-18: mfíkana h I mushnktrnt n» ja'murtt waispda '■loin ,.. mntmm w'imiu tiiii><!/nlfí Utilit man ammin bí-Slňiil, :Il is not for polylhi isis to lend llu*mosques of(iod ••■ Onlv lie who believes in God shall tend the mosques ol God'. 141. Qur'an 11.114: wa-mn tt£kitn> mimmnn mana'a ma-iajida Slain■ an xiidtikitrajt ha .•nnu-Sui, 'and who dots ltm'Ucr wrong than lie who forbids |entrance to J] 1 lie mosques ol (iod lest his name should be commemorated there*. The next \crse slis™c.sK that Christians andjo.vs are here intended.
TIIK 'HOUSE OF  I'll F. PROl'H I'.'v' A.\D THE CONCEPT OF ti1k MOSOl't
different religions communities could be clistinguislicd from each oilier. That this was indeed the case is suggested bv the appearance ofmasajidat the end of a list of Hi life rent types of places of worship suwami\ bvya', salaivat. and masajid which seems to imply some son of differentiation, !l~ At the same time, however, the (hir'dn uses tfie word masjid of so many clilfcrcnt sanctuaries (hat it is unlikely that it was yd associated with any particular architectural type.
This short discussion of mosques in the Qur'an has helped 10 clarify die ambiguity surrounding die ierm masjid. It has shown that although masjidcan indeed mean any pla.ee of prayer and need not imply a. built structure, still less the architectural type of the Muslim mosque, the word was commonly used to refer to pre-lslainie place, of worship and, in (he (htr'an, is most commonly tised ofspeeilie pre-Jslamic sanctuaries. During the life of the Prophet, the Muslim community already possessed and laid claim to more mosques lhan the Masjid al-IJamm and the Masjid al-.Kably. AJlhough the physical characteristics of early Islamic mosques are not described in the (jhn'dii, i( seems that the)' could be distinguished in some way Irom other types of religious buildings, some of which were associated with Jews and Christians. If the miifassirs' account of the masjid al-dirar is accepted, then the mosque (or mosques] ol the Prophet acted as an architectural svmbol of his authority, a weapon in the war againsl all who challenged it. In die Prophet's Mosque, from its very beginning, there was no separation of divine and earthly authority.
Seen in this light, the lack of interest shown by early Muslim writers in the origins of the mosque becomes easier to understand. What drove the the first Companions out of Mecca to pray in the surrounding ravines [slii'ab) was not. thai (he masjid had yet to evolve or to be created, but rather that they had been instructed to withdraw from the polytheists in the Masjid at-llaram.11:' The Qur'an associated the Masjid al-Haram with Ibrahim, and thus revealed the primordial origins of the mosque.11 f Bui the Qur'an also revealed dial the Muslims had themselves possessed several masajid from the earliest days of the community. Like Islam itself, die masjid was aboriginal.
THE CONVERSION OF THE MOSQUE
The preceding discussion of the mosques of the (hir'dn has raised the possibility that the earliest mosques in Islam includedjahilt mosques that had simply been converted to Islam. Tli ere is a useful parallel for this sort of conversion. 1 romblcy has recently studied the Christianisation of Syria in Late Antiquity, as part of a much
i-4','. Qur'(i)i. xxn.411.
143. llisham liiy'f. p.ih(i; Tabaii 1879, volu, p. 1 dip,. See also Qui1 an XI.e^.
144. Qur'an n.iBj V27, ui.yti ctf.Thcsc jRissages, of"
course, refer not to nl-Mnyld al llardm bin to r,i-bavt. almost universally understood as tii-linrl <i!-H<uinn, the Ka'ba at tlie centre of at Masjid ai-ilamm.
93
JEREMY JOHNS
broader discussion of die Christianisation of Hellenic religion.' 1 ' The process of acculturation thai he describes bears many resemblances to the Islamicisation of Arab society as reported in the Islamic historical tradition. His account of Rilmdimlamsn-vat/gmvhidcs, for example, the Christianisation of incubation and of rogationary prayers for rain, which may be compared, respectively, to the Islamicisation of istikham and islisqa'.1 " His discussion of the conversion of pagan temples into C christian churches is therefore of potential relevance to this inquiry.' '' Trombley argues that temple conversions    'the demolition or partial dismantling of a sacred edifice and its modification into a church or marlyrion' - were 'the logical consequence of the theological tendency to reeategori/e pagan deities into destructive daimxme.s'.'"; On the borders of Arabia, the demotion of deity to demon was still common in the early full century. At Tzra in the Hawran, for example, an inscription off, 15 celebrates the transformation of a pagan sanctuary into a church dedicated to StCieurge:
The: abode ol'dcniotis has become die house of'God. The light o( salvation shines where darkness caused concealment. Where saerifitcs to idols occurred, there are now choirs of angels. Where (rod was provoked, now He u-. propitiated.1 '''
Exactly the same theological tendency to reclassify jalullgods and goddesses (sing. Hah, ilaha) as malicious demons (sing, shaytan. nhavtawi; sing, punt, is encountered in Islamic tradition. In an account of the expedition sent by the Prophet to destroy the shrine of al-'Lzza at Nakhla, the great /rTfe'/* goddess appears to Khalicl ibn al-Wahd as a demon {shaytaruV- in the form of'an Abyssinian woman with dishevelled hair [sic. but see note] and her hands placed on her shoulders, gnashing and grating her teeth ... [He] dealt her a blow which severed her head in twain, and lo, she crumbled into ashes'.1'"'
The authority for this transformation ofpl/uli deity into Islamic demon is Qur'anic. In the (Mr'an, the word shaylan occurs six times in the indefinite with the apparent meaning of'a demon'; in the plural, al-shavatin occurs eighteen times and always denotes malicious beings of some sort, although it is seldom clear whether humans or jmn are intended. '' In one instance, however, shay tan clearly refers to a jdhiti deity:1 '' al-jinn is twice used in a similar way.1 '5
!.[.",. Tnimblev 199 ;.
i.lb. talid 1974a: I'alid 1974b.
i.|.7.   I mmhlcy 1993. vol,1. pp. 111 17.
1 1 Si. Trombley 1993, voi.T. p.iub.
141). Waddingloii 1870, vol. 11, 110/2 pttt. pp.30';) 70.
Pivim ice I ()(18. no..I ",7a. I Volnlit:'v 1993. vol I. ] 1.104,
•'•I a. pp.;-vi?> '■',">•
17,0.  Ibn al-Kalhi K)7)>. p."i  = Ibn al-lvalbi 1924. pp.147 lis 1 ler hair was net 'dishevelled' but bristling, (hilleil-np. fug liair 'reading: jn-iilhri liwi'd hi huhfislnviiii!! i:rtj:\h(il>n >ha'ru-h>7s perhaps as may be
seen 011 ruck-carvings ol dancing-girls: 11 recalls the 'hairy demons' \<ririm': of 1 be Old'I estanieiu: e.g. Leviticus xva.7: ei. Robertson Smith 1894, pp.120   1. For Islamic demons, van Yloten 1893 conveniently ■>umnnirizc,> the account ol Jahi/. 111 St! IlmdvSdiL
faint 11)98: Rippin 199b. band's assertion that ' 1 here are numerous passages where Mm)xlti>l denotes the deilies of paganism' is to be treated with care. ly. Qttr'an iv.: 17. 133. Qui'an vi. 100 &. xxxiv.40 1.
94
THE 'ilOl SF, OF Tilt PROPHET'" AND TH I' CONCEPT ()!•' THE MOSOl.'E
Muslim scholars went one step further than the Christian opponents of paganism; not only did they reclassify~jrlkilj deities as malicious demons, hut they also reclassified idol-worship as a rile which had developed out. of the aboriginal religion of Abraham, bul had subsequently been perverled. Thus, by the ^rd/cjlh century, both jahill idols and, indeed, the very practice of idol-worship had been incorporated into islamic sacred history. According to ibn al-Kalbi. from earliest times, every pre-lslamic descendant of Ishmael who performed the hajjwould carry away a stone from the Ka'ba and:
erect that stone and circumambulate it in the same manner he used to circumambulate the Ka'bah, seeking thereby its blessing and aflirming Ins deep allection lor the Holy House ... In time, this led them to the worship of whatever look their iiinry, and caused i hem to forget, their former u orship.
In this way, the Arabs grew:
passionately fond of worshipping idols. Some of them took unto themselves: a temple around which they centred their worship, while others adopted an idol to which they olfered their adoration. The person who was unable to build himself a temple or adopt an idol would erect a stone in front of the sacred I louse or in front ot'any oilier temple he might prefer, and then circumambulate it in the same manner in which lie would circumambulate the Sacred House ... The Arabs were wont to oiler sacrifices before all these idols, bam vis and stones. Nevertheless, they were aware of the excellence and superiority of I be Ka'bah. 10 which they wen I on pilgrimage and visilalion. What tbev did on their travels was merely a perpctuaiion of what they did at the Ka'bah, because of their devotion to it.1,11
Thus, Arab idolatry, like Judaism and Christianity, was accommodated within Islamic sacred history as a perversion of the aboriginal religion of Abraham, and the idols themselves were convened back into the stones of Abraham's Ka'ba.
If, with the coming of Islam, jahill deities became mere demons, and the rites observed at their shrines were revealed to have been but perversions or rituals once practised by aboriginal Muslims at the Ka'ba, it should have been relatively easy to convert the pre-lslamic religious buildings of Arabia into Islamic mosques. This, of course, is exactly what is said to have happened at Mecca. Tradition reports that the Prophet expelled from the shrine, and then destroyed, the 3Go idols, bul then purified the sanctuary and converted \^ jahill rites back to the dm Ibrahim..*"
But what of other shrines? Was there a widespread Islamicisation o[jahill sane I italics on the Mcccan model ? We have already seen that several passages in the (Jur'an suggest that the early Muslims laid claim to more pre-lslamic masajul than
134. Ibn al-Kalbi u^i. pp.4, 28 q:'» Ibn al-Kalhi       iraditional account, see Hauling i()H'j with U)'J4, pp.'/. 33!. Cf. Ilm Hisham i<!y">. pp.31 2. comprehensive bibliography.
135. lor a radical rciitlerprelatiun oi'lhis
95
j i.REMy JOHNS
lite Masjid al-Haram.1 ''' ['here are a few olher references to the conversion of sanctuaries inlo Islamic mosques: for example, Ihn al-Kalhf mentions the iclol-block [sakhra murabba'a\ of Allat. an I lie place of the left-hand side of the minaret of the present-day mosque of I a'if\ ant) the white quartz idol of Dhul-Khalasa, 'now the threshold of I he gate of the mosque atTabala'.' " If these are temple-conversions, tlien they arc alone in the whole Kitab al-Asnam; but it would also be possible to argue that these idols were incorporated into mosques either as stones of the Abrahauiic KaTia reconverted to Islam, or as trophies symbolizing the victory of Islam.
A degree of caution is necessary here because no systematic catalogue or study has yet been made either from written sources, or from the archaeological and ethnographic record, of pre-Islamic sanctuaries converted to Islam. That said, my distinct impression is that Muslim writers, apparently against the testimony of the Oiir'an itself, believed thai die first Muslims did not generally convert/I/»7f sanctuaries into Islamic mosques. This is unexpected: either pious Muslims in the 2nd and 3rd centuries had forgotten (or chose to ignore) thai manyjdhill sanctuaries had been converted into Islamic mosques, or such conversions were truly rare. I cannot resolve the enigma, but I point it out as a promising subject for further research.'
To dispel that rather inconclusive note, it is worth repeating that there is ample Qiir'anic authority for the conversion oijahili shrines; that die theological preconditions were fully satisfied; and that the Masjidal-llaram stood as a supremely conspicuous example of both the possibility and the legitimacy of such conversions. This justifies pursuing the quest for the pre-lslamic models for the concept of the mosque.
THE FAMILY OF THE MOSQUE'. SYNAGOGUE. CHURCH, AND BAYT AL-ARAB
Iii the centuries immediately preceding Tslam, different Near Eastern religious building types made use of a peristyle forecourt giving onto a covered space, arranged symmetrically about the central axis of 1.1 ic. complex. 1 hey may therefore be briefly reviewed in search of the origins of the concept of the mosque.
Tn its final phase (245-2561. the synagogue at Dura-Europos was a rectangular enclosure, divided into a forecourt with a single colonnade on three sides, and an assembly hall (fig. 20). In die middle of the west wall of the hall, a niche directed the worshippers towards both the Torah-scrolls and, approximately. Jerusalem.
!;",(>. (hif'im 11.11 j. [x.17-1(1, lxxti.i'II. 157. Dhul-Klialasa: Ibn abKalbi KrjJ, ]>. i() \ ' II111 al-Kalbl 1*124.1'-'t'; Label i<)68, pp.bi 8, esp. p.62 n.y Alia!: Ibn a.l-Kalbi 193a, p.yjG : - lbn al-Kalbi if)-1!, p.31}; Filhtl iqbii, pp.11   jo, esp. p. 1-20 & n.j.
if,)-!. Thai. a. number of;^'/? fortresses .in; . ulum, hisn) were transformed into mosques a tier Islam ;see Leeker K)()5. pip.lb   17; might suggest that, here at least, something other limn sacred space nas being convened.
THE 'HOUSE (JE IHK PROPHET' AND THE CONCEPT OF  THE MOSQUE
20. Plan and isometric reconstruction of 1.1 ic s\ na^ogue al Dura-Fjuropos [;i.fIit Kracling icj-aj;
Immediately, 10 the right of the niche, against the west wall, was the live-stepped 'seat of honour'. In the northwest corner of the forecourt was a water-source, presumably for ritual ablulion.1'''1
The formal parallels between the synagogue at Dura-Europos and the mature stage of the mosque are immediately apparent: orientation upon a direction of prayer [qiblaf the Torah-niche (imhrdb); the 'seat of honour' ',mmbaf;,: the ablutions-facility in the forecourt (mida'a): and, most important here, the arrangement ofi.be whole complex (hall, haram, and axial peristyle forecourt, sahri).1'*' Some or all of these features are found in other Diaspora synagogues, such as Priene"'1 and Sardis (fig.at' in the synagogue atjerash"'3, and at Belli Alpha near Scylhopolis in Palestine.'1'1
The feature of greatest potential significance to this search for a model for the concept of the mosque is obviously the axial peristyle forecourt. But this was the exception, not the rule, in synagogue architecture. It occurred earlier and more regularly in the Diaspora than in Palestine. In Palestine and Arabia, it seems to
l"){.). Kracliny; icj,")(». |>|>.;? ;•}•}. Qffi •>■>: Scaler kh)2. too. Lamberi 1950.
1G1. Kraabel Hjpy pp.107 <), al-o cites earlier bibliography: axial toreeotirt (no colonnade:, orientation, lorah-nichc, ablution basin. t(>2. Kraabcl 1995. pp.101 (i, also cites earlier bibliography: axial peristyle forecourt, apse (west:, aedicula lor 1 »ra.li scrolls added to east, end iorien-
laled onjernsalem 1 in final stage qtti cenuiryi, alihitions fountain in centre ol'lbieeotitl. See also Seagcr 1992.
it);j. Kracling n);jl_!, pp,■.>;$•! - -| 1: axial perist}Ic f oreeourt, recumgnlar apse oi'ientalcd on Jerusalem. 1(14. Siikcnik 19;;^: axial loreconrl (ito rolonnarlei, apse and 1 01 a.h-shrinc orieiited onJeiLisalem.
97
JERF.MV JOHNS
IsomcTrir nvoiislniction nitho svnaengue Lit Sordid (alter Gutman íqiii
have occurred in ihosr regions where traditions of Roman public building were strongest.Although it hits been said that the combination ofasscmblv hall and courtyard was 'an almost invariable lea I arc of [the] synagogues [of CalilceJ V1'1' the claim is doubly misleading. First, the presence of courtyards has been generally inferred on the model ol Capernaum, but there is no archaeological evidence that most synagogues really did have adjoined courtyards. "'■ Second, as the example of Capernaum illustrates, courtyards for which there is archaeological evidence are usually attached to one of the sides of the assembly hall, and are thus neither axial nor forecourts.
At about the same time that some synagogues combined an apsiclal hall with a colonnaded Ibiecourt, so that both elements were arranged symmetrically about I he direction of prayer along the central axis of the complex, many7 churches developed a similar arrangement Tig.22). The basilical church with atrium became a widespread and standard form during the fourth century. A glance at the plan of the final (ph-cent 1 try'; stage of the svnagogue at Sardis ÍFic.yi; demonstrates its formal similarity to this type of church. Indeed, the architectural layout of the synagogue and the church may occasionally be so similar that only artefactual evidence can distinguish one from the other: what was once regarded as the synagogue at Stohi is now identified as a
[ijy Srat;iT tijcu. pp.07 8. McviTs '.()~t>. 107. .Soaker ujp'j. p.q.p
nib. oiiah K)()l, cols. 17tí 7. t()8. Sealer h;<)i>, pp.iy-; ' 5. ( 'apeniaurvi: (lorbo 1970.
9«
THE 'HOUSE O I7 THE PROPHET' AND THE CONCEPT OT THE MOSOt I'.
church,"'"while the 'house-church' at Pricnc has become a synagogue.170
Whether the synagogue with axial peristyle forecourt influenced the development of the basilica! church with atrium, or vice versa, or whether each influenced the other, need not concern us here: scholars arc divided as to whether the two developments are interrelated or merely analogous.171 It is far more important to recognize that the development of both building types was informed by a common concern to emphasise the direction ol prayer; a point of obvious relevance to the Islamic mosque, to which I shall return at the end of this section.
Thus far, f have considered the possible influence upon the concept of the mosque of synagogues and churches buill in the lands thai were to be conquered by Islam, but what of Arabia? No archaeological evidence has yet been found to indicate what the synagogues of the Arabian Peninsula before; Islam might have looked like.1 - We are better informed about the churches of pre-Tslamic Arabia. The basilical
ihq. Kraabrl lOCc,, |). i) 2.
170. Kraahe] i<i<)5, p.107.
171. Kraelin" n}jd: p.2i. n.cji. Seagcr p)[)2, PP-97
172. l.uerarv rvidnicr adesis [0 ilicir cxislence: lni" example, see the iMedinan svnago^ics 111 the slon cited b\ Lceker ic)C)r): pp-4i -■
99
JKKhM V JOHNS
church with atrium occurred widely throughout the limes Arabic us, and may well have been carried beyond the frontier by Christian Arabs, although none is yet known. The recently discovered churches at al-Qusfir Taylaka!, al-Jubayl, and Sir Bam Yas attest rather to the dominant influence of the Ncstorian church in the Culf.iv;i In the Yemen, the church built at San'a' by Abraha, and called al-Qahs (i.e. err/i'sia), seems to attest to direct Byzantine influence (fig.23;. The basilieal nave and domed sanctuary seem to have referred to the Church of the Nativity and to the I loly Sepulchre. The nave and sanctuary-' were decorated with mosaics, which may confirm the presence of the Byzantine craftsmen reported by '1 abarT.''1 Al-Qalls stood within a massive enclosure, but may have been, a haram on the Arabian model, rather than an axial forecourt.1 ■"'
Synagogues and churches were trot alone in employing a peristyle forecourt axial to the covered space. 1 his arrangement may be found, too, in the buy lit al-'arab. the prc-Islamic temples of.Arabia, both north and south. To begin in the south; it has long been recognised that the temple at al-lluqqa offers a strikingly close formal parallel to what I am calling the concept of (lie mosque (ftg.24'i;;t" to l.Iuqqa can now be added others such as Waddum Dhu-AIasma im and the tomb-temple at al-Masajid (fig.25':.''' These temples are separated from the rise of Islam by a chronological gap that may be as wide as half a millennium and which current archaeological research is not yet able to bridge.1'" It is now clear, however, that the yth-een-tury Hijaz was not the architectural desert envisaged by Creswell but, on the contrary, lay on the northwestern edge of an architectural koine 1 crated in the Yemen which stretched as far as Sassanian Iran and Abyssinia.1'"' In this context, it is intriguing that the term MS'CD is attested in Epigraphic South Arabian. ■""
At the other end of the Peninsula, as we have already seen, the Aramaic word [mxgd\ from which the Arabic masjid is likely to derive, occurs with some frequency in Nabatacan inscriptions of the tst century. Although continuity of terminology need not indicate continuity ofareluleclural form, this nonetheless directs the search for possible prototypes of the mosque towards Nabataca."11 I he meaning of the term msgd' in Nabatacan inscriptions indicates not a built place of worship, but a stele,
17;;. Al-Ous:nr: Bernard. Callol & Sall<s 10,0.1: Honiard & Salles 1991. Al-Jubayl: bangleldl 1994. Sir Bam Väs: Kill»' i'l<)/. for Christianity in die Arabian Gulf, see l'otts 1990. voI.it, pp. ifjO 3, 240        :::;: p.
174. Tahan 1879, vol.i, p.035.
175. A/.niqi 1838. p.89. Ci. Serjeant & l,c\ver>rk <<><"»v l'TM-1 <>>
17(1. Ralhjens & von Wissnrwui iu'}2, pp.2", 77,.
Serjeant 1939. pp.. 14,2 3.
177. Schmidt Kj<'2a: Schmidt ujffch: Schmidt 10,&_•<•.
178. :\1-Huqc|<i was destroyed by volcanic aeiiviiv no earlier 1 li.ui die 3rd century: Ralhjens & von Wissman 1932, pp.14 - 16. Current research: Dor 19H3; Ansaty 1981: King 1991: Kcal! 1993.
179. Creswell 1969. vol.1.1, pp.3 3; Grabar 1983. Finslcr 1973, Finster 1991. King lone see also Lammens 1922. p.18 p
180. liceston 1982, p.125, s.v. S'GI).
lib. I owe this obsci valion to Gerald Hawtinjr.
IOO
i'iil 'not si: of nn: propiift' axd the • mm i t • i of ihk mosouf
2[>t. Serjeant— rcvoiitmrtion of Abnilijim's church ot al-Qahs at San'a ialioi Serjeant & Lewcock J<)1>;>'.. 'I he measurements arc given in dhirg'' — 0.4.8 m;. I'he walls were ij diar/l' thick
statue, or' allar dedicated to a god.1'''' In inucli die same way, the kaba of Dhu'l-Shara at Petra, seems to have been an idol-block, not a building.I;s:) As to the temples themselves, in some in the Hawran that of P>a'1 Shamln at Shi' (late 1st century bo}, and those a 1 Siir and Sahr the cella is positioned approximately in the centre of a rectangular peribolos. the front half of which is occupied by an axial peristyle forecourt fh.. . ' The forecourts have benches arranged around three sides under the roof of lite colonnade, and Focrster therefore argues that thev were the prototype for the assembly hall of the Galilean synagogue, and that the cella was .similarly transformed into the Torah-shrine 'which was conceived as a miniature temple, an imitation of the Temple in antiquity :."!'' It seems preferable, however, to regard both the Nnbata.cau temple forecourt and the Galilean assembly hall as independently related to the large family ol Hellenistic-Roman public assembly halls with benches: die forecourt at Shi' is called a thmirtm in the Nabataean foundation inscription.111"
In short, some prc-Islamic temples in both Nabataea and the Yemen, some synagogues especially in the Diaspora, and basilical churches with atria, all bear a more or less distinct formal resemblance to what 1 am calling the concept of the mosque, in that an axial peristyle forecourt leads to the covered space at the rear of the complex. Hut even when the formal resemblance is extraordinarily close, as it is in the Dura-Furopos .synagogue, it does not necessarily demonstrate a direct connection; indeed, in that particular case, chronological and geographical distance clearly preclude it. As for other synagogues, the type with axial peristyle forecourt is too rare, and generally occurs too early and too far away in lite Diaspora, to have influenced the development of the mosque. Similar considerations must, for now-, rule out the
Sniudrl icjji, pp.iir, ti. 1'uu see note 110 above.
103. hpiphaniiF t<ji.r), vol.11, pp.jllli 7 (51 nr. hut following the reading o!'< lelilei
i860, pp.'} 14 5. The report is discussed injolins :1b 1 "lltcnniiiig bi with full bibliography.
1M4. Bnller Hjoy, pp. .;b5 <!-, ''Shi'':, pp.428 30 ;Sui\, pp.441 3 .S.ihi'l. C'T. Sourdcl 10     pp . •: \. ■ «"v~>- Foci sin IuiSj.
l.ilimann 1004. p] >.ijy no: biluiiamiu):71-, PP-7(> 7-
IOI
jkrkmy johns
prc-Islamic temples of Nabataca and the Yemen. Future archaeological discoveries may, of course, bridge diese gaps.
The case of the atrium church is rather different. In the first place, il was undeniably well-placed in space and time lo have influenced the early development ol'the mosque. In the second, that is precisely what it did. Many Umayyad mosques were built by Christian masons and decorated by Christian craftsmen who had learnt their trade in die churches of"Syria. Many churches were converted into mosques. Later wriiers. such as Muqaclda.si, claimed that 'And al-Malik and al-Walid had intended iheir religious monuments lo outshine the churches oi Syria.'"' Hut could the basilical church with atrium have been the prototype of the concept of the mosque':' It is simply the wrong shape. In the church, the central axis is typically three or four times longer than the. width of the struclure. In the mosque, the length and width of the structure arc typically equal, or nearly so; unlike the church, the mosque may be wider than it is long. These strikingly different proportions constitute one of the chief characteristics which distinguish the mosque from the church.
That none ol'the bay! ai-'amb, the synagogue, and the church seems to have been the immediate ancestor of the concept of the mosque, but thai all shared the arrangement of the axial peristyle forecourt, serves to focus attention upon the architectural function of that common feature. First, although in each the courtyard was used for different purposes, in all il acted as a zone of transition between the public sircei. and the sacred space at the far end of ihe enclosure. Second, by emphasising the axial symmetry of the complex, the forecourt contributed to the architectural pointers that orientated the congregation towards the direction of prayer. I 'he courtyard of the mosque also performed precisely these functions. 1 lie mosque therefore seems lo belong to what is best regarded as a family of religious building types widespread throughout the Near Fast in the centuries preceding Islam.1"" That family and, indeed, tin- com crsion of its members to Islam - seems to be recognised in a passage in the (ha-'ah justifying war against the enemies of Tsla.nt: 'For had not Cod driven back one group of people by means of another, there would surely ha\e been torn down saiulrru' [retreats of Christian hermits?], biya' | Christian churches or Jewish synagogues? |. salawat [places oi prayer], and mauijid, in which the name of God is abundantly commemorated'.111" I he precise relationship of the mosque to the other members of this family cannot now be reconstructed, but the family resemblance is readily apparent.
The attribution of the corn ept of the mosque to a Fate Antique family of reli-
187. MtHjiiddusi 1877. pay). hn'da-twm bihu'di't ia-huddum! swmi'it wa-bm'm
iJ5i'l. The idea is 11111 a new one: see uu-saku1ttutt ttw-mfoqiithtmilliktiitijt hS mu HHIn
Lambert ujy & uj^l). kiUhlran. ■Hi). Lhir'u}!xxu.40: nw-fcr III rial'it tta'u h-iwm
i 02
l'IIĽ 'IIUUSK OF TUK ľROPTtV/ľ' AND TUK CON01.IT OK THE MOSOCĽ
giotis building types lias not, however, brought us any closer to identifying the
immediate origins of that concept. This line of inquiry peters out in the absence of archaeological evidence for the mosque in the Hijäz during the jä/iilí and Prophetic periods. That all my attempts to trace the evolution ol the mosque have ended in failure, persuades me to retrace my steps and pick up a thread left hanging towards the beginning of this article, when it was suggested that the crucial question is whether the mosque gradually evolved from pre-existing architectural forms, or whether i I was created by the new Islamic elite.
'I'llľ. prophet's mosouf.
Although the Prophet's House can not have been the origins of the mosque, it remains possible that the Prophcľs Mosque was the prototype of the mosque in Islam, If so. as we have already seen, it is likely to have been a much more elaborate structure than the empty hypaelhral courtyard described by Islamic tradition. In other words, we have yet to explore the possibility thai the Prophet created the concept of the mosque.
There is broad agreement in Islamic tradition as to the main points of the story of the foundation of the Prophet's Mosque. On entering Medina, the Prophet gave free rein to his camel until she stopped and 'knelt at the door of the mosque [baraka/ 'aula bäbi t-masjidiy. This was a mtrlmd owned by the orphans Sahl and Stihavl. Their guardian, As'ad ibn Zttrara, had used the enclosure as a masjidfor prayers and Friday congregation before the lujra. The Prophet summoned the two boys and ollered to buy the mirbad. They sought to make him a present of it. but lie refused and eventually bought it from Ibn Zttrara who compensated his wards. The Prophet then began to build his mosque.11"' This story echoes other foundation myths: the choice of site by the Prophet's she-camel recalls the role of animal guides in the foundation of many cities, including Antioeh, Rome, Thebes, and Trov. Put die story-" is also a caique upon the foundation of the Temple in Jerusalem The masjid oľQur'äri xvii./i.
According lo u Samuel i>4, Satan tempted David 1.0 conduct a census olYhe people oflsrael, against God's will. God punished him by sending a plague borne by an angel. In the middle of the destruction of Jerusalem, God staved his hand 'and the angel of the Lord was by the threshing-place of Araimah the Jebusite'. God then commanded David to creel an altar on the threshing-floor. David sought to buy it from Araimah who thereupon offered it as a gilt; David refused, and eventually succeeded in purchasing it. David built an altar on the site, and it was here that Solomon later constructed the Temple.11"
itjo. Samluidi if-liiii. pp.ioíi ŕi; Sanihu<Ii :cj'iti, 191. a Samuel ■> [. 1 Chronicles 21, a (lironieles 3.
vol.i, pp.'.ijy 33.
103
jFRKMY JOHNS
'['Uli 'lIOlľSL OF '111ľ I'ROI'HKt' AMD 'FHt C O NX. kpt OF THE MOSQUE
SI (
iSEElA' \
-A
it ' .'Clí
' ?:
V V:'
■ ^
6a
2()C
2 ]. Plan and rt:ooiiNlriieliou of the temple at al-Huqqa (after Railijem & von Wissman 1932.'' 25. s01.1I.i1 Arabian Icrnplcs from Yemen (after Schmidt 1982 a c): (a—b) (ila.n and isometric reconstruction of temple at YVaddum Dhu Masma'im; {c• plan oftomb-I.emple al al-Masa|id. 26. \abatacan temples trom I lie Havvräri (ailor Butler 1907]: (a; Si', temple of Ba'al Shamln; (IT Sahn 'e) Sur.
JEREMY JOHNS
This story was known to the collectors of traditions about the Prophet's Mosque, and a version of il appears in the accounts of the enlargement of the mosque under 'Umar ibn al-Khattab. Significantly, in this version, David eventually purchases the site of the' I emple not from Araunah but from two orphans, the counterparts of Sahl and Sidmyl in the traditions concerning the foundation of ihc Masjid al-j\~abty.]'''
There are thus four striking similarities between the foundation stories of the Temple :'as retold in Islamic tradition', and of the Prophet's Mosque:
i.  the site wa.s chosen by a divinely-inspired guide and not by
the human founder ;the angel of the Lord; the Prophet's shc-catneh;
•i. it was previously an agricultural enclosure \gorarf. mirbad\\
3.  it was owned by two orphans {i.e. according to the Islamic version of the story of David; Sahl and Suhayl);
\. from whom, but only after a struggle, the lounder succeeded in purchasing il.m:i
'I'he dimensions of the Prophet's Mosque may also be significant. Although
Samhtldi does report a divergent tradition, attributed to Kharija ibn Zaycl ibn
1'habit, that the mosque originally measured 70 by bo cubits, most authorities agree
that the enclosure measured 100 cubits square.1:11 These were the dimensions of the
conn of Solomon's Temple as given in Ezekiel xl:4/.
Another Islamic tradition concerning the construction of the Prophet's Mosque
refers not to the foundation of the Temple by David but, indirectly, to its dedication
by Solomon. The tradition describes the building of the zulla as follows:
"The columns were ol'palm trunks and the roof of palm fronds. And it was said to him 11 lie Prophet |, 'Why do you not roof it |i.e. properly]?'. And he said, 'It is a booth like the booth of Moses made of twigs and grass: tlic affair ]i.e. the end of the world] uill happen sooner than that ''amhtmka-'arishiMusu. Iduisfwbdtun iui tftummnmi.
tii-dinita a'j(d.(i mm dittibkaY.1!''
This tradition has been studied by Kister, who observes that it must ha.ve been widespread as early as the late second or early 3rd century. It contains a reference to Leviticus xxni, Moses declaring the feasts of the Lord, and precisely to the beast of the Tabernacles (SukkoO, the Jewish Harvest Festival. The ephemeral booths or
tabernacles constructed on this festival both celebrate the harvest and commemo-
iui>. Sam In i< 11 1 !!e!!. p. i;.]_'. lines 19 -jy Samhfidi [<)<)". vol.!. p.';4'J. line ly   1>.;:;4~>- 'ul<" n>-11);;. The story also hears su'ongsimiku'iiies to the siorx ol die purchase ol die cave of Maehpelah at 1 lebron \>\ Abraham in tjenesis ay).
1 ob
it)]. Samhüdl :'itl(if), p. 107, lines 0  10 and 190'f \ol.T. p.238, lines 12- 14). Sec also note 39 above. K);(. Ihn Sad 11)04, vol.i. ■_>. p.L>. 196, Kister 191)2.
TIFF 'UOUSF OF THF rROt'Hl.'l' AND THK CONCEFT Oh  1'HE M OS 01'J'.
rale the booths in which the Israelites dwelt during the Exodus.1'1'' In addition, because Solomon chose the least of the Tabernacles for the dedication of [.lie Temple, the booths also refer to the encaenia. Moreover, jusl as the Islamic tradition contains an eschatological reference, so docs the feast of Sukkoi have special eschatological significance.15,11
That the traditions surrounding I lie foundation oi the Prophet's Mosque contain a. reference to the foundation of the Temple is apparently confirmed by the fact that, according to most accounts, the mosque was founded on the anniversary of die Dedication of the Temple, for example, following the chronology given by Ibn Ishaq, the Prophet left Mecca on Friday 9 Rabl' 1 and, after three nights in the cave on Mount Thawr, reached Qiiba' on the following Monday 12 Rabl' 1. After resting there four nights, he set out for Medina on Friday 16 Rabl' 1 2!! Seplcmber (122, Julian;. He prayed at midday in the Wad! Rantina'. and it was in the late afternoon or evening that his exhausted camel knelt down at the spot where the Prophet ordered the mosque to be built. In this account, the days of the week do not correspond to ihe date of the month; thus, 16 Rabf 1 fell on Monday, not Friday. Assuming that Ibn Ishaq mistook the date but not the day. and intended the Prophet to arrive in .Medina on the Friday, this would have corresponded precisely to the anniversary of the Dedication of the Temple on 15 Tishri (25 September. Julian), which began at sunset on 13 Rabi' i.1'1"
The traditions surrounding the foundation and construction of the Prophet's Mosque thus contain the fragments of an artful literary confection closely modelled upon accounts of the foundation and dedication of the Temple in Jerusalem. 1 his elaborate version is clearly early, and must predate the versions given by Ibn Ishaq and Ibn Sa'd; they or their sources suppressed, or more probably failed 10 recognize (he reference 10 the foundation of the Temple embedded in the Islamic narrative. With its strong eschatological vein and its preoccupation with the Temple, the original version could well belong to the time of'Abd al-Malik, to the mid yos/bqos.
The fact that the only accounts of the construction of the Prophet's Mosque to have survived arc based upon an elaborate literary7 confection which is apparently Marwanid in date docs not necessarily mean that they are without historical content. There is, however, a. lar stronger reason for doubling the historicity ol the account of the Prophet's Mosque given by Islamic tradition. It is the objection raised by Caetani, that the mosque could not have sprung fully-formed from the Prophet's
K}".  T hat the ■arucUu'r had in he ephemeral niav explain whv the mosque was said to have hern buiil i)l mud-brick, a material, as King has pointed out : Kjiit), pp.7^  /();, that \cas foreign to Medina where stout; was ll ic i isual medium ol const ruction.
n.|t>. Sliaelei iii;r-;a& Shadcr ioq,jb; but see al-o Riibenstein iqq(>.
1 cjcj. [bn I hsliam 18",'}. p|).;j2u lor uncertainties over the precise dale ol the hifm. sec Akkouch iq'jj, pcjtt-P Rubin it)rj_>. pp.iqo |.
107
JEREMY JOHMS
head on die very lirsl. day that he setfoot in Medina. To believe that the Prophet had the foreknowledge, before the formation of a Muslim community', to build a structure large enough to accommodate a congregation ofsorne 3,000, and before the development of lite Islamic rite, to create an architectural form which, with only minor modifications, would fulfil all the liturgical needs of mature Islam, requires an acl of faith of which I am incapable. (Others will certainly disagree, and not just those who see the hand of God in historical processes, for one student of comparative religion has already pointed out to me that it is precisely the founders of new religions who create, ex trihilo, grandiose new structures - scriptural, liturgical, and architectural.)
It remains possible, of course, that the Prophet, didhuild the mosque that became the prototype for the mosque in Islam, but that he did so, not on the first day that he arrived in Medina, but very much later, after the Muslim community had grown large and after there had developed the essential components of the Islamic rice; for example, after the victory at Hunayn. This Prophet's Mosque could have survived just, long enough to have served as the prototype for the first mosciues to lie built in the conquests, before it was demolished and rebuilt by 'Umar. Islamic tradition, admittedly, contains nothing that might support this hypothesis; indeed, there is nothing in tradition to suggest that the Prophet's Mosque was so specially revered that it could have served as the model for the mosque in Islam.
(riven the rudimentary design and humble fabric of the Prophet's Mosque, it could not have been its architecture or its decoration that would have caused it to be adopted as the model for the mosque in Islam; rather, it would have been its association with the Prophet. It is thus highly peculiar that, despite the obsessive search for awail. tradition makes no attempt to identify the Prophet as the creator of the mosque. Moreover, there is little suggestion, at least not until it was rebuilt by al-Walid, that the fabric of the Prophet's Mosque was regarded with any special reverence, still less as lite prototype of the mosque iti Islam.200 What is more, tradition betrays no embarrassment that, the early caliphs treated the Prophet's Mosque with complete disregard. First 'Umar completely demolished, enlarged and rebuilt the Masjidal-J\'(rfny, replacing the palm- trunks with, columns of timber and mud-brick. Then 'Uthman destroyed 'Ulnar's mosque, and built his own in cut stone and plaster. It was only when al-YYalTd completely demolished and rebuilt the Prophet's Mosque for a third time that voices were raised in protest, although al. the destruction not of the mosque but of the hujurat. The sacralisation of the Masjid al-M'alny dms seems to belong to the Marwanid process of commemoration of the Prophet examined below by Barry' Flood.
ion. Peelersen (1989, p d. '. col. a) claim; that Taban 'expressly emphasised' that the plan ofKfifa I was ':m rs:ari reproduction nl ill.a nl the mosque in Medina'. 1'abarT does no such Uiíjisí, neither al
thepoinl riled hy Peilersen    [abaci i ří/tj, vola, p.2489, line.- 4 II". — nor, =o far a.s I can -ee Iron) A thorough reading ol the Knglish translation, elsewhere.
108
THE 'HOUSE OF Till': PROrilET" AND THE CONCEPT OF THE MOSOUE
There arc, in short, strong reasons to douhl dial die original Masjid. al-JVably in Medina could have served as the model for the mosque in Islam. While these by no means rule it out completely, they do incline mc to suspect that the plan of the Prophet's Mosque, as reported in the written sources, may have been reconstructed retrospectively in the 2nd or 3rd century by pious traditionalists who took as their model the mosques in which they themselves were accustomed to pray.
t'MAR IBN AL-KHATTAB AND THE CONCEPT OF THE MOSOUE
The concept of the mosque was not the product of gradual evolution after the hyra, but was created or adopted by the Muslim elite as the template for the religious architecture of Islam. While it remains possible that the Masjid cd-Nably was the model for this concept, there are reasons to doubt the accuracy of the description of the Prophet's Mosque given by Islamic tradition. That the concept of the mosque was not yet established during the lifetime of the Prophet is indicated by those passages in the Qur'an which suggest that the word masjid was not yet associated with a. particular architectural type. In litis context, it is surely not insignificant that the (ha'an makes no explicit reference to the Masjidal-Nabiy. If the Prophet was not himself the creator or the adopter of the concept of the mosque, surely it must have been a. caliph, but which ?
Archaeological evidence comes too late to do more than suggest that the concept was probably established before the accession of the Umayyads: it certainly dictated the plan of Wash in 84/703, and may have been followed at Kufa in 50/670 and at the Aqsa in the early 4OS/6O0S. The earliest written evidence for the mosque in Islam, independent of Islamic tradition, is the yth-ccntury Georgian pious tale, republished by 1'iusin m Bay tat Maqdisi, which places the construction of the first Aqsa mosque in the patriarchate of Sophronius, who died on 11 March 639 or (340.-"- If this Aqsa mosque - the Georgian text uses the Arabic word midzgitha - followed the concept of the mosque - and it must be stressed there is no evidence that it did then the choice of caliphs is narrowed to two: Abu Bakr and TJmar ibn al-Khattab.
Abu Bakr is said to have had a masjid outside his house at Mecca before the Jiijra^ and to have renewed the worm-eaten palm-trunk columns in the Prophet's Mosque, but he does not otherwise feature in Islamic tradition as a builder of mosques.1'"" In complete contrast, 'Umar is portrayed as an almost obsessive mosque-builder. 1 Ie is said to have enclosed the haram at Mecca, to have rebuilt the Prophet's Mosque in Medina, to have built the first Masjidal-Aqsd, to have commanded and supervised
201. l'lusin 1992. pp. 13t   [.. S;-imhfi(l! itjiifi, vol.!. p.34 j.
202. Thn HisliSm ifijH, P.24IJ- Siiinliudi li'b'o,
109
I: k ; \: y |: 111 \ >
from alar [lie Ibimdaiioii of mosques in die amsar. including Basra, Kula, and bustat, and in the conquests, including Alexandria, Damascus, Mada'in, and Mosul: in addition, popular tradition ascribes to him literally hundreds of mosques, probably more tba.n to any other figure excerpt die Prophet himself.
This image of TJmar the Mosque-builder is sustained by the following passage in Tabari's History. After describing the foundation of the hist mosque at Kuia in according to 'Ulnar's instruetions, Tabari continues, apparently in his own \ nice: wa-ka-dlullika kdnati 1-ni.asajidu makhala kmasjida l-haramaja-kami la yusltahbiluma bi-hi 1-rnasajida lawman li-hurmati-lu ; which Juynboll translates In the same manner, other mosques were laid out, except the Mosjid al-Hamm\ in those days they did not try to emulate that out of respect for its holiness'.In other words, 'Tabari believed that the caliph T'mar ibn al-Khaliab decreed that all congregational mosques should be laid out according to a common standard. Given our present ignorance of the archaeology of the mosque before the oos/yoos. this is our best evidence for the origin of the concept of the mosque.
SUMMARY
In conclusion, it may help to give a brief summary of this long article. Before the end of the tst century', the mosque had spread iboughout the conquests. Despite the many variations in plan, elevation, construction, and decoration, which tend to reflect tire indigenous traditions of the conquered cultures, all congregational mosques con-lonrtcd to a common standard that I have called the concept of the mosque. There is firm archaeological evidence that all congregational mosques were built according to this concept by the 8os//OOs, and there is less reliable evidence that this standard was already in use at Jerusalem in the 40s/660s and at Ktifa. in the 50s/670s.
Creswell believed that the Islamic mosque was the product of "trivial facts', of a casual process of evolution starting with the open courtyard of the Prophet's House in Medina. The critical analysis of C re swell's reconstruction of the Prophet's House demonstrates that the origin of the mosque in Islam cannot have been that simple domestic courtyard. Islamic tradition is clear that the Prophet built a mosque for the early community, and an analysis of the mosques of the Qur'an suggests that, during the lifetime of the Prophet, the mosque was already an established institution that could be distinguished from other types of religious buildings. The Qur'an uses the word masjid of so many different sanctuaries, however, that it seems unlikely that it was yet associated with a particular architectural type.
latian i\>/\.). \ol.l. p.^iSu; j'al)an-]u\nbull l()o<l, p.fii).
110
THK 'HOUSE 01  1HI: PROPHKT' AXD 1"H 1. OONOKPT OF Till'. MOSOD'.
j>".'lliiM"i«ii,'iiitiiriiin|ii'*1     v"in'-""inrinn'".......imnn 'hum     «piiii............'ithi1...........111 iniirnjtg     s,"iiiiWiir'i|iiiTnifT-"qi5 ^rrmnvmir%nfun-rt,-Ki^
| ^4utimi4uiiii .iJimiil.l'5     ji-liil....., ililluHii'i i-i,"*iuii-iiJ!     ?r!i..,j.iiuii- I i-Ii^juliiiiiliiliiilii Li^bJUiiiJ!     &' i,iiuuiiL.iuUiii.,.,iiiid     ikj.u Jimui.iium.jitj^ J
I    Marble column* Itthti Hfra
- arrow - cast-
- arrow - cast--
■■ illinhtlliHIir^      ijSfl'l Biim'IFTFnilTnL^      ^nT-rriiivnTi'rriiirY^      ijinifTiPn'i'h«^ %r
2.7. Kul'a 1 ;afuT (.'irsvvcll
7-7......,„„.....^ j
Paint traces of the prehistory of the mosque are dimly perceptible before Islam. The open enclosure aligned upon a ifibln seems to have been intrinsic lo die concept of the mosque, arid may, conceivably, be traced back lo an ancient tradition of hypaelhral enclosures associated with die veneration of the fertile earth. On firmer ground, the concept of the mosque clearly belongs to a family of Late Antique religious buildings, including the church, the synagogue, and the Arabian temple, all of which arranged a colonnaded courtyard and a covered sanctuary7 symmetrically a.boui a central axis. It is not yret possible to demonstrate when, where, or how die mosque became clearly distinguished from other members of this family. But it is clear that, some time after the hijra, probably after the death of the Prophet and belbre the accession of the Linavyads, the concept of the mosque was adopted as the common standard to which all congregational mosques were built.
The description of the Prophet's Mosque given by Islamic tradition is of a building which conforms to the concept of the mosque, and it remains possible, if unlikely, that the Alasjid al-j\abiy determined that common standard. But there are strong reasons to doubt that this was so, and to suspect, that the description of the original Alasjid al-Nahiy may have been composed reirospectivcly by pious Muslims who took as their model the mosques in which they themselves were accustomed Lo pray, long after the Prophet's Mosque had disappeared under the successive reconstructions bv'Umar, TJthman, and al-Walicl.
111
j erf. my johns
On the contrary, there is what appears lo be explicit testimony that 'Umar ibn al-Khattab decreed thai all congregational mosques should be constructed in the manner of the mosque built at Kul'a in the year 17/638 (fig.27). In the absence of archaeological evidence for the crucial formative period, this is the most likely moment at which the concept of the mosque was removed bom the family of Late Antique religious buildings in which it had hitherto developed, and was established as the common standard for the religious architecture of Islam.
112