Primitive Islam and Architecture in East Africa

Download as pdf or txt
Download as pdf or txt
You are on page 1of 15

Primitive Islam and Architecture in East Africa

Author(s): Mark Horton


Source: Muqarnas, Vol. 8, K. A. C. Creswell and His Legacy (1991), pp. 103-116
Published by: Brill
Stable URL: http://www.jstor.org/stable/1523158
Accessed: 03-04-2017 06:57 UTC

REFERENCES
Linked references are available on JSTOR for this article:
http://www.jstor.org/stable/1523158?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

Brill is collaborating with JSTOR to digitize, preserve and extend access to Muqarnas

This content downloaded from 69.167.4.1 on Mon, 03 Apr 2017 06:57:14 UTC
All use subject to http://about.jstor.org/terms
MARK HORTON

PRIMITIVE ISLAM AND ARCHITECTURE IN EAST AFRICA

THE ORIGINS OF ISLAM AND


K. A. C. Creswell never visited East Africa; indeed if he
had he would have seen very little that would have illu-THE SHANGA EVIDENCE
minated his understanding of early Islamic architec-
The
ture. In the 1930's archaeological investigations hadearly
not Islamic history of the East African
poorly by
begun, and most of the monuments were covered documented. Traditional chronicles an
thick and impenetrable bush. But within the last thirty
history provide a narrative for the arrival of Isl
years much has been done to clear the principal monu-
these were only written down in the nineteenth c
ments and make plans and elevations.' Archaeological and are generally held to be unreliable.4 The only
excavations provide a chronology, extendingwitness back into
descriptions by Arabic writers are those
the early Islamic period and detailed evidence for Mascudi
tim- and Ibn Battuta. Al-Mas'udi5 desc
ber as well as stone structures. The result is an impor-
coast largely pagan in 916, with one Muslim roya
ily residing in Kanbalu. Ibn Battuta6 describe
tant group of buildings that illuminate our understand-
ing of the use of Islamic architecture beyond the hiscenters
visit in 1331, a devout and wholly Islamic
of scholarship and craftsmanship of the central extending
Islamic from Mogadishu to Kilwa. Other d
lands. tions can be shown to be based on hearsay eviden
Creswell's comments that "Arabia, at the rise of Is-
often with little basis in fact. Idrisi, for exampl
lam does not appear to have possessed anything worthynot mention any Muslim communities, but dated
of the name of architecture"' is a view that could betterbic inscriptions survive from some fifty years be
apply to East Africa, south of Ethiopia. Indigenouswas writing.7
stone structures that are known from the region, such as The most recent archaeological work undertaken by
those from the Zimbabwe plateau and in the highveld of the British Institute in Eastern Africa has gone some
southern Africa, belong generally to the later Iron Age, way to establish the chronology of Islam in the region.
that is, after A.D. 1000; archaeological investigations An eight-year program of excavations at Shanga, in the
Lamu archipelago (fig. 1), has been followed by exten-
have as yet failed to identify an African tradition in sub-
stantial timber architecture. sive surveys and excavations in Zanzibar and Pemba
A common view of the genesis of Islamic architecture and a reassessment of previous work at Manda, Mafia,
and Kilwa. The broad conclusion from this work is that
is that, as it spread into centers of civilization and city
life, so it adopted and modified the existing architectu- ideas of Arab "colonization" of the African coast for
old
ral forms of these regions. One exception to this thesis the
is purpose of trade have been largely replaced by a
the coast of East Africa, where there was no such archi- broad process of conversion of indigenous coastal com-
tectural tradition. Therefore one might expect to find
munities to Islam through contact with the monsoon
traces of "primitive Islam" as defined by Creswell trading system of the western Indian Ocean.8
transmitted directly from the Arabian peninsula and Excavations at the site of Shanga in the Lamu archi-
unaffected by contact with complex societies. While pelago
the between 1980 and 1988 revealed a sequence in
phase of primitive Islam may have lasted only a very the center of the site with which we can begin to identify
short time in Arabia, it was able to continue in the more
these processes of Islamization. Shanga was a minor
remote areas of the Swahili coast.3 Thus the possibility
Swahili trading center, abandoned in the early fifteenth
of making significant archaeological discoveries iscentury,
so but which enjoyed an early importance, with
much the greater there. During excavations at Shangastratified occupation levels up to five meters in depth,
in the Lamu Archipelago, precisely such remains wereextending back to ca. 750, dated on the basis of import-
discovered, which are of interest to our understanding
ed pottery and radiocarbon dating.
of the development of Islamic architecture. The primary occupation lies directly upon white

This content downloaded from 69.167.4.1 on Mon, 03 Apr 2017 06:57:14 UTC
All use subject to http://about.jstor.org/terms
104 MARK HORTON

N1

ate L t suniga

rw a0Manda 0 0

K0ipungani kl
O LAMU ARCHIPELAGO

S 9th -:10
O O Later sites

Ungwana o0 5 0 10 20

Fig. 1. Location map of early sites in the Lam

beach sand. The trances on its west and eastassemblage


pottery sides. The enclosure was
pottery of the Tana
preciselytradition, but
cardinal north-south in orientation, and inthe
the
of imported Sasanian Islamic
geometric center was the and
well. Later reconstruction of
jars, as well as Chinese stonewares
this well damaged its early form, but a concreted sur-
from the outset face
Shanga was in bycont
of sand leading down to the well, formed the re-
soon trading systems.
peated sloppingThe house
of water onto stru
the surface, indicated
ber post-hole construction and
that it was either an open hollow circula
or shuttered with tim-
The layout of thisber in aearly
temporary manner.settlement
Five meters to the east of i
istic. Excavations the
revealed a burnt-out
well was a large tree, whose series of
tree stump
into the sand, which had
was excavated, and much been redug
iron slag was found around it.
Elsewhere within
and which apparently the enclosure there was littlea
contained trace of
fen
be shown to have habitation,
enclosed aandrectangu
only short gullies slots for temporary
mately 100 meters
structures,by 80with
often associated meters,
craft activity such as w

This content downloaded from 69.167.4.1 on Mon, 03 Apr 2017 06:57:14 UTC
All use subject to http://about.jstor.org/terms
PRIMITIVE ISLAM AND ARCHITECTURE IN EAST AFRICA 105

shell bead manufacture and the preparation


African kaya isof shell
a local manifestation of a wider settle-
ment plan found within both Bantu-speaking and
cones. Immediately outside the enclosure were found
Cushitic-speaking
arcs of post holes suggesting round huts, bread ovens,communities of the region, and may
and a much higher proportion of animal well and
have extended
fish bones.
up to the Red Sea coast of Ethiopia.
While no direct
The enclosure appears to represent demarcated link can be postulated, early Muslims
areas
that reached
of specialized activity, with defined entrances the African
whose po- coast may well have found
forms
sition is retained throughout the next six with whichyears
hundred they were already familiar in Arabia.
of the settlement. The importance of the well in the cen-
ter, the large tree nearby with its iron slag, but ARRIVAL
no fur- OF ISLAM
naces, suggest an area of ritual. The cardinal alignment
of the enclosure must also be significant,
At Shanga,
and astrono-
Islamic practice can be identified alm
mical alignments are not unknown among rectly above the
various horizon of white sand with its
early
pastoral groups in the region.9 debris of craft activity. Direct evidence comes
The ethnographic and ethnohistorical number of burials cutfor
parallels into the sand and laid out in the
these ritual enclosures are well recorded in East Africa. conventional way that continues to the present day -
The best known are the kaya of the Mijikenda who occu- that is, lying east-west on the side, with the head in-
py the coastal strip inland from the Swahili.'o The typ- clined northwards to Mecca. They are found in levels
ical kaya is a double rectangular enclosure with massive dated from around 800.
gates set within the forest. The gates are associated with More controversial are the timber structures that
specific clans, and buried below the gate is a fingo pot,were built directly over the burnt-out tree stump,
said to contain the magic carried from Shungwaya, thethe center of the site. They each lie directly below the
traditional homeland of the Mijikenda. The central rit-prayer hall of the later Friday mosque and thus, on th
ual area has no domestic occupation, only burials and
grounds of continuity, would appear to be mosque
the moro" where elders meet. Each miji or tribe has its
themselves. Each is laid out, using the dhira of 518 mm.
and they have a consistent qibla, albeit some 500 awa
own kaya and is modeled on the original archetypal kaya
of Shungwaya, the homeland, according to tradition. from the true direction of Mecca. What is interestin
Indeed the Mijikenda and the Swahili speak closely re-about this orientation is that it does not follow that of
lated languages, and the Mijikenda appear to represent the central enclosure, again suggesting that the timber
the non-Islamicized component of a common society.'2structures, or mosques, are not primary to it. Support
The Swahili also built similar enclosures, although that the deviant qibla line was perceived as the correct
within an Islamic context. The plan of Takwa, built line
in comes from the burials which line up with the tim-
the eighteenth century, clearly shows how the mosqueber mosques and not the central enclosure.
was set within a larger enclosure with entrance gates. InA total of seven timber mosques and one stone
the Comores, where many traditional Swahili practicesmosque (figs. 3-5) were found in the excavations below
have continued, a central communal area known as thethe prayer hall of the Friday mosque which itself was
constructed around 1000. The earliest structure lies di-
fumboni is found in a number of villages, again with clan
and moiety associations with gateways. There are even rectly above the burnt-out tree stump and was built on-
the remains of afumboni in the center of the Shangani to the horizon of trodden white sand and ironslag. The
quarter in Zanzibar Old Town. Drawings of nine- following is a summary of the sequence:
teenth-century Mogadishu show an open enclosure at
the point where the two moieties meet by the mosque of Mosque A. This structure, rectangular in plan, was
Fakhr al-Din (fig. 2). Fingo pots have also been found marked by traces of a thin wall trench supported by an
under doorways in a number of Swahili settlements."3external row of post holes. The entrance was somewhat
Of course this East African practice is reminiscent of
eroded, with traces of an emplacement for a timber step.
descriptions of pre-Islamic Mecca, with its walled sanc-
This mosque had two kinyokae floors, made of a green or-
tuary of Ka'ba, containing within it the sacred well ofganic mud collected from the mangroves, and two phas-
the Zemzem and settlement without. The suggestions ofes of a post hole wall, suggesting a rebuilding. It was
Askumite, and thus African, connections in the archi-aligned 3100 with internal dimensions of 1.64m. by
tecture of the Ka'ba that were made by Creswell might
2.59m. (almost precisely 9 cubits by 5 cubits using
extend to the form of the sanctuary itself.'4 The east
518 mm.).

This content downloaded from 69.167.4.1 on Mon, 03 Apr 2017 06:57:14 UTC
All use subject to http://about.jstor.org/terms
106 MARK HORTON

Fig. 2. The central enclosure in Mogadishu, located


of the moiety gates into the enclosure. To the left is

Mosque B. A clearly outside.


defined The plan
oblong is o
area a
beach pebbles directly
celledoverlay mosque
structure, with
apparently no associated post holes
larger northern an
room
dence for a superstructure. Itsof
the center measurem
the west
the
approximately 6.4 m. bybase
4.0 of
m. a stair. Th
marked by square ho
Mosque C. This was a well-defined
frame. rect
Unlike E, this
hole. Its
ture, with a wall trench internal
and dim
post-hole
side, a kinyokae (14 and
floor, cubits by 8 cubits
a central post h
ture was a little larger than mosques A
placed a little to the east, with
Mosque a very
E. This was a
3080. The internalcomplete
dimensions of 5.10
rebuilding
gives a size bywereofalmost
5.5 cubits 10exactly
using 518c
cut away by the tren
Mosque D. This structure
phases tofollowed alm
this buildin
same lines as mosque C, with a
kinyokae, qibla
two of fo
cuts 30
displaced to the east and by
ening slightly larger
the demoliti
defined kinyokae floor
ond and a wall
phase. The trench
walls w
holes actually set into the wall
40 mm. trench, ra
in diameter an

This content downloaded from 69.167.4.1 on Mon, 03 Apr 2017 06:57:14 UTC
All use subject to http://about.jstor.org/terms
PRIMITIVE ISLAM AND ARCHITECTURE IN EAST AFRICA 107

MOSQUE A (phase 1) MOSQUE A (phase 2) MOSQUE D MOSQUE E (phase 1)

io0 O '9 i ...


t"
0 stumpii
q ) I . . .I "
'00I I( i!_
* I
II o:7
'I - -
1iI . I i IiO
I LO __I
0 C)
-- -
o 0

CL9

MOSQUE B MOSQUE C MOSQUE E (phase 2) MOSQUE F

- - - - - - - - - - -L
I - - - - - - - -
I - - . . . . ..

9, I

bpebble
surfacete

0 5 t
4 0

~0

Fig. 3. Plan Fig.


of the4. Plan of
foundati
Shanga. The post holes wh
dence for the exact positio

duce an even were surface. 7.25 m


T
substantial 518 mm.).
post associa
ing that it had a flat, r
mensions Mosque G. A wooden
of this building was defined by a number
struct
actly 12 cubits
of large post pits, 0.80 m. in diameter
by and up to 1.08
m. cu
deep. As the posts had been robbed out, their exact po-
Mosque F.sitionThis
could not be discovered. The floor level of this
rectan
ture, on a building
new had also been qibla
destroyed. However, the basic
lin
a larger plan was of a rectangular building that
northern and was undivided
on axis from the
internally. Two smaller south
posts suggested a porch or en-
way. It was
trance on theconstruct
east side. The qibla was 329', and the ap-
tained posts,
proximate externalasdimensionswell
were 9.7 m. by 5.4 m., a
posts. A that is 18 by 10 post
large cubits, using 540 mm. in and not the
t
be for a 518 mm. of the previous structures.
mihrab, as th
southern wall required
plaster Mosque H. The
floor wasfirst stone mosque was built directly over
found
holes suggests
the post holes of mosque G. that t
It was constructed of neatly
building and
shaped porites coral replacem
(that is, undersea coral quarried by

This content downloaded from 69.167.4.1 on Mon, 03 Apr 2017 06:57:14 UTC
All use subject to http://about.jstor.org/terms
108 MARK HORTON

MOSQUE G MOSQUE H MosqueJ. Mosque J is the Friday Mosque; it still sur-


vives (fig. 6). It was built in stone, which was still in use
when the town ----
- ----_- -------- was abandoned
i in ca. 1425. The mosque
had three plaster floor levels resting upon a platform of
white sand, 1.5 m. deep. The mosque was entered by
steps in the center of the west side. The east and west
walls had an arcade of flat-headed doorways with small
rectangular windows above. On the east side, there was
a veranda of the same height as the internal floor. The
mihrab that was present in the final phase was inserted
through the north wall. The earlier arrangement was
destroyed, but it was apparently a much smaller niche,
possibly set within the thickness of the north wall. The
MOSQUE J
external walls had side pilasters, which were built up-
wards to support a thatched roof. Internally the prayer
hall was divided into a large northern and a small
o o
southern room, but this dividing wall was demolished
in the second phase and replaced by columns. Many re-
used stones were included in its walls. The qibla was
342 .

0 0
DATING

In the absence of inscriptional evidence, dating i


by radiocarbon or by association with local or im
pottery. The pottery comprises three basic a
blages:
Thefill ofMosqueJ. The floor was built over a platform of
white sand, which contained a large number ofsherds of
Fig. 5. Plan of the foundations of timber mosq
pottery, all of which were recovered, through total siev-
stone mosques H (all phases) and mosque J (p
ing. Sherds included Sasanian Islamic, white glaze, and
luster pottery, the later forms of the earliest local pot-
divers and shaped tery, but not a single sherd
while still of sgraffiato pottery. This
wet), bond
a white plaster face. In
group can be closely datedplan, the
to around 1000, providing a bui
into three parts - construction a square date for mosque J. prayer hall,
room, and a southern courtyard
Construction levels beyon
associated with Mosque H and G. These
nex was added during contain fewerthe
sherds, similarlife ofJ but
to those at Mosque the bu
ter of the north wall were the remains of a salient that with no luster pottery and with white glaze, and earlier
presumably contained the mihrab. On the west side forms of local pottery, dating to 850-900.
were steps, probably the base of a staircase minaret. Levels
No associated with Mosques A-F. There was very little
floors survived, but a layer of white sand sealing variation
the in the content of these layers, and pottery was
construction level suggested that there was a raised generally rare. The very earliest types of local pottery
floor, probably capped in plaster. The span of the were associated with Sasanian Islamic, eggshell wares,
prayer halls was too wide and must have been support- unglazed storage jars, and single sherds of Changsha
ed by a central column, but this too was robbed out. stoneware, Dusun and black stoneware. White glaze
Constructional evidence showed that there were three was entirely absent. Date range is 750-850.
phases of building. Internal dimensions of 5.10 m. and A sequence of charcoal samples was recovered from
wall thickness of 0.54 m. show that a cubit of 540 mm. the stratified mosques providing direct radiocarbon
was used, with a qibla of 3290. dates; the dating was undertaken by Dr. R. Switsur of
the Godwin Laboratory, Cambridge University. Six

This content downloaded from 69.167.4.1 on Mon, 03 Apr 2017 06:57:14 UTC
All use subject to http://about.jstor.org/terms
PRIMITIVE ISLAM AND ARCHITECTURE IN EAST AFRICA 109

Fig. 6. Excavations within the prayer hall of mosqueJ, showing the plan of mosque H, and the
three principal qibla lines are visible.

Table 1. Radiocarbon Dates from Shanga Mosques

Sample Archaeological Context Estimated Radiocarbon Calibrated date range, cal AD


Reference Date Age BP 68% 95%

SA1982 Fill of tree stump, cleared to build 750 1235 ? 35 710 to 745 685 to 880
mosque A 760 820
835 855
SA1647 Post hole in wall of mosque C 790 1180 ? 40 790 to 890 720 to 735
765 965
SA1512 Central post hole to mosque E, 825 1170 ? 45 785 to 895 720 to 735
associated with first of three phases 925 935 765 965
SA0659 Post hole of mosque F 850 1100 ? 50 890 to 980 815 to 840
850 1020
SA0616 Spread associated with early use of 900 1060 ? 45 900 to 915 895 to 1025
mosque H 950 1020
SA0590 Spread below floor of mosque J, above 1000 985 ? 35 998 to 1042 980 to 1062
robbing of mosque H 1095 1115 1070 1125
1105 1150 1135 1160

This content downloaded from 69.167.4.1 on Mon, 03 Apr 2017 06:57:14 UTC
All use subject to http://about.jstor.org/terms
110 MARK HORTON

samples were chosen to represent


mosques excavated. Samples were
ed features such as OO 0
post holes or
floor levels. Table 1 shows the resu

oo o 4 5
i i ,
with our estimates of the date arri
ic analysis and general evidence of
diocarbon ages obtained from thes
correctly to their stratigraphic or
diocarbon dates provide an indi

-LI,
1%
mosque is unlikely to date from
even at 95 percent probability le
sible date for the tree stump w
mosque C 720; at 68 percent the
785. A date for mosque A would, o

* i~
within the range 750-850. The a
ramic evidence would point to th
range. A possible dating scheme w
the second half of the eighth cent
buildings of the flimsy structures
and generationalFig. rebuilding of
7. Plan of the timber hall at Shanga the
excavated in 1
foundations
E, F and G to ca. of the porites
900; building shown
the stoneabove. The p
of the timber posts (shown as solid) could be defined wit
900-1000, replaced by mosqueJ
post pits.

SECULAR BUILDINGS
west by a short flight of steps. On one side of these steps
Contemporary with and very near to thesewas early
a basintimber
presumably for washing feet before enter-
mosques, on the west side of the well, another timber
ing. Inside the plaster floors were raised on a platform of
white
structure, which could be described as a hall, wassand, and as the building was heavily robbed,
found
and excavated (fig. 7). The post holes were there
of twowas notypes,
direct evidence for the positioning of door-
ways.
an outer row of small posts and an inner grid ofThere were three rows of rooms, the first un-
massive
post pits up to 2 m. in depth and up to 1.6divided,
m. deep.possibly an internal courtyard, the second di-
The
vided into three,
ghosts of the posts themselves suggested timbers up to and the third into two. At the east end
40 cm. in diameter. In plan, the post holes formed
was another, much a shorter flight of steps to a rear space.
grid, five by four of equally spaced posts, There
withwas theno evidence
ex- of domestic occupation on the
ception of two central posts, suggesting the structure
floors, or indeed in the external courtyards. Some me-
had a symmetrical internal courtyard and a ters to the west
portico onwas a second, much smaller, porites
each side that ran north-south (fig. 8). Thebuilding,
alignment of
of two rooms, again with raised floor levels.
this structure was that of the central enclosure and not These may have been stores.
of the mosques to the east. Its scale suggests a ceremo-
nial function, and it was replaced, significantly, by a ARCHITECTURAL ORIGINS
stone building on the same alignment.
This new building was constructed of porites coral
Finding parallels for these buildings is a task be
bonded in mud with plaster facing. Its positioning,
difficulty, not least because such ephemeral
slightly to one side of the timber hall, shows that the structures have never before been excavated either in
builders were aware of the location of the post pits.
East Africa or in Arabia. However some suggestions
Otherwise the two structures were almost identical in can be made, although I realize they are based upon ve-
size and shape, and the stone building must be viewed
ry slender evidence.
as a replacement in stone of the timber hall below. Dat-
ing evidence suggests that this took place around 900.The timber mosques. The reconstruction of the elevation
The porites building (fig. 9) was entered from thefrom the pattern of post holes suggests that the build-

This content downloaded from 69.167.4.1 on Mon, 03 Apr 2017 06:57:14 UTC
All use subject to http://about.jstor.org/terms
PRIMITIVE ISLAM AND ARCHITECTURE IN EAST AFRICA 111

Fig. 8. The timber hall with the principal post pits excavated out.

ings were flat-roofed rather than gabled.


were It is
simply note-
walls of woven wattle att
worthy that the side post holes arespaced
exactly the same
timber size
uprights, which were for
as the corner posts. Mosques D and E, for
support a example,
timber joist that held th
have no structural supporting timbersshaped
on theroof of wattle
corner or and daub. By
on the central axis, which initially led me
posts, toareas
large conclude
could be roofed in th
that they were not roofed at all. Themum evidence
span forforeach
a cen-
wattle being about
tral post, however, indicates that some form of roof
constructions was
of these tembe houses in
placed over the structure, but it is seum in Dar esto
very difficult Salaam,
see provide a pos
how a gable roof was supported with reminiscent
no major postsof theatShanga structures.
the north and south ends to support the timber.
However, I think,
such flat-roofed structures
therefore, that the evidence pointssuitable for the
to flat-roofed East African coast,
struc-
tures of very simple construction. much heavier than in the Masai pla
There are parallels for the flat-roofed
stonehouses
housesinare East
widespread, but the
Africa, known as "tembes," found and
in run-off channels,
the Masai plains which of cour
with wattle-and-daub
of Tanzania in the area of Lake Manyara.'5 Sometimes roofs. But there
they were dug into the ground, that even domestic
but more commonly mud houses of the

This content downloaded from 69.167.4.1 on Mon, 03 Apr 2017 06:57:14 UTC
All use subject to http://about.jstor.org/terms
112 MARK HORTON

Fig. 9. The porites building at Shanga durin


robbed wall foundations. The steps, with a

roofed during
the
timeeighth and nin
the clapping o
Chittick ligious
pointed
out that ceremony'
the patter
on daub fragments from
ological Manda f
evidence w
their position onearly mosques.
the roof, rather
Similar daub has The
alsosouthern
been found
room
most likely that tion. Some of the
the flat-roofed bu
troduced from an arid area,
others either
do not, so t
i
ner in which
or the Arabian peninsula. its th
Clearly do
ly, given the Islamic context
was there for th
partly to
A feature of particular
ea. It has nointere
washin
mosques is the use of rounded
at the northwest peb
c
mosque B, a practice followed
entrance. But it in
sti
Amr in Fustat in 641.
used forIt ablutions,
was also
665; about that Themosque
absence Creswe
of a m
marker,
"when people prayed, in hands
their the cen b
dust, which they used was
mihrab to remove
not wi
caused Ziyad to say,
tury. 'IIn
am afraid
the Shang t

This content downloaded from 69.167.4.1 on Mon, 03 Apr 2017 06:57:14 UTC
All use subject to http://about.jstor.org/terms
PRIMITIVE ISLAM AND ARCHITECTURE IN EAST AFRICA 113

with mosque F, where there appears to be a


central large axially
column does not remain, but w
located post that was not answered sentby another
because atroof
the the span is too great
south end. This large post may have been partly
supported. Anotherhol-
feature at Shanga
lowed out to provide a shallow recessed
of stepsniche.
on the east side, probably for a
(fig. 10), again found also in the Siraf
amples.'9 architecture, the
The stone mosques. Turning to the stone
evidence is much clearer. The earliest stone
Although mosque
these parallels at
between Shanga and the Si-
Shanga fits closely into a group ofrafmosques
small aremosques
quite precise, known
they should not necessar-
from Samarra and Siraf,'8 whichilyare imply also dated
the direct to
importation of the
Sirafi architecture to
tenth century. These structuresthedo
East not
Africancontain the
coast. The earlier timber mosques at
courtyards that are characteristic of the
Shanga have congregational
no courtyard, and the separate southern
mosques of the period, but they have a tripartite
rooms suggest divi-from a wooden to a
some local continuity
sion along the qibla axis of the prayer hall, back
coral architecture. The useroom,
of porites coral, a Red Sea
technique,
and courtyard, in exactly the same way points
astoatYemeni connections and stone
Shanga.
They have square prayer halls divided
structures by a transverse
with similar plan are known from the Ye-
arcade. As the floor level does not survive
men,20 at recent
although more Shanga in date. Very little is
(and was probably raised on a platform of
known about the sand),
small mosques ofthe
the Abbasid and late

Fig. 10. The recessed minbar at Mbui Maji, Tanzania. It has been claimed that this arrangement ofmihrab and minbar is a survival of early Is-
lamic practice in East Africa, but it probably dates to the 18th or early 19th century.

This content downloaded from 69.167.4.1 on Mon, 03 Apr 2017 06:57:14 UTC
All use subject to http://about.jstor.org/terms
114 MARK HORTON

Umayyad period;against
they this interpretation.
may Burials stratified
have below this
bee
dard plan throughout the
structure give a clear indicationIslamic
of the perceived qibla wo
by the early- toline at this date of around 3100, whereas this structure
mid-eighth-centu
Muqatil in Iraq.2'
faces 2670, virtually due west. This is too great a di-
The final mosque at
vergence to accept, even Shanga,
in remote East Africa. da
seems to be a localThe traditional
development derivation of courtyard mosques
deri
cessors. The prayer
from the plan
hall of domestic
ishouses,
rectangul
in particular that of
Muhammad, is
of 2:3, rather than square; it isparticularly significant.27 Reconstruc-
divid
columns. Again tionthere
of the house of theisPropheta souther
in Medina gives a rect-
floor was raisedangular
1.5 building,
m. with anabove
open courtyard having
the a
that entry was southernsteps
via portico of three bays,
from similar in somea
wayswas
to
the west side. Both the east and west walls have an ar- the timber hall at Shanga. The portico was supported
on palm trunks, which may well have been the material
cade of three flat-topped doors. No trace of the earlier
mihrab arrangement survives, although the recess must used at Shanga, as the post diameters are too wide for
have been set into the wall above floor level. A partic-mangrove poles. Such courtyard houses may have been
widespread in the Arabian Peninsula and copied by
ular feature of the mosque was that it has pilaster strips
Muhammad for his own house. A quasi-domestic or
on the outside walls, which end in short pillars that sup-
ported a thatched roof, which was presumably gabled. ceremonial function for the Shanga building is possible,
Some of the features of this mosque are reminiscent perhaps
of associated with the ruling group who adopted
the earlier timber and stone structures at Shanga; this building form from Arabia.
others, such as the pilasters, are new features.
The stone mosque at Shanga is of particular impor- The porites building. The case for a ceremonial function
for the timber hall is strengthened when we consider the
tance as it provides the prototype for a group of early
mosques in East Africa, which are remarkably uniform.stone building which replaced it. It survives on a gran-
This group includes Mbui,22 Kaole west,23 Kizimkazi,24
der scale and represents the adoption of stone building.
Kisimani Mafia I, II, and III,25 and Sanje ya Kati.26 The
All technique that uses porites coral appears to have
these structures date to the eleventh or early twelfthbeen introduced from elsewhere, and the Red Sea
would be the most likely candidate, given the distribu-
century. They often have side pilasters with prayer halls
in a proportion of 3:4. On the east and west walls tion
are of similar corals in this area and the continued use
paired doors, with a centrally placed southern door; the
of porites coral in building to this day.28
roof is supported on four, or sometimes two, circular co-As the building only survives to floor level, its eleva-
lumns, which often were in timber. Only Kisimani Ma- tion is difficult to reconstruct. The thickness of the
fia III has a southern room. Otherwise the prayer hallwalls,
is especially the internal walls, and the quantity of
freestanding with a raised floor level, supported on a
rubble suggested that it may not have been a single-
storied building; it may have risen to several floors -
platform of white sand. The walls are constructed of po-
rites coral. the entry through a flight of steps suggests a grand front
elevation, which would not have fit in easily with a low,
The timber hall. The post holes of the timber hall support- squat building. I suggest that this was a tower building
ed major structural timbers, and are too widely spaced several stories in height. Examples of such a building
to have supported a tembe-style roof. The outer wall be-style can be cited on both sides of the Red Sea, from
yond these post holes does not survive, but may havewhere the use of porites coral seems also to have been
been of daub construction. A suggested reconstruction derived.29 The tower palaces of the kings of Askum,
is of a square building, with an internal courtyard and a which continued to be built at least until the seventh
portico on the west side. The arcade would have beencentury, provide a particularly close parallel.3o In Asku-
entirely of timber, of which no details have survived. mite architecture stress was given to monumental stair-
Similarities between this building and early court-way entrances, which finds an echo in the Shanga build-
yard mosques can be suggested, albeit on a very muching. At nearby Manda a similar porites building was
smaller scale. Both have the portico on one side with a found with an entrance stair of four steps."' The Asku-
central courtyard. Could this structure be a small early mite palaces are not that much larger than the East
courtyard mosque? Evidence for its orientation isAfrica examples - that at Ta~akha Mariam was only

This content downloaded from 69.167.4.1 on Mon, 03 Apr 2017 06:57:14 UTC
All use subject to http://about.jstor.org/terms
PRIMITIVE ISLAM AND ARCHITECTURE IN EAST AFRICA 115

Monumental

,onument--. stone building


stone buildings 1 --'
L ggver.av
well gr ea "-

street

house Mosque
N

0 10 20 30

metres

Fig. 11. The complex of monum

21 m. by 25 the
m. early Um
compared
ga and 16.2 Creswell
m. by and
12.2 m
The openinglied
up upon
of doc
tradin
Africa and forms
the Red of Isla
Sea in
context for come
the for ar
transmiss
southwards.more
The criticall
porites
whole group of similar st
enclosure; British
their Institu
wooden f
ones. At the London,
same Engla
date th
also replaced in stone. T
covering up to 8,000 m2 s
complex of monument
NOTES

mosque in the center (fig


1. P. S. Garlake, The Early Islamic Architecture ofthe East African
mite palace complexes c
British Institute in Eastern Africa, Memoir no. 2 (Na
enth-century Dongur,32
1966), provides the most comprehensive corpus of plans a w
ings and courtyards
vations. filled
These 2. K. A. C. Creswell, A Short
examples showAccount of Early Muslim Architecture,
that ed.
James Allan (London: Scolar Press, 1989), p. 3.
ea in which to understand
3. Ibn Battuta first uses the term Swahili Coast, and travel writers
tecture, both because the
until the nineteenth century refer to the African Muslim inhabit-
tion, as the ants earliest
as the Swahili. The Swahili Coast extends fromcen
Mogadishu
abandoned, and because
in the north to Mozambique, in
and includes offshore islands such
tigraphy is wellas the Comores, Zanzibar, Pemba, Mafia, and northern Mada-
preserved
gascar.
the transition from wood
4. Only the fragments of the Kilwa chronicle that survive in a Por-
ephemeral of buildings. W
tuguese prtcis written down by De Barros (Da Asia, Dec. 1, book
tend back into the
VII) predate the nineteenth century.sevent
On the unreliability of
may push back these traditions,the origin
particularly the claimed Umayyad expeditions

This content downloaded from 69.167.4.1 on Mon, 03 Apr 2017 06:57:14 UTC
All use subject to http://about.jstor.org/terms
116 MARK HORTON

to the Swahili, seegregational


M. Tolmacheva,
Mosque "Th
and Other Mosques of the Ninth to Twelfth Centuries,
mascus in Syria," The British Institute of Persian
International Journal Studies, Siraf III
of (London,
Af
12.2 (1979): 259-69. n.d.).
5. Muruj al-Dhahab, 19. J.1:
Schacht, 112;
"Further Notes on 3:the Staircase
31. Minaret," Ars Orien-
talis 4 (1961): 137-41, argues that the2:
6. Les Voyages d'Ibn Batoutah staircase179-96.
minaret is a relic G.
The East African Coast, of early Islamic practice
Select on the East African Coast. Its occur-
Documents
1975), contains translations of
rence at Shanga places its use there as the main
early as the tenth century, sour
Africa early during Islamic
but staircase minarets the
continued to be used in the period.
Gulf for many
centuries, as D. Whitehouse shows,
7. Idrisi states, in the seventh "Staircase Minarets on the
section of the
Persian Gulf," Iran 10 (1972): 155-58.
Rujar, that Barawa (Somalia) was Schacht's argument
the about last of
recessed minbars' being another Umayyad
fidels, who have no religious creed, relic survival in East
but tak
anoint them with fish oil and bow down before them. But in the Africa is less convincing, with no known examples predating the
Friday mosque at Barawa, there is an inscription dated A.H. 498 late eighteenth century; see J. Schacht, "An Unknown Type of
(1104-5). Minbar and Its Historical Significance," Ars Orientalis 5,2
8. M. C. Horton, "Early Muslim Trading Settlements on the East (1957): 149-73. Garlake, Early Islamic Architecture, pp. 74-75,
African Coast," Antiquaries Journal 67, 2 (1987): 290-323; idem, G. R. Smith, "A Recessed Mimbar in the Mosque at Simam-
"Asiatic Colonization of the East African Coast: The Manda baya," Azania 8 (1973): 154-56; and H. N. Chittick, "The
Mosque at Mbuamaji and the Nabahani," Azania 4 (1969): 59-
Evidence"Journal of the Royal Asiatic Society, pt. 2 (1986): 201-13.
9. R. Soper, "Archaeo-Astronomical Cushites: Some Comments," 60, show that Schacht's dates are often too early.
Azania 17 (1982): 145-62, points out that some of the20. R. B. Lewcock, "Architectural Connections between Africa and
detailed
claims for astronomical observation by Cushitic groups cannot Parts of the Indian Ocean Littoral," Art and Archaeology Research
be sustained. Papers 9 (1976): 15.
10. A. Werner, "The Bantu Coast Tribes of the East Africa Protec-
21. Creswell, Short Account pp. 221-22.
22. Unpublished fieldnotes.
torate," Journal of the Royal Anthropological Association 45 (1915):
326-54; T. Spear, The Kaya Complex (Nairobi: East African Liter-23. Garlake, Early Islamic Architecture, fig. 3.
ature Bureau, 1978); H. Muturo, "An Archaeological Study24. ofIbid, fig. 50; H. N. Chittick, "Preliminary Report on the Excava-
the Mijikenda Kaya Settlements on Hinterland Kenya Coast," tions at Kizimkazi Dembani, Zanzibar," Tanganyika Antiquities
Ph. D. diss., University of California, Los Angeles, 1987, espe- Report for 1960 (Dar es Salaam, 1962), pp. 17-19.
cially plan on p. 170. 25. H. N. Chittick, Kisimani Mafia: Excavations at an Islamic Settlement
11. Lit. "Cattle byre," a possible clue to the pastoralist origins of the on the East African Coast, Antiquities Division Occasional Paper
Mijikenda. no. 1 (Dar es Salaam, 1961), and "Report on the Excavations at
12. D. Nurse and T. Spear, The Swahili (Philadelphia: University of Kisimani Mafia and Kua." Tanzania Antiquities Report for 1964
Pennsylvania Press, 1985). (Dar es Salaam, 1966), pp. 15-16.
13. T. Wilson, "Spatial Analysis and Settlement on the East African 26. H. N. Chittick, "Kilwa, A Preliminary Report," Azania 1 (1966):
Coast," Paideuma 28 (1982), fig. 2; M. C. Horton, "The Early 3.
Settlement of the Northern Swahili Coast," Ph. D. diss., Cam- 27. Creswell, Short Account, p. 4.
bridge University, 1984; C. Revoil, "Voyage chez les Benadirs, 28. J. P. Greenlaw, The Coral Buildings of Suakin (London: Oriel
les Qomalis et les Bayouns, en 1883," Le Tour de Monde 56 (1888): Press, 1976); S. M. Head, "Corals and Coral Reefs of the Red
385-414. Sea," in A. J. Edwards and S. M. Head, eds. Key Environments,
14. Creswell, Short Account, p. 3. Red Sea (Oxford: Pergamon Press, 1987), pp. 130-33.
15. S. Denyer, African Traditional Architecture (London: Heinemann,
29. F. Varanda, Art of Building in Yemen (London: Art and Archaeol-
1978), pp. 62-63. ogy Research Papers, 1981), pp. 81-99.
16. N. Chittick, Manda, Excavations at an Island Port on the Kenya Coast,
30. Y. M. Kobishchanov, Axum (University Park, Pa.: Pennsylvania
British Institute in Eastern Africa, Memoir no. 9 (Nairobi, State University Press, 1979), p. 141.
1984), p. 18. 31. Chittick, Manda, pp. 41-44.
17. Creswell, Short Account, p. 9. 32. F. Anfray, "L'archeologie d'Axoum en 1972," Paideuma 8 (1972):
18. Creswell, ibid, p. 414; D. Whitehouse, "The Smaller Mosques at 60-76.
Siraf: A. Footnote," Iran 12 (1984): 166-68; and idem, The Con-

This content downloaded from 69.167.4.1 on Mon, 03 Apr 2017 06:57:14 UTC
All use subject to http://about.jstor.org/terms

You might also like