See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/273293351

Inside and Outside the Dry Stone Walls: Revisiting the Material Culture of
Great Zimbabwe

Article in Antiquity · December 2015

DOI: 10.1017/S0003598X00097726

CITATIONS READS

83 11,596

2 authors:

Shadreck Chirikure Innocent Pikirayi

University of Oxford University of Pretoria
148 PUBLICATIONS 2,357 CITATIONS 83 PUBLICATIONS 1,396 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Innocent Pikirayi on 23 March 2015.

The user has requested enhancement of the downloaded file.

Inside and outside the dry stone walls:
revisiting the material culture of Great
Zimbabwe
Shadreck Chirikure1 & Innocent Pikirayi2

‘Any study of Great Zimbabwe has to rely a great deal on re-examining and re-assessing the work
of early investigators, the men who removed all the most important finds from the ruins and
stripped them of so much of their deposits’ (Garlake 1973: 14). The authors have here done us
a great service in reviewing the surviving archaeological evidence from this world famous site.
They challenge the structuralist interpretation – in which different parts of the site were allocated
to kings, priests, wives or to circumcision rituals – and use the architectural, stratigraphic and
artefactual evidence accumulated over the years to present a new sequence. The early enclosures
on the hill, the Great Enclosure and the valley enclosures now appear as the work of successive
rulers, each founding a new residence and power centre in accord with Shona practice.
Keywords: East Africa, Great Zimbabwe, second millennium AD, structural, symbolic
archaeology

Introduction
Great Zimbabwe (Figure 1) is one of more than 200 sites in southern Africa (Garlake 1970;
Beach 1998) which display the architectural tradition of those monumental but mortarless
walls that have continued to attract archaeologists and the public alike (Ndoro 2001; Fontein
2006). With that attraction have arisen speculations and debates about the identity of the
site’s builders and the function of the walled enclosures (see Hall, R.N. 1905; Garlake 1982;
Hall, M. 1995; Huffman 1996). Since the late nineteenth century several research issues
have dominated the archaeology of Great Zimbabwe, including its origins and dating (Bent
1892; Hall 1905; 1910; MacIver 1906; Caton-Thompson 1931; Summers et al. 1961;
Collett et al 1992; Chipunza 1997) and its purpose and significance (Garlake 1973; 1982;
Huffman 1986; 1996; Beach 1998; Thorp 1998).
Architectural studies by Schofield (1926), Summers and Whitty (1961) and Whitty
(1961) proposed a relative sequence in which the walls were constructed over time. Other
studies have deduced the economic basis of the state based at Great Zimbabwe and
the broader landscape setting and settlement hierarchy around stone built monuments of the
Zimbabwe type in general (Sinclair 1987; Pwiti 1996). And yet others, particularly the more
1
Department of Archaeology, University of Cape Town, Rondebosch, Cape Town 7701, South Africa (Email:
shadreck.chirikure@uct.ac.za)
2
Department of Anthropology and Archaeology, Humanities Building, University of Pretoria, Pretoria 0002, South
Africa (Email: innocent.pirikayi@up.ac.za)
Received: 30 October 2007; Revised: 29 May 2008; Accepted: 22 June 2008
ANTIQUITY 82 (2008): 976–993
976
 Shadreck Chirikure & Innocent Pikirayi

recent ones, have focused on the politics

of the past and the conservation and
management of the site (Ndoro 2001;
Fontein 2006; Chirikure & Pwiti 2008).

Research
More recently the site has been interpreted
from symbolic, structuralist and cognitive
standpoints (Huffman 1981; 1982; 1985;
1986; 1996; 2007).
Research at Great Zimbabwe has, right
from the onset, been highly politicised.
Between 1870 and the 1930s, the site
attracted considerable controversy, much of
it based on antiquarian speculative beliefs
and European colonial attitudes (Bent
1892; Hall & Neal 1902; Douslin 1922;
Caton-Thompson 1931; Burke 1969; Hall
1987; 1995; Pikirayi 2001). The Rhodesian
settler community did not accept an
Figure 1. Map of southern Africa showing the location of African authorship of the monuments.
Great Zimbabwe.
Even the highly scientific investigations by
Robinson et al. (1961) were overshadowed
by partisan claims (Wainwright 1949; Jeffreys 1954; Dart 1955). The radical white
Rhodesian Front exiled professional archaeologists and hired non-archaeologists to re-write
and popularise the antiquarian version on the origins and identity of the builders of the site
(Bruwer 1965; Gayre 1972; Garlake 1982; Pikirayi 2001). Thus between the 1960s and
1980, when Zimbabwe attained independence, research on Great Zimbabwe remained far
from impartial.
Fairly recent work from Collett et al (1992) and Chipunza (1994) has continued the study
of the architectural development, but post-1980 research has suffered from a moratorium on
excavations placed by the cultural management authorities, in favour of the conservation of
dry stone walls and earthen structures. Whilst conservation is a laudable development albeit
with its own problems (Ndoro 1994; Fontein 2006), it is worrying that no new generation
of scholars seems to be taking an active interest in the archaeology of the site (Chirikure
2007a). This creates the false impression that we have exhausted all possible avenues of
investigating the monument. Yet there are huge gaps in our knowledge. Since 1980, there
has never been an integrated archaeological research programme on Great Zimbabwe, only
isolated and often fragmented approaches (see for example Chipunza (1994) on stone
architecture, Matenga (1998) on soapstone birds and Chirikure (2007b) on metalwork).
This fragmented approach somewhat frustrates attempts to develop a coherent history of
the different activities carried out at the site as revealed through artefact studies.
Using archival and published data, we seek here to review the archaeology of Great
Zimbabwe, integrating assemblages with stratigraphic and architectural sequences. Making
sense of some of the existing data is not easy if one considers the history of sustained plunder
which has destroyed important deposits without adequate recording (Garlake 1973: 14).

977
 Inside and outside the dry stone walls: revisiting the material culture of Great Zimbabwe

Figure 2. Site plan of Great Zimbabwe (modified from an original plan by National Museums and Monuments of
Zimbabwe).

Nevertheless we are convinced that old data has the potential to bring new insights. In
particular, we feel we are in a position to challenge the prevalent current structuralist model
and replace it with another that has time depth.

Great Zimbabwe: the sequence (see Figure 3 and Tables 1 and 2)

The site of Great Zimbabwe consists of stone-walled enclosures on the hill (the Hill
Complex) and in the adjacent valley (the Valley Enclosures), together with other unwalled
areas (Figure 2). Two perimeter walls demarcate the inner and outer limits of the town
(Sinclair 1987: 106). The Upper Valley includes the Great Enclosure, famous for its exquisite

978
 Shadreck Chirikure & Innocent Pikirayi

Research

Figure 3. Stratigraphy of a section of the Hill Complex together with pottery classes and radiocarbon dates. The chronology
of the site is based on the excavation by K. Robinson (1961a).

979
 Inside and outside the dry stone walls: revisiting the material culture of Great Zimbabwe

Table 1. Great Zimbabwe: summary of the main periods of occupation (modified from Summers et al.
1961 and Huffman & Vogel 1991).
Period Focus Pottery Architecture Date

I and Ia Hill Complex Class 1 – no stone walling 5th-8th centuries

Gokomere/Ziwa/
Zhizo pottery
------------ Abandoned ------------ 9th-12th centuries
II Hill Complex Class 2 – no stone walling; mid 12th-early
Gumanye dhaka house 13th centuries
pottery floors
III Hill Complex, Western Class 3 pottery, P stone walling; early-late 13th
and Eastern Class 3 substantial centuries
Enclosures influenced by dhaka houses
Class 4a pottery
IVa Hill Complex; Great Class 3 influenced P and PQ late 13th-early
Enclosure; Upper by Class 4a walling 14th centuries
Valley Enclosures pottery; Class 4
pottery
IVb Great Enclosure; Lower Class 4b graphite PQ, Q and R early 14th-mid
Valley Enclosures burnished ware walling 15th centuries
IVc Lower Valley Enclosures Class 4c graphite Q and R walling mid 15th-mid
burnished ware 16th centuries
------------ Abandoned ------------ mid 16th-19th
centuries
V Lower Valley Enclosures Class 5 – Karanga R walling 19th century
pottery

workmanship in stone. This has attracted looters and treasure hunters who, for centuries,
have stripped it of metres of archaeological deposit. In 1902 Hall and Neal ‘found huge
piles of soil debris deposited within the ruins by a succession of explorers, both authorized and
unauthorized’.
Over the years, the chronology of Great Zimbabwe has been built up within a combined
framework of stratigraphy, pottery sequences, radiocarbon dates and architectural history
(Summers et al. 1961; Figure 3). The sequence is divided into five periods (I-V) ranging from
the sixth to the nineteenth century AD. This periodisation is strengthened by the existence
of notable differences in material culture such as glass beads and local pottery belonging to
the individual periods (Robinson 1961a, b & c). Accordingly, the local pottery has been
categorised into five classes (Classes 1 to 5), corresponding with the stratigraphy (Figure 3).
The material culture of Period I, assigned to early farming communities, dates to between
AD 500 and AD 800 and comprises characteristic pottery known as Gokomere-Ziwa and
Zhizo (Class 1) (Robinson 1961b). After a lengthy hiatus, the site was occupied from the
early second millennium AD (Period II) by an ancestral Karanga people who made pottery
known as Gumanye (Class 2) (Huffman 2007) and lived on the hill.
Periods III and IV which follow constitute the main era of the stone walls and of the
flourishing of Great Zimbabwe as a central place. Similarities in material culture strongly
indicate that Period III (Class 3) evolved out of Period II (Class 2) (Sinclair 1987). Dating

980
 Shadreck Chirikure & Innocent Pikirayi

Research
Figure 4. Types of wall fabric (modified from Whitty 1961, Garlake 1973 and Collet et al. 1992).

between the twelfth and early thirteenth centuries on the hill, Period III is associated with
substantial clay plastered houses and may have been associated with the first stone walling on
the site (Robinson 1961b; Pikirayi 2001). Period IV covers about two centuries stretching
from the late thirteenth to the early sixteenth century (Collett et al. 1992). It has been
subdivided into three phases – IVa, IVb and IVc. Phase IVa developed out of Period III and
the local pottery (Class IVa) has features of both Class 3 and 4 and it is sometimes referred
to as transitional pottery (Robinson 1961a). This period witnessed the first expansion of the
Zimbabwe culture settlement from the hill into the valley. Period IVb is associated with
the florescence of the Zimbabwe culture up to the mid-fifteenth century. The local pottery
(Class IVb) is lavishly burnished with graphite. Period IVc is associated with later settlement
in the valley and subsequent abandonment of the site. The recovery of Ming Dynasty
porcelain from the Lower Valley enclosures indicates that Period IVb ended in the early
sixteenth century (Collett et al. 1992). Lastly, Period V represents the re-occupation of
Great Zimbabwe more than three centuries later. The pottery is akin to that used by the
nineteenth-century Karanga peoples who lived around the site (Garlake 1973).
The architectural chronology of Great Zimbabwe is also consistent with this sequence
(Schofield 1926; Whitty 1961; Garlake 1970; Chipunza 1997). According to Whitty (1961)
the earliest, poorly coursed, stone walls, which he termed P and PQ, were found on the
Hill Complex, followed by the neatly coursed Q walls in the valley, the site of the elliptical
Great Enclosure, and the other valley enclosures. The uncoursed R walls were attributed to
the later occupation (Figure 4). However, this model was weakened by the existence of P

981
 Table 2. Radiocarbon dates from Great Zimbabwe. The radiocarbon dates were calibrated using the Radiocarbon Calibration Program (Calib Revs
5.0.2. Copyright 1986-2005 M. Stuiver and P.J. Reimer). Modified from Huffman & Vogel 1991 and Stuiver & Reimer 1993.
Lab. No. Date AD Cal Age +
− 1 sigma Cal Age +
− 2 sigma Context Comment

Inside and outside the dry stone walls: revisiting the material culture of Great Zimbabwe
The Hill Complex
Periods 1 and 2 (development of farming communities)
M-914 1075 −+ 150 1040-1290 900-920 (1%) Test I, level 13. Gumanye Chivowa Hill dates 970 +
− 35
950-1410 (99%) pottery, end of Period 2 (Pta-1919), 1060 +− 40
(Pta-1922) and Gumanye
Hill 1020 +
− 35 (Pta-1916)
Pta-1983 670 +
− 40 721-741 (16%) 682-891 Test V, Western Enclosure, Zhizo Huffman & Vogel 1991 Period
770-873 (84%) pottery Ib
M-913 320 +− 150 348-369 (5%) 136-772 Test V, Western Enclosure, Huffman & Vogel 1991 Period
378-651 (95%) charcoal with fragments of Ia
dhaka, Gokomere/Ziwa
pottery
982

Periods 3 and 4 (development of stone walling and solid clay (dhaka) floors and clay-built houses)
M-915 1440 +− 150 1320-1350 (11%) 1220-1690 (96%) End of Period 4, Test I, level 5, Huffman and Vogel 1991,
1380-1520 (59%) 1730-1860 (4%) floor c Robinson 1961a
1540-1630 (30%)
Pta-2706 1370 +− 50 1328-1337 (11%) 1311-1360 (25%) Test I, level 7 Robinson 1961a
1391-1438 (89%) 1379-1452 (75%)
Pta-1192 1310 +− 50 1312-1359 (66%) 1293-1420 Colophospermum Mopane Robinson 1961a, Huffman
1380-1405 (34%) lintel, covered passage and Vogel 1991
Pta-1986 1310 +− 45 1314-1357 (67%) 1295-1417 Test I, floor g Robinson 1961a, Huffman
1381-1404 (33%) and Vogel 1991
Pta-2704 1280 +− 45 1304-1362 (80%) 1288-1402 Test I, level 11, floor h1, sorghum Robinson 1961a, Huffman
1377-1391 (20%) and Vogel 1991
Pta-745 1280 +− 30 1306-1329 (37%) 1297-1395 PWD face above Test VI Huffman and Vogel 1991
1336-1361 (42%)
1378-1391 (21%)
Pta-1985 1260 +− 45 1294-1323 (40%) 1283-1397 Test I, level 11 floor i Robinson 1961a
1346-1388 (60%)
 Pta-2705 1190 +
− 50 1233-1245 (11%) 1218-1324 (76%) Above Test V: South Wall, dating Robinson 1961a, Huffman
1264-1311 (68%) 1344-1389 (24%) early P walling and Vogel 1991
1359-1379 (21%)
Pta-1984 1100 +
− 40 1211-1270 1162-1172 (2%) Test I, floor j Huffman and Vogel 1991
1175-1281 (98%)
The Great Enclosure
Trenches 3-6 and lintels from one of the adjoining walls
SR-47 1380 +− 90 1320-1360 (31%) 1280-1510 (96%) Earliest dhaka lens and P-walling, Summers 1961, Huffman and

Shadreck Chirikure & Innocent Pikirayi

1380-1460 (69%) 1580-1620 (4%) Enclosure 1 Vogel 1991
Pta-1594 1310 +− 40 1316-1355 (67%) 1298-1413 Lintel Summers 1961, Huffman and
1382-1402 (33%) Vogel 1991
Pta-792 1300 +− 50 1309-1360 (72%) 1290-1415 Lintel Summers 1961, Huffman and
1378-1400 (18%) Vogel 1991
Pta-2694 1250 +− 40 1292-1319 (44%) 1283-1392 Trench 8, level 6, layer sealing Summers 1961
983

1351-1385 (56%) midden

Pta-2693 1240 +− 45 1292-1319 (44%) 1276-1394 Trench 5, level 9, earliest dhaka Summers 1961
1351-1385 (56%) lens and P-walling, Enclosure 1
The valley and peripheral structures
Pta-2429 1410 +− 40 1410-1442 1391-1458 Basal early dhaka floors, Trench 3, Huffman and Vogel 1991
Nemanwa Ruin
Pta-2423 1400 +
− 50 1401-1444 1318-1353 (10%) Camp Ruins, Z4, midden, 30-43 Huffman and Vogel 1991
1384-1462 (90%) cm
Pta-1208 1350 +
− 50 1322-1347 (34%) 1302-1364 (40%) Camp Ruins, Z1, House 35 Huffman and Vogel 1991
1387-1426 (66%) 1376-1444 (60%)
Wits 774 1350 +
− 40 1324-1343 (29%) 1312-1359 (35%) Camp Ruins, Z4, midden, level 4 Huffman and Vogel 1991
1389-1427 (71%) 1379-1441 (65%) (71-83 cm), underneath
Pta-2423

Research
 Inside and outside the dry stone walls: revisiting the material culture of Great Zimbabwe

Figure 5. Monumental walls in the Hill Complex. The eastern wall of the Western Enclosure bordering the Eastern Enclosure.

and Q walling on both the hill and in the valley, tentatively suggesting that both parts of
the site were built according to a preconceived plan, which was then elaborated over time.
The P and PQ walls are supposed to have been built in Period IVa (Robinson 1961a;
Huffman 2007) while the Q walls were constructed in Period IVb (Whitty 1961). On the hill,
Chipunza (1994) concluded that P and PQ courses of the Western Enclosure mark the focus
of settlement before it expanded to the Eastern Enclosure (Figure 5). Chipunza (1997) also
considered the spatial connectivity, construction affinity and stylistic succession of the walls
on the hill and concurred with Whitty (1961) that these structures had evolved over time.
A similar architectural development is noted in the valley. According to Summers and
Whitty (1961), the construction of the Great Enclosure developed in stages as shown by
the mixture of P and Q coursing (Figure 6). Wall construction began around the P-coursed
enclosure (no. 1 on the plan) and from this initial core area, additional PQ walls were built
over time (e.g. no. 15). In contrast to these early features, the most architecturally elaborate
walls in the Great Enclosure such as the girdle wall and the conical tower were built in Q
style (Figure 7). These Q walls have butted joints, revetments and rounded entrances.
Adjacent to the Great Enclosure, the Upper Valley enclosures were almost exclusively built
in P and PQ style (Collett et al. 1992), but the Lower Valley enclosures were exclusively
built in Q style and contain typical Q walling features. Therefore, these Lower Valley

984
 Shadreck Chirikure & Innocent Pikirayi

Research
Figure 6. Plan of the Great Enclosure showing the constituent ruins, after Lance Penny (National Museums and Monuments
of Zimbabwe).

enclosures are late additions to the site (Collett et al. 1992). Interestingly, they reproduce on
a smaller scale the trademark features of the Great Enclosure and in particular the conical
towers, curved batter walls and round entrances (Caton-Thompson 1931: 15).
Some of the walls of Great Zimbabwe were decorated with motifs which give them a strong
visual appeal. For example, the girdle wall of the Western Enclosure of the Hill Complex is
capped with stone monoliths and conical towers (Garlake 1973; Huffman 1996). There is a
herringbone design inside the Water Gate, and a dentelle pattern in the Western Enclosure
(Hill Complex). Vertical grooves (see Huffman 1996) also exist in the same enclosure and are
fairly common in the valley enclosures. Perhaps the most spectacular decoration is the chev-
ron design richly adorning the exterior girdle wall of the Great Enclosure (Figures 6 and 8).

On the missing link: the material culture of Great Zimbabwe (periods

III and IV)
The assemblages of artefacts and other cultural material are crucial, not only for
endorsing the architectural and stratigraphic sequences, but for the proper interpretation

985
 Inside and outside the dry stone walls: revisiting the material culture of Great Zimbabwe

of the different activities in each part of the site. However, this key aspect has been
relatively neglected up to now. Only the material culture relating directly to the builders
of Great Zimbabwe, i.e. belonging to Periods III and IV, will be considered here.
Period III pottery (Class 3) was
only found on the hill. It was well
fired, finely polished, with occasional
graphite burnishing (Caton-Thompson
1931; Robinson 1961b; Garlake 1973).
The most characteristic vessel types were
shouldered jars with everted rims or straight
necks. Decoration is very rare. With time,
Class 3 pottery developed new features
and this has been designated Class 4a or
transitional pottery (Period IVa). Class 4a
was found in virtually all the P and PQ
walled enclosures. Eventually, transitional
pottery gave way to Zimbabwe Pottery
(Class 4b) which was prevalent during the
state’s florescence. Class 4 pottery is lavishly
burnished with graphite and it consists
of shouldered pots with tall necks and
flared rims and spherical pots with very
short necks and heavily rolled rims (Garlake
Figure 7. The conical tower in the Great Enclosure in Q
style (courtesy of Innocent Pikirayi). 1973: 112; Huffman & Vogel 1991).
Metalworking evidence and metal objects
have also been recovered from different areas. An iron smelting furnace (Hall 1905)
and a gold melting furnace (Bent 1892) were recovered on the Hill Complex. Iron
slag was recovered on the hill (Robinson 1961a), in the Great Enclosure (Willoughby
1892; Hall 1905; Caton-Thompson 1931) with the Valley Enclosures yielding both
iron slag and iron blooms (Collett et al. 1992). Crucibles, ingots and casting
spills were also found in Great Zimbabwe’s walled enclosures (Bent 1892: 221).
Additional evidence of metalworking includes tuyeres, iron hammers, chisels, pincers and
wire drawing equipment (MacIver 1906; Hall 1910; Caton-Thompson 1931). Iron tools
(chisels, knives, hoes, arrowheads, spearheads and axes) form part of the inventory of
utilitarian objects. The assemblage of copper, bronze and gold objects largely consisted
of beads, thin sheets and other objects used for ceremonial and decorative purposes.
Therefore, metalworking evidence and metal objects were present in most of the constituent
walled enclosures making up Great Zimbabwe.
A wide array of soapstone objects range from the famous birds to bowls (Figure 9),
some imitating the architectural motifs already mentioned. Fragments of decorated and
undecorated soapstone bowls have been found on the hill (Bent 1892) and within the
Great (Willoughby 1892) and Valley enclosures (Caton-Thompson 1931). Perhaps
the most important category of soapstone finds is the collection of eight soapstone

986
 Shadreck Chirikure & Innocent Pikirayi

Research
Figure 8. The decorated outer girdle wall, Great Enclosure showing the chevron design (courtesy of Innocent Pikirayi).

Figure 9. Soapstone bowls from Great Zimbabwe: 9a) complete soapstone bowl decorated with herringbone pattern (Iziko
Museum of South Africa); 9b) fragment of soapstone bowl decorated with animal motifs, Hill Complex (housed in Iziko
Museums of South Africa). Photographs by Shadreck Chirikure.

birds. Seven of the birds were recovered from the Hill Complex while the eighth was
recovered in the Valley Enclosures (Matenga 1998). On the hill, only one bird was
found in the Western Enclosure while the remaining six were recovered from the Eastern
Enclosure. The Eastern Enclosure yielded meagre amounts of cultural debris and the
existence of platforms and monoliths has suggested the use of this enclosure for priestly
functions (Garlake 1973; Chipunza 1994; Huffman 1996). The flakes of worked soapstone
found in the Hill Complex middens would suggest that soapstone working was practised
here. The recovery of spindle whorls suggests weaving was an important activity carried
out at Great Zimbabwe. Pottery and soapstone spindle whorls were recovered in varying
proportions on the hill, and in the valley and Great Enclosures.

987
 Inside and outside the dry stone walls: revisiting the material culture of Great Zimbabwe

Figure 10. Period IVa glass beads from Great Zimbabwe courtesy of Marilee Wood.

Imported artefacts constitute a significant part of the material culture recovered at Great
Zimbabwe. Glass beads (Figure 10) have been recovered in fair amounts at the site with the
largest number being associated with the hoard found in the Renders Ruin in the valley
(Hall 1905; Caton-Thompson 1931; and see below). Modest amounts of Chinese celadon
and Near Eastern earthenware were found across the site. However, Arabian glass was only
recovered in the Great and Valley Enclosures. The Lower Valley Enclosures are the only
places to have yielded sixteenth-century Ming Dynasty pottery. Overall, the distribution of
imports suggests that the occupants of the stone walls had access to exotic goods.
Material culture also includes immobile features. Gravel floors make their first appearance
towards the end of Period III on the hill (Robinson 1961a), and become an integral feature
of virtually all the stone enclosures in Period IV. A deep succession of such house floors was
excavated in the Western Enclosure on the hill (Douslin 1922; Robinson 1961a; Figure 3).
Despite coming in different sizes, some common features of the house floors include low
benches, fire places and pot stands (Huffman 1996). Because of the history of plundering
on the site, one cannot establish the development of house types used over time. House
floors similar to those on the hill were found in the Great Enclosure and in the adjacent
valley (MacIver 1906; Hall 1910; Caton-Thompson 1931; Summers et al. 1961).
While these categories of material culture were found in almost every enclosure, Richard
Hall (1905) discovered a spectacular hoard in the Renders Ruin which contained gold
wire, iron spoons, a lamp stand, copper box, two finger rings, several hundred thousand
glass beads and several kilograms of wire, cowrie shells and coral. This probably signifies
the presence of a resident trader at Great Zimbabwe. This practice has been historically
documented in the Mutapa state, one of the successors to Great Zimbabwe (Pikirayi 1993;
Chirikure et al. 2001). By extension, this hoard represents royal control over trade and
exchange relationships in the Zimbabwe state.

Discussion
According to Huffman (1981; 1982; 1985), it is possible to understand the spatial correlates
of Great Zimbabwe’s dry stone walls using a binary-coded cognitive framework supported by
ethnography. Huffman concluded that the kings at Great Zimbabwe resided in the Western
Enclosure of the Hill Complex while the Eastern Enclosure served as a ritual centre. The
Great Enclosure in the valley was interpreted as a centre for initiation (see Huffman 1985),
while the Valley Enclosures were the residences of the royal wives (Huffman 1996).
Huffman’s model has been criticised for presenting the picture of a society in stasis
for 200 years (Beach 1998). Like Sinclair (1987), Beach (1998) made recourse to Shona
ethnography and history of political succession to argue that the ruler’s residences had more

988
 Shadreck Chirikure & Innocent Pikirayi

likely changed during Great Zimbabwe’s 200-year florescence. Thus the Great Enclosure
was not an initiation centre nor were the valley enclosures residences for royal wives: they
were centres adopted by successive rulers.
However, Sinclair and Beach did not support their hypothesis with archaeological

Research
evidence. If we were to do so, we can bring both the archaeological sequence and the
archaeological spatial distribution to bear in order to create a new model. This does indeed
endorse the idea of a shifting focus during the Great Zimbabwe periods III and IV and
effectively eliminates Huffman’s structuralist hypothesis.
The combined archaeological sequence and architectural chronology is consistent with
an expanding and shrinking settlement. Because no other part of the site has Period III
remains, the early foundations of Great Zimbabwe must be located on the hill. The earliest
monumental walls at Great Zimbabwe are the P and PQ walls of the Western Enclosure (Hill
Complex) (Chipunza 1994; 1997). Because the contemporary walls in the valley were less
monumental (Summers & Whitty 1961), this is where the first rulers of the state lived. From
this formative area, settlement then expanded into the areas with Period IVa occupation,
including P and PQ walling in the Great Enclosure and the Upper Valley. The exclusively
Q style Lower Valley enclosures were added in Period IVb. Over time, the development of
the elaborate and monumental Q styled girdle walls of the Great Enclosure suggest that
political succession was passed to an individual living there. Thus, there was a shift in the
centre of power from the hill to the Upper Valley, and from the Upper Valley to the Lower
Valley (Summers 1961; Garlake 1973).
It would seem that the rulers who were based in the Great Enclosure presided over the state
during its most affluent period. Subsequently, the centre of power moved to the exclusively
Q walled Valley Enclosures (Collett et al. 1992). This was the last place to be abandoned,
long after the Great Enclosure and the Hill Complex had been deserted, or were serving
lesser stately functions (Pikirayi 1993; Chirikure et al. 2001). Oral traditions claim a direct
link between Great Zimbabwe and the Mutapa state. Therefore, the sixteenth-century Ming
Dynasty porcelain from the lower valley would suggest that a small population in the valley
continued to occupy the site after most of the inhabitants had moved to other areas (Collett
et al. 1992).
The thesis of changing rulers’ residences is adequately supported by the distribution of
material culture found inside the stone walls of Great Zimbabwe. Although the dates of the
assemblages change, there is a remarkable similarity in the range of objects and activities
carried out in the earlier and later enclosures. Each assemblage covers not only utilitarian
and ceremonial objects but also ritual and craft production activities. Imported material
culture also has an all-encompassing distribution, just like the locally-produced objects.
This almost homogenous distribution of imports in all the major areas would suggest some
form of equal access to resources on the part of those living there. If the residences of rulers
changed over time as we contend, then this similarity in assemblage is hardly surprising.
In the Shona world, political succession does not follow the principle of primogeniture
(Bourdillon 1976). Instead, succession follows the system of ‘houses’ whereby if the founder
of a state has many sons, political succession alternates in all these houses starting from
the eldest to the youngest, and then reverts back to the house of the eldest son (Holleman
1952; 1969). Often, when a ruler dies, his successor does not move into the deceased’s

989
 Inside and outside the dry stone walls: revisiting the material culture of Great Zimbabwe

residence. He usually rules from his present homestead and depending on his power and
influence can extend its grandeur. Therefore, it would seem that the Hill Complex, the
Valley and Great Enclosures were, at one time or another, residences of rulers during the
200-year long florescence of the state. The homestead of the founder of the state invariably
assumes a religious significance (Gluckman 1937) and this probably explains why the Hill
Complex has always remained the site of an important shrine (Ndoro 2001; Fontein 2006).
It can be noted that the hypothesis of a changing centre of power is also applicable to other
Zimbabwe type sites on the plateau. In northern Zimbabwe, the stone-walled enclosures are
actually named after particular rulers such as Mutota (Mutota’s Zimbabwe) and Rusvingo
waKasekete (after Kasekete) (Beach 1980).
So far, the story presented by the material culture from Great Zimbabwe is consistent
with the major parts using identical material culture and being spaces for typical male and
female pursuits. This is at odds with structuralist interpretations of the site. Huffman (1996)
confidently argued that royal wives occupied the valley for the duration of Great Zimbabwe’s
florescence. He based his argument on the existence of ‘female symbols’ which primarily take
the form of vertical grooves on some of the stone structures. In a more recent publication,
Huffman (2007: 405) argues that royal wives lived together under the authority of the first
wife, and maintained that the valley complexes best served this function. The expectation is
that the material culture recovered from within the Valley Enclosures should be consistent
with an exclusively female domain. Yet, the material culture reveals the presence of both
male and female activities, as is common in Karanga societies (Aschwanden 1982). In
particular, the presence of metalworking slag and iron blooms all falling within the domain
of male activities show that there was a sizeable male presence in the lower valley enclosures.
Furthermore, that the lower valley enclosures were the last place to be abandoned raises
serious questions regarding why royal men would leave royal wives behind when abandoning
the site. This further casts doubts on the royal wives hypothesis and strengthens the point that
the lower valley housed the rulers at a time when Great Zimbabwe’s influence was waning.
Using Venda ethnography, Huffman (1996; 2007: 407) also argues that the Great
Enclosure was used for circumcision and acted as a pre-marital school for boys and girls
known as the Domba. He cited the existence of symbols for different age groups from
the young to the old and ritual objects that supported his initiation centre hypothesis.
Nevertheless, an examination of the distribution of material culture at Great Zimbabwe
shows that the so-called ritual objects are found in stratigraphic contexts in the valley and
on the hill (Matenga 1998). Apart from containing modest amounts of metalworking
evidence such as wire drawing plates, some slag, gold cake and utilitarian and non-
utilitarian objects, the Great Enclosure also possessed a fair share of imports, some of
which have never been recovered on the hill (for example the Islamic glass) (Summers &
Whitty 1961).
More importantly, the Domba did not take place regularly because it was dependent on
the number of young people ready to participate and the nature of the harvest (Stayt 1931).
As such, initiation centres tended to be impermanent structures built of perishable materials.
The objects of instruction consisted of symbolic objects which were kept in the chief ’s hut
when not in use (Stayt 1931). Archaeologically, the Domba institution is unlikely to have
left significant fingerprints. By contrast, the Great Enclosure is a permanent building whose

990
 Shadreck Chirikure & Innocent Pikirayi

construction took place over a long time. It had a broad-based material culture that included
local pottery, spindle whorls, symbolic objects, metalworking evidence and lavish imports.
This assemblage is similar to that found on the Hill Complex and in the valley; areas which
Huffman agrees were not initiation centres. The presence of platforms and figurines (which

Research
Huffman used to support his hypothesis) is equally consistent with Shona cultural beliefs in
which each household has a place to propitiate the ancestors (Bourdillon 1976; Aschwanden
1982; Gelfand 1973).
Identifying the ceremony and rituals of initiation among the ancient Karanga also remains
tenuous in the absence of supporting written and oral evidence (see Ashwanden 1982).
Although the Venda, who settled in northern South Africa in the seventeenth century, were
once part of the Karanga people, they soon interacted with neighbouring Sotho-Tswana
communities, creating a Venda identity which is in many ways different from the Karanga
(Beach 1980). According to Stayt (1931: 125), the cultural practice of circumcision was not
an indigenous institution amongst the Venda (see also Blacking 1985). It was introduced
through centuries of interaction with the Lemba and Sotho-Tswana societies. As such, the
Domba had no historical derivation from the Karanga, north of the Limpopo (Aschwanden
1982; Blacking 1985; Beach 1998). Nevertheless, the Venda retained some elements of the
Karanga such as the concept of sacred leadership. This must question the suitability of a
hybrid Venda culture as an analogy for the Karanga worldview.
Holl (1996) suggested that the structuralist interpretation should be synchronised with
the chronological development and by implication, the material culture of the site. While
Great Zimbabwe’s expansion was probably situated in existing mental templates (Huffman
1986; Holl 1996), it is difficult to apply a structuralist model to understanding the meaning
of space at different periods during the site’s development. For example, the model cannot
be used to understand the organisation of space while the lower valley was the only area
inhabited. Does this mean that interpretations based on structuralist theories have no future
at Great Zimbabwe? Far from it, but they must be supported by artefacts, architectural
history and chronology.

Conclusion
A critical assessment of the chronology, architectural history and material culture has shown
that Great Zimbabwe emerged from local farming communities as a series of aristocratic cen-
tres succeeding each other in a manner consistent with Shona systems of political succession
and chiefly politics. The focus of power moved from the Western Enclosure on the hill in
the twelfth century, to the Great Enclosure, the Upper Valley and finally the Lower Valley
in the early sixteenth century, where Great Zimbabwe’s prominence was destined to end.
This sequence invalidates structuralist hypotheses which assume that different parts of the
ruin were active at the same time and could thus be dedicated to different activities, rituals
or genders. Symbolic beliefs are part of the broad social structure which connects disparate
facets of human experience within a chronological and cultural framework (Hodder 2007).
The priority at Great Zimbabwe is to give more value to the existing data and finds.
Great Zimbabwe’s archaeology is currently elite archaeology; more work needs to be done
on the commoner areas that formed part of the settlement. We still await the publication

991
 Inside and outside the dry stone walls: revisiting the material culture of Great Zimbabwe

of the excavations conducted during the early 1970s. These results will throw some light
on issues of production and the circulation of goods, as well as the use of space in non-elite
residences. Pursuing some of these issues will awake the archaeology of Great Zimbabwe
from its temporary siesta of the last 30 or so years.

Acknowledgements
This paper benefited from ideas by Simon Hall, Edward Matenga, Munyaradzi Manyanga, Webber Ndoro,
Gilbert Pwiti, Judith Sealy and Joseph Vogel (AltaMira Press). We would like to unreservedly thank them for
sharing their thoughts with us. We are also indebted to the late David Beach, Tom Huffman and Paul Sinclair’s
insights which created an arena for healthy academic debates. Paul Lane, an anonymous referee and the editor
of Antiquity offered suggestions that improved the scope of this paper. Lastly, we thank our students past and
present for the lively discussions that we have had over the years.

References CHIRIKURE, S. & G. PWITI. 2008. Community

participation in archaeology and heritage
ASCHWANDEN, H. 1982. Symbols of life: an analysis of the management: case studies from southern Africa and
consciousness of the Karanga. Gweru: Mambo Press. elsewhere. Current Anthropology 49(3): 467-85.
BEACH, D.N. 1980. The Shona and Zimbabwe CHIRIKURE, S., I. PIKIRAYI & G. PWITI. 2001. A
900-1850. Gweru: Mambo Press. comparative study of Khami pottery, Zimbabwe, in
–1998. Cognitive archaeology and imaginary history at F. Chami & G. Pwiti (ed.) Southern Africa and the
Great Zimbabwe – comments and reply. Current Swahili world (Studies in African Past 2). Dar es
Anthropology 1: 47-72. Salaam: Dar es Salaam University Press.
BENT, J.T. 1892. The ruined cities of Mashonaland. COLLETT, D.P., A. VINES & G. HUGHES. 1992. Dating
London: Longmans. and chronologies of the Valley Enclosures:
BLACKING, J. 1985. The Great Enclosure and Domba. implications for the interpretation of Great
Man 20: 542-3. Zimbabwe. African Archaeology Review 10: 139-61.
BOURDILLON, M.F.C. 1976. The Shona Peoples. Gweru: DART, R. 1955. Foreign influences on the Zimbabwe
Mambo Press. and pre-Zimbabwe eras. Native Affairs Department
Annual 32: 19-30.
BRUWER, A.J. 1965. Zimbabwe: Rhodesia’s ancient
greatness. Johannesburg: Hugh Keartland. DOUSLIN, H. 1922. Recent explorations at Zimbabwe.
Proceedings of the Rhodesian Scientific Association 20:
BURKE, E.E. 1969. The Journals of Karl Mauch.
61.
Salisbury: National Archives of Rhodesia.
FONTEIN, J. 2006. Silence of Great Zimbabwe: contested
CATON-THOMPSON, G. 1931. The Zimbabwe Culture:
landscapes and the power of heritage. London: UCL
ruins and reactions. Oxford: Clarendon Press.
Press.
CHIPUNZA, K. 1994. A diachronic analysis of the standing
GARLAKE, P.S. 1970. Rhodesian ruins: a preliminary
structures of the Hill Complex at Great Zimbabwe
assessment of their styles and chronology. Journal of
(Studies in African Archaeology 8). Uppsala:
African History 11: 495-513.
Societas Archaeologica Upsaliensis.
–1973. Great Zimbabwe. London: Thames & Hudson.
–1997. A diachronic analysis of the standing structures
of the Hill Complex at Great Zimbabwe, in –1982. Great Zimbabwe described and explained.
G. Pwiti (ed.) Caves, monuments and texts: Gweru: Mambo Press.
Zimbabwean archaeology today (Studies in African GAYRE, R. 1972. The origins of Zimbabwean civilisation.
Archaeology 14): 125-41. Uppsala: Department of Salisbury: Galaxie.
Archaeology and Ancient History, Uppsala GELFAND, M. 1973. The genuine Shona. Gweru:
University. Mambo Press.
CHIRIKURE, S. 2007a. Archaeology beyond borders: GLUCKMAN, M. 1937. Mortuary customs and the belief
Zimbabwean archaeology in a regional perspective. in survival after death among the South-Eastern
Keynote address, National Forum on Zimbabwean Bantu. Bantu Studies 11: 117-36.
Archaeology, 6-8 September 2007, Harare.
HALL, M. 1987. The changing past: farmers, kings and
–2007b. Metals in society: iron production and its traders in southern Africa 200-1860. Cape Town:
position in Iron Age communities of southern David Philip.
Africa. Journal of Social Archaeology 7: 72-100.

992
 Shadreck Chirikure & Innocent Pikirayi

–1995. Great Zimbabwe and the Lost City: the cultural PIKIRAYI, I. 1993. The Archaeological identity of the
colonisation of the South African past, in P.J. Ucko Mutapa State: towards an historical identity of
(ed.) Theory in archaeology: a world perspective. northern Zimbabwe (Studies in African Archaeology
London: Routledge. 6). Uppsala: Societas Archaeologica Upsaliensis.
HALL, R.N. 1905. Great Zimbabwe. London: Methuen. –2001. The Zimbabwe culture: origins and decline of

Research
–1910. Prehistoric Rhodesia. London: Fisher Unwin. southern Zambezian states. Walnut Creek (CA):
AltaMira.
HALL, R.N. & W.G. NEAL. 1902. The ancient ruins of
Rhodesia. London: Methuen. PWITI, G. 1996. Continuity and change: an
archaeological study of farming communities in
HODDER, I. 2007. Symbolic and structural archaeology.
northern Zimbabwe AD 500-1700 (Studies in
Cambridge: Cambridge University Press.
African Archaeology 13). Uppsala: Department of
HOLL, A.F.C. 1996. A diachronic analysis of the Archaeology, Uppsala University.
architecture of the Hill Complex at Great
ROBINSON, K.R. 1961a. Excavations of the Acropolis
Zimbabwe by K.T. Chipunza. African
Hill. Occasional Papers of the National Museums of
Archaeological Review 13(1): 77-85.
Southern Rhodesia 3(23A): 159-92.
HOLLEMAN, J.F. 1952. Shona customary law: with
–1961b. Zimbabwe pottery. Occasional papers of the
reference to kinship, marriage, the family and the
National Museums of Southern Rhodesia 3(23A):
estate. Cape Town: Oxford University Press.
193-226.
–1969. Chief, Council and Commissioner: some problems
–1961c. Zimbabwe beads. Occasional Papers of the
of government in Rhodesia. Assen: Afrika
National Museums of Southern Rhodesia 3(23A):
Studiencentrum.
227-35.
HUFFMAN, T.N. 1981. Snakes and birds: expressive
SCHOFIELD, J.F. 1926. Zimbabwe: a critical
space at Great Zimbabwe. African Studies 40:
examination of the building styles employed. South
131-50.
African Journal of Science 23: 971-86.
–1982. Archaeology and ethnohistory of the African
SINCLAIR, P.J.J. 1987. Space, time and social formation: a
Iron Age. Annual Review of Anthropology 11:
territorial approach to the archaeology of Zimbabwe
133-50.
and Mozambique c. 0-1700 AD (Aun 9). Uppsala:
–1985. The Great Enclosure and Domba. Man 20: Societas Archaeologica Upsaliensis.
543-5.
STAYT, H.A. 1931. The Bavenda. London: Oxford
–1986. Cognitive studies of the Iron Age in southern University Press/International Institute of African
Africa. World Archaeology 18: 84-95. Languages & Cultures.
–1996. Snakes and crocodiles: power and symbolism in SUMMERS, R. 1961. Excavations in the Great Enclosure.
ancient Zimbabwe. Johannesburg: Witwatersrand Occasional Papers of the National Museums of
University Press. Southern Rhodesia 3(23A): 236-88.
–2007. Handbook to the Iron Age: the archaeology of SUMMERS, R. & A. WHITTY. 1961. The development of
pre-colonial farming societies in Southern Africa. the Great Enclosure. Occasional Papers of the
Scottsville: University of KwaZulu-Natal Press. National Museums of Southern Rhodesia 3(23A):
HUFFMAN, T. N. & J.C. VOGEL. 1991. The chronology 306-25.
of Great Zimbabwe. South African Archaeological SUMMERS, R., K. ROBINSON & A. WHITTY. 1961.
Bulletin 46: 61-70. Zimbabwe excavations. Occasional Papers of the
JEFFREYS, M.D.W. 1954. Zimbabwe and Galla culture. National Museums of Rhodesia 3(23A): 15-332.
South African Archaeological Bulletin 9: 152. THORP, C. 1998. Kings and commoners at Great
MACIVER, D. 1906. Medieval Rhodesia. London: Zimbabwe (National Museums and Monuments of
MacMillan. Zimbabwe Memoirs). Harare: Trustees of the
MATENGA, E. 1998. The soapstone birds of Great National Museum and Monuments of Zimbabwe.
Zimbabwe. Harare: Africa Publishing House. WAINRIGHT, G.A. 1949. The founders of Zimbabwe
NDORO, W. 1994. The preservation and presentation of civilisation. Man 49: 62-66.
Great Zimbabwe. Antiquity 68: 617-23. WHITTY, A. 1961. Architectural style at Zimbabwe.
–2001. Your Monument, our shrine: the preservation and Occasional Papers of the National Museums of
presentation of Great Zimbabwe (Studies in African Southern Rhodesia 3(23A): 289-305.
Archaeology 19). Uppsala: Department of WILLOUGHBY, J. 1892. A narrative of further excavations
Archaeology and Ancient History, Uppsala at Zimbabwe. London: George Philip.
University.

993

View publication stats