Cambridge Archaeological Journal http://journals.cambridge.org/CAJ

Additional services for Cambridge Archaeological Journal:

Email alerts: Click here Subscriptions: Click here Commercial reprints: Click here Terms of use : Click here

Beyond Static Models: An Evaluation of Present Status and Future Prospects for Iron Age Research in Southern Africa

Per Ditlef Fredriksen and Shadreck Chirikure

Cambridge Archaeological Journal / FirstView Article / May 2015, pp 1 - 18 DOI: 10.1017/S0959774314001115, Published online: 06 May 2015

Link to this article: http://journals.cambridge.org/abstract_S0959774314001115

How to cite this article: Per Ditlef Fredriksen and Shadreck Chirikure Beyond Static Models: An Evaluation of Present Status and Future Prospects for Iron Age Research in Southern Africa. Cambridge Archaeological Journal, Available on CJO 2015 doi:10.1017/ S0959774314001115

Request Permissions : Click here

Downloaded from http://journals.cambridge.org/CAJ, IP address: 84.215.22.142 on 07 May 2015 Beyond Static Models: An Evaluation of Present Status and Future Prospects for Iron Age Research in Southern Africa

Per Ditlef Fredriksen & Shadreck Chirikure

To what extent do we need structuralist cognitive settlement models such as the Central Cattle Pattern and the Zimbabwe Pattern for future research and understanding of Iron Age social life in southern Africa? How will alternative approaches enable us to progress beyond the present status of knowledge? While the three last decades of debate have un- derpinned key aspects of archaeological inquiry, notably questions of social change, gender dynamics, analytical scale and the use of ethnographic and ethnoarchaeological insights, the sometimes entrenched nature of the debate has in other respects hindered development of new approaches and restrained the range of themes and topics scholars engage with. In this article, we identify the issues of analytical scale and recursiveness as key to the development of future approaches and present an alternative framework through empirically grounded discussion of three central Iron Age themes: ceramics and the microscale, the spatiality of metal production and the temporality of stonewalled architecture.

To change analytical scale is to change perspective on 2007; 2008) have resulted in an impressive list of publi- social life. Issues of scale exist at a fundamental level cations. However, critics have pointed out that discus- of archaeological interpretation, influencing our per- sion of the models has restrained the range of themes ception of space and time. These issues are found at and topics with which scholars engage. For example, the core of the debate of the epistemological founda- the focus on spatial organization of settlements, and tion and use of the models for Iron Age settlement how this allows for certain reconstructions of ide- space in southern Africa. The Iron Age spans nearly ology and worldview, has directed attention away two thousand years, from the Early Iron Age (ad 200– from more ‘mundane’ aspects of Iron Age societies, 900), via the Middle Iron Age (ad 900–1300) to the Late such as subsistence strategies, patterns of resource Iron Age (ad 1300–1820), which ended with colonial- procurement, sequence of reconstruction, or reasons ism (Huffman 2007; 2012a). The era is associated with for abandonment (see Bonner et al. 2008, 11–12; Lane the beginning of crop agriculture, metalworking, pot- 2010, 312). tery making and settled life. Since their introduction The key focus in this article is how microscale in the early 1980s, the models known as the Central variation of archaeological data, remains of every- Cattle Pattern (CCP) and the Zimbabwe Pattern (ZP) day or ‘mundane’ aspects, may inform regional or (see below) have been instrumental in broadening our interregional frameworks in novel ways. We argue understanding of the time period. And applications of for recursiveness through mutually informing move- these ethnographically derived cognitive models and ment between analytical scales. Acknowledging the the ‘direct historical approach’ (Huffman 1986; 2001; significance of including context-specific insights in

Cambridge Archaeological Journal page 1 of 18 C 2015 McDonald Institute for Archaeological Research doi:10.1017/S0959774314001115 Received 06 Mar 2014; Accepted 17 Nov 2014; Revised 20 Oct 2014 Per Ditlef Fredriksen and Shadreck Chirikure

macroscale modelling,1 we have two main aims. The CCP and the ZP: issues of scale and temporality The first part of the paper is devoted to recog- nizing the root causes of the current situation and The introduction of structuralist modelling to under- to comprehend their implications for our under- stand the organization of settlement space (Huffman standing of the Iron Age. In the second part, we 1982) was a decisive turning-point in Iron Age stud- aim to demonstrate how we can use these insights ies. The combined emphasis on ideology and belief to develop alternatives. We do this by presenting systems enabled an approach where behaviour was a set of approaches to societal change, continuity not only integral to human cognition, but also pro- and variability in the archaeological record: to ce- vided understanding of how key ideological prin- ramic production in households, spatial dimensions ciples were inscribed onto settlements and land- of metalworking and the temporality of stonewalled scapes. Aided by radiocarbon dating, the approach architecture. was groundbreaking in allowing a turn from more A core argument is that lack of recursiveness classical culture-historical questions to processually between analytical scales, and thereby also between oriented archaeological analyses of past behavioural empirical, methodological and theoretical levels of in- patterns. As is well known to Iron Age archaeologists, quiry, is a key contributor to the present status for Iron both models are ethnographic constructs. Anthropol- Age research in southern Africa. After more than three ogist Adam Kuper (1980; 1982) had observed what decades of debate, in which the models’ inventor and he called the southern African Bantu Cattle Pattern, main proponent, Thomas Huffman, is prominent, the which described the late nineteenth-century Nguni models remain influential—not least in the sense that settlement organization. Following the direct histori- we will use here; that critics engage with them when cal approach, the application of the model, to be re- proposing alternatives. The partly entrenched nature named the Central Cattle Pattern (CCP), was extended of the debate contributes to the current epistemic stand- back to the Early Iron Age (EIA). still. By this we mean that, while archaeological use The CCP model is principally associated with of the models may produce interesting and nuanced Nguni and Sotho-Tswana speakers, and the set- results, often applying sophisticated field and labo- tlement pattern represents a ‘cultural package’ re- ratory methods, a critical theoretical reading reveals stricted to groups of Eastern Bantu speakers shar- that the epistemic foundation for the models, and ing certain distinct features (see Huffman 2001; particularly the treatment of ethnography by the di- 2012b, 124). A typical organization consists of an rect historical approach, remains largely unchanged. arc of houses around a central cattle corral. Gen- Rooted in structuralism and a belief in archaeology- der is an important structuring principle. Opposi- as-science, the models and their largely ahistorical tions of male/female, pastoralism/agriculture, ances- epistemic foundation, including a retreat from present tors/descendants, rulers/subjects and cool/hot are politics, have been characterized as outmoded, even represented spatially, either concentric or diametric. unethical (Meskell 2005; see also Fredriksen 2011; In this manner, the CCP also makes statements about Pikirayi & Chirikure 2011). rank and status (Huffman 2001; 2008; 2012b). The standstill may be hard to perceive without The Zimbabwe Pattern (ZP) is seen to have devel- a long-term discursive perspective on a time-span oped from the CCP, and may be represented through with significant societal change. We therefore evalu- the same kind of concept, but with a different result. ate the present status against the backdrop of the last Paul Lane (2005, 31) notes that the ZP has no his- three decades of Iron Age research in southern Africa. torical analogue and that all known examples there- In this period, countries such as South Africa, Zim- fore are archaeological. However, Huffman (2011, 38) babwe and Mozambique have seen profound societal has recently pointed out that Venda society presents transformation, while the concurrent wider archaeo- a twenty-first-century version. The main sources of logical discourse has felt the impact of postproces- analogy come from sixteenth-century Portuguese doc- sualism. Recognizing the significance of postcolonial uments and elements of recent Shona and Venda critique of notions that the past may exist indepen- ethnography. The ZP has a more restricted spatial dently of present politics (see e.g. Lane 2011), we seek and temporal distribution than the CCP. And, un- answers to two interlinked questions: to what extent like the CCP, which is associated with both elite and are the CCP and the ZP models needed for devel- commoners, the ZP was restricted to elite settlements, oping future understanding of social life in Iron Age while commoners retained a basic CCP settlement lay- southern Africa? And precisely how may alternative out. A key difference between the two models is that approaches enable us to move beyond the present the cattle enclosures in the ZP have been removed status? from elite areas and replaced by a court or assembly 2 Beyond Static Models

area. This is seen to reflect an increased emphasis on range of empirical data. Alternative viewpoints, per- political decision-making. Huffman (2011, 37) lists five haps in particular by younger researchers influenced components that each Zimbabwe capital needed to by various versions, and to varying degrees, of post- function: a palace, a court, a compound for the leader’s processualism, have been weighed against the exist- wives, a place for commoners, and a place for guards ing models by their main proponents and deemed (see also Huffman 2014). as failures because they do not explain the data acti- Are the models adequate as explanations about vated by the models as well as the models themselves. the Iron Age past? It would seem that some critics In our opinion, the ensuing epistemic tension (see be- have not yet fully grasped that the models are norma- low) has contributed to a somewhat unfruitful gap tive, apply at a general level and are not intended to which is at least partly generational, where alterna- account for variation below a certain analytical scale tive approaches, especially with regards to theoretical (see also Whitelaw 2012; 2013, for recent discussion position and analytical scale, could have been given relating to the CCP). An example, to which we will more time to breathe. return in more detail, is the cognitive interpretation of The subtext also has another overlapping dimen- the use of space at Great Zimbabwe, a prominent de- sion. The narrowing of themes and topics has im- bate point since the mid 1990s (Beach 1998 with com- plications for the range of factors to be taken into ments and reply; Huffman 1996). Different enclosures account when seeking multi-scalar understanding of are allocated to wives, initiation ceremonies and rit- past social dynamics. In particular, the subsuming of ual sisters by reading symbols on stone. A more recent small-scale variation and difference into generalized study revealed a more-or-less homogenous distribu- structuralist models favours the large-scale and cogni- tion of material culture, with objects typically made tive templates over the bodily experienced everyday, and used by both men and women, found across the thus limiting the possibilities for integrating individ- site. Contrary to the interpretation informed by the ually contextualized African experience into models ZP, the distribution suggests that the Valley Enclo- for understanding material culture and societal con- sures are not a royal wives’ area; neither is the Great tinuity, change or innovation (but see e.g. Huffman Enclosure an initiation centre (Chirikure & Pikirayi 2012a and Whitelaw 2013 for recent alternative argu- 2008). On a general level, the cognitive models’ fo- ments). Also, while possibly not unique in the wider cus favours horizontal distribution according to struc- discipline of archaeology, it is interesting to note how turalist binary principles, thereby overlooking types an approach inspired by processualism has generated of information that relates to detailed material culture more interest in cognitive and symbolic matters than distribution and the potential variability between sites in aspects such as household dynamics, economy and in such distributions (see Anderson 2009; Fredriksen subsistence. 2007). A practical consequence of the cognitive settle- As indicated, to work recursively is not merely ment models’ subsuming of variation is a tendency for a question of shifting between analytical scales. The generalizing from studies of important sites to those outcome is production of different forms of archaeo- that are not yet investigated. For example, research has logical knowledge. In our opinion, there is need for to date only been done on a handful of ZP sites, such as more studies of historical development at sites such Mapungubwe, Great Zimbabwe and Khami (Fig. 1). as Great Zimbabwe. Studies that may identify unique These sites represent a small percentage of known ZP characteristics of individual sites will enable closer sites, numbering over 300 (see Pikirayi 2013, 286). Yet scrutiny of layers of cultural development and inter- the ecological, political and socio-economic environ- action, thereby providing a more nuanced past (see ments of the sites differed from one another, which is Chirikure et al. 2012; 2013a). why they rose and declined at different points in time, with some enduring longer than others (see Chirikure et al. 2013a). It may be questioned whether the gener- An alternative knowledge production: alizing models are necessary for further development recursiveness three-fold of this rich and varied history in the future, as the sub- suming of variation is justified on the basis that sites The withdrawal of theory from debate and retreat were authored by people that held similar worldviews from present politics hints, to readers less intimate (see Chirikure et al. 2012; 2013a). Our approach can with southern African archaeology, that there is a po- thereby be seen in close relation to previous critiques litical subtext below the question about a regional in- of the direct historical approach, and the CCP and ZP terpretation. The particular view of archaeology-as- models as leaving us with ahistorical accounts and to science underpinning the models activates a certain invite circularity between ‘the ethnographic present’ 3 Per Ditlef Fredriksen and Shadreck Chirikure

Figure 1. Key sites and locations discussed, including the two comparative research areas in South Africa: the Magaliesberg and the Mpumalanga Escarpment. and past social life (Hall 1984a,b,c; 2005; Lane 1994/95; tion is therefore not merely about difference in scale. 1998; 2005). To move between macroscale and microscale is to Alternatively, as recently argued by Innocent change perspective on social life. More precisely, un- Pikirayi (2013, 296), the construction of settlement derstanding of small-scale household dynamics re- layout should be seen as intimately involved in the quires an engagement with the bodily-experienced process of constructing the worldview itself. Thus, everyday. Thus, to work recursively between scales rather than the structuralist models’ envisioning of requires a shift from the CCP and ZP models’ em- people enacting a predefined plan, where architec- phasis on spatial cognition, where binary codes are ture and material culture become passive expressions heuristic devices and the models comprise clusters of of power and social categories, the intimate, always interlocked principles (Huffman 2012a, 121), to spa- ongoing, recursiveness between the material and the tial experience. And an outcome of this shift is the pro- mental should be emphasized. The use of macroscale duction of different forms of knowledge about the mental templates exposes a problematic idea deep- past. rooted in structuralist thought, namely its unidirec- Critical to the concept of modernity (e.g. tionality. Templates work from mind to matter, not the Gonzalez-Ruibal 2006a,b; Schlanger 2004; Schnapp other way. This implies a pre-existing form that is sim- et al. 2004; Witmore 2006; Webmoor & Witmore 2008) ply impressed upon a passive material world (Ingold and informed by cosmopolitan perspectives on ar- 2000, 340–41). Underlying our alternatives is an on- chaeological research (Gonzalez-Ruibal 2009; Meskell tological shift, termed by Tim Ingold (2000, 216) as 2009), our main argument is that recursiveness must one from generalizing detachment to localizing en- be acknowledged as three-fold: 1) there is a constant gagement. This requires an awareness of the implica- dialectic between the past and the present; 2) relations tions of one’s choice of analytical scale; of how to an- between scales of analysis should be dialectical and alytically capture, for example, the dynamic links be- interactive; and 3) intimacies between the human tween the shifting materialities and practices of small- mind and the material world do not work in unidi- scale everyday life and changes to more large-scale rectional ways, i.e. from mind to matter, but are a regional economies. As indicated above, the ques- constantly recursive process.

4 Beyond Static Models

A political critique: from detachment to using unnecessary metaphors of rivalry and conflict. engagement For example, the models are to be ‘evaluated against rival hypotheses’ (Huffman 2004a, 69) in order to test Africanist archaeologists are presently debating whether alternatives ‘undermine the validity of cog- the complex moral and ethical issues related nitive models’ (Huffman 2009, 40). Consequently, in- to power/knowledge dynamics in post-colonial stead of a positive debate about data sets, methods and contexts (see, e.g., Giblin et al. 2014; Karega-Munene theoretical perspectives that could spiral questioning & Schmidt 2010;Lane2011; Schmidt 2009;Smith and analyses into novel domains, there are only two 2014;Stahl2010). For South Africa, Nick Shepherd possible outcomes: success or failure. That alterna- (2003, 843–4) notes that the idea of archaeology as a tives to the models are deemed failures is perhaps not science has been attractive to archaeologists in pe- surprising when the models’ theoretical foundation is riods of societal transformation. This has enabled withdrawn from debate and the ground rules are set in a retreat from society as an enduring feature, and advance: ‘To challenge [the CCP] model successfully, the present epistemic situation may be compared to critics must propose an alternative that interprets the that of the earliest days of postprocessualism. Well data better at the same scale of abstraction’ (Huffman known to many is the general meeting of the South- 2001, 31). In other words, critics cannot deviate from ern African Association of Archaeologists (SA3)2 held the set scale of abstraction (see below). in Gaborone on 5 July 1983. Called on to denounce The epistemic outlook of this form of cognitive apartheid by their African counterparts, the South archaeology provides grounds for separating the aim African members, who also constituted the major- of archaeology from its social practice, where differing ity, hesitated. The view that present politics had no schools ‘can be evaluated in terms of their effective- place in the pursuit of archaeological knowledge ness in disclosing a real past’ (Huffman 2004a, 69). about the past was, of course, rooted in the domi- This separation necessitates a particular definition of nant positivist paradigm (see Bonner et al. 2008). Mar- political archaeology that many professionals world- tin Hall (1984a,b,c) challenged the apolitical stand. wide would question. For most, the term implies com- This epistemic critique is still relevant, as Hall (2005, plex entanglements of past/present political practices 181–5) reiterated its essence more than two decades and ideologies, but in this instance (e.g. Huffman 2007, later, while also identifying key factors for the mod- 19–22) it is related exclusively to past hierarchies and els’ perpetuation: 1) the predominant interest in the processes of centralization. On this basis it is per- truth value of the ‘archaeological record’; 2) a set haps paradoxical that, in a response to critics argu- of values that favours fieldwork over other forms ing that this standpoint, grounded in a ‘modern view of inquiry; 3) an interest in systems of classifica- of science’ (Huffman 2004a, 66), allows for little un- tion; and 4) that these values have been underscored derstanding of past social change, it has been argued by a relative absence of counterbalancing intellectual that the critique represents a cultural appropriation trends. belonging to a specifically western worldview and Among the replies to criticism of the CCP and ideal. This, we learn, ‘denies Africa to speak for itself’ ZP models (e.g. Huffman 2001; 2007; 2008; 2010; 2011), (Huffman 2001, 32). one article (Huffman 2004a) in particular presents the A concrete legacy of the modernist science view view of theoretical differences. The archaeological dis- is the use of the term ‘Iron Age’. As an import of early course is grouped into two opposing schools, the Pro- and mid twentieth-century European models it im- cessual and the Postprocessual (Huffman 2004a, 66–7) plies an evolutionary stage bounded by ethnicity and and Colin Renfrew’s 1982 work Towards an Archaeol- culture, thus carrying a problematic heritage (Bonner ogy of Mind is still held to form the basis of a ‘cognitive et al. 2008). However, there is also an often overlooked archaeology’ (see also Huffman 1986). After convey- dimension in the inherent spatial stasis and its im- ing complex convolutions as a dualism, cognitive ar- plications for analytical scale, as the concept rests on chaeology is presented as the viable ‘middle way’. In the implicit view of the world as an ‘assumed isomor- our opinion, this simplification has two closely related phism of space, place and culture’ (Gupta & Ferguson outcomes. Firstly, as observed by Paul Lane (2005, 1997, 47), The statement that ‘Iron Age archaeology is 33–4), there is a misrepresentation of critics’ views Bantu archaeology’ (Huffman 1982), perennial to the and the level to which critique is directed. To mis- retrospective direct historical approach, is therefore represent is to change the premises for debate and is inseparable from a specific notion of time, develop- therefore an effective way to channel the discussion ment and history. In our opinion, a two-step progres- in a certain direction, even derail it. Secondly, the de- sion is helpful in order to enable future alternatives bate is presented as consisting of a constant either/or, that question the isomorphism. 5 Per Ditlef Fredriksen and Shadreck Chirikure

Firstly, we may instead ask: Is Iron Age archae- generalizing detachment and experiential localizing ology really only Bantu archaeology? This implies a engagement. shift in attention from specific groups of people, which We do not expect these two points to raise many requires research on certain fixed analytical scales, to eyebrows among archaeologists with a postprocessu- the multi-scaled spaces people, animals and things alist bent. There is a critical awareness that commu- dwell in. This way, it becomes clearer that archaeolog- nities of all scales and types must be understood in ical studies are not only about various Bantu groups relation to local meanings and histories (e.g. Meskell and their interaction in time/space. By looking at re- 2005, 73; Segobye 2005, 80; Weiss 2007, 415). This is an gions, spaces and places as meeting grounds for in- acknowledgement of the past as potentially having teraction between various groups and/or individuals been as dynamic as the present: layers of complexity with differing subsistence strategies, practices, belief are entangled in the definitions of terms like ‘com- systems and knowledges about the material world, munity’ and ‘indigenous’ (Chirikure & Pwiti 2008, we are constantly engaging archaeological evidence 468; see also Fredriksen 2011; Green 2008; Kuper 2003; by questioning what and who we are studying. This Shepherd 2002; 2003). may have epistemological importance, for example by contributing to erase old mental frontiers separating Theme I: ceramic studies, scale and agency western pastoralists from eastern agriculturalists (see, e.g., Fauvelle-Aymar & Sadr 2008,3;Marks2011, 130). Concurrently with this growing present/past aware- Significantly, similar theoretical points are raised by ness, ceramic studies in southern Africa seem to have recent critiques by East Africanists of the notion that undergone a transformation. Innocent Pikirayi (2007, the EIA equals initial appearance of Bantu-speakers in 288) observes that, since the early 1990s, archaeolo- the region. Although specific archaeological argumen- gists have shifted focus from identity construction tation may differ, the emphasis on ‘moving frontiers’ to studies of group interaction (see Calabrese 2007; (see Lane 2004a) and ethnic mosaics (e.g. Kusimba & Schoeman 2006). This emphasis is also found in more Kusimba 2005; LaViolette 2008;Stahl2004)sharethe recent ceramic work by Huffman (e.g. 2002; 2004b; epistemological concern with erasing unfruitful men- 2007), seeking to clarify the question of analytical scale tal boundaries. and social formations. The theoretical framework is, Secondly, we may move from abstract, ahistor- however, still rooted in structuralism (see Huffman ical cognitive modelling to developing approaches 2007, 115). Thus, while ‘ethnicity’ is now replaced that focus on localized experience, thereby enabling by ‘large-scale identities’, it remains a normative ap- us to work recursively between scales. Since the CCP proach, which leaves us with a paradox, as pointed and ZP models are ‘not designed to investigate the out by Fredriksen (2007, 127): since the more imme- detail of daily behaviour and dynamics’ and have diate social, economic and ritual contexts of ceramic to ‘subsume such differences in order to extract the production and use are rendered invisible, the subjec- common underlying principles’ (Huffman 2001, 24), tivity of the makers and users of the material culture they are rendered with little ability to understand are to a large extent invisible. We depend upon under- change arising from other factors than large-scale ex- standing the dynamics between subjectively experi- ternal forces (Lane 2005, 33). While this underscores enced social identities in order to understand changes the analyst’s choice of scale (e.g. Hall 1995; 1998; to ceramic technology.3 Hall et al. 2008), it also demonstrates Bassey W. An- In a response to this critique, Huffman has modi- dah’s (1995, 159) observation that even studies that fied his standpoint on the concept of agency (Huffman are well intended and broadly conceived have an 2008; see also Huffman 2012a, 123; Whitelaw 2012). outcome that is often essentializing and reductive. While defending the use of the normative models This spatial appropriation means that lived-in envi- there is also reference to Giddens’ (1984) theory of ronments as spheres of experience and intimate in- structuration. Huffman admits that the tension be- teraction are downplayed and even disappear alto- tween ideal norms and daily action—between the gether (see Fredriksen 2011). Importantly, these in- pressure for continuity and the potential for change— sights add to our understanding of the resistance to is an important social dynamic. It is, however, some- engage in aspects of social life on less general lev- what puzzling to find a largely phenomenological els, and also to the epistemic standstill for Iron Age critique of a structuralist approach to be described research. What is argued to be a mere question of de- as ‘fundamentally structuralist’ (Huffman 2008, 40). veloping new models to address the microscale actu- This recent acknowledgement of agency and norma- ally implies an epistemic shift. A recursive approach tive/daily tension is an interesting epistemic turn, as requires an ontological move between cognitive a significant factor for archaeologists who have voiced 6 Beyond Static Models

postprocessual critique seems to be the degree of influ- ideas throughout the LIA. As already pointed out, the ence from sociological thought, particularly Giddens use of cognitive models encourages generalization in- (1984) and Bourdieu (1977). This critique has high- stead of investigation of individual sites. To classify lighted two aspects in particular: how social practice settlements into versions of the CCP is to prioritize uni- changes over time (e.g. Davison 1988;Hall1986;Lane formity over difference. By collapsing meaning on the 2005; Maggs 1993) and an attention to microscale and ethnographic/historical source-side and the archae- gender dynamics (e.g. Delius & Schoeman 2008;Hall ological subject-side (Lane 2005, 30–34) onto a gen- 1998; 2000;Lane1998; Schoeman 1998a,b; Segobye eralized level, context-specific meanings from certain 1998). investigated sites are transferred to sites with simi- lar spatio-material features. Importantly, this is not Collapsing scales: when households engage only to invite timelessness: it is to embed abstract, landscapes appropriated space into the models’ temporal circu- An example of the kind of questioning that post- larity. The outcome is a mental template being ap- processual critique has introduced to Iron Age re- plied through a top-down approach wherever LIA search is the discussion of ceramics and Late Iron circular stone structures are found. The template has Age (LIA) stone walling. This offers the opportu- been instrumental in centring attention on the cat- nity to combine the first two forms of recursiveness tle/court/male head triad. To be clear, this is not we have discussed thus far, present/past and mi- to question its importance for understanding polit- croscale/macroscale, with the third: that between the ical development during the LIA, nor that the core human mind and the surrounding material world. of the CCP is based on the ideological centrality of In South Africa, it is well established that the use cattle (cf. Huffman 2010, 165). Rather, we stress that of stone by ancestral Tswana speakers during the LIA to make such broad generalizations is to leave sig- was intimately linked to changes in the landscape lo- nificant synchronic and diachronic differences in the cation of settlements and agriculturalists’ movement dark. Illustrating this point is Lane’s (1998) argument onto the interior highveld (e.g. Boeyens 2003;Hall that the adoption of stone walling and changes to in- 2007; Huffman 1996; 2007; 2008;Widgren2000). New ternal and external house layout might have been re- forms of interactions (Hall 2000; 2007) and economic, lated to changes in male–female power relationships social, and technological transformation (e.g. Boeyens as the wider socio-political context of proto-Tswana 2003; Huffman 1996; Maggs 1976) followed, including societies changed. Kathryn Fewster’s (2006) identifi- an increased emphasis on cattle keeping after c. ad cation of an apparent ‘adoption’ of a CCP principle 1700. The move was probably instrumental for sub- for settlement layout among Khoisan ‘Basarwa’ near sequent political development, contributing to a sys- Serowe in the twentieth century has additional impli- tem with increased opportunities for accumulation of cations for understanding agency. Specifically, Few- personal wealth (Reid 2004; Segobye 1998). However, ster’s work demonstrates the capacity of people con- while its origin is an ongoing matter of debate (e.g. stantly to re-negotiate perceived rules in order to make Sadr 2012; Sadr & Rodier 2012), the implicit notion sense of their lives. The implication is that structure that building in stone articulates some form of male cannot be disengaged from individual agency. Rather, power has gone largely unquestioned since Alinah the post-colonial persistence of cultural elements de- Segobye’s (1998) call for a re-examination of the CCP. scribed in the CCP testifies to the interplay between She identified the underlying assumption that women structure and agency. were passive commodities in social processes char- In our opinion, an example of an alterna- acterized by wealth and power-seeking male heads tive framework that considers intimate present/past, (see also Fredriksen 2007; 2011;Hall1995; 1998; Lyons microscale/macroscale and mind/material recursive- 2006). Recent analyses of crucible sherds from several ness is offered by a synchronic comparison of in- Iron Age sites (Chirikure et al. in press) have sup- teraction between two LIA agropastoral communi- ported her observations: the crucibles are basically ties and their surrounding landscapes in the South identical to domestic pottery made by women, but African interior: in the Magaliesberg region and the used in a typical male activity. This evidence indicates Mpumalanga Escarpment (Fig. 1). that women contributed significantly to a technical Focusing on what spatial and material changes and transformative process that underpins their im- to hearths mean for social interaction in households, portant role in reproduction and thus societal renewal Fredriksen (2007; 2011; 2012) has sought to under- and growth. stand Moloko pottery and microscale environments Another important aspect to consider is whether in associated settlements. The Moloko ceramic se- LIA stonewalls always articulate the same cultural quence (see Evers 1983; 1988; Huffman 2002; 2007)is

7 Per Ditlef Fredriksen and Shadreck Chirikure

dated from about ad 1300 and associated with ances- permanence through stone architecture. Women spent tral Sotho-Tswana speakers in southern Africa dur- significantly more time in the fields (Hall 2007)and ing the LIA. Concurrent with a significant decline this increased intimacy between the female body and in Moloko stylistic intensity (Hall 1998), significant soil and earth implied that the presence of her an- transformations occurred in the architectural envi- cestors became relatively more dominant in the fields ronment, household spaces and material surround- as well as in dwelling spaces associated with women ings of the people living in Moloko settlements in the and the processing of agricultural produce. In this Magaliesberg with the introduction of stone walling way, a change of awareness of the relationship be- around ad 1700. The early assemblages demonstrate tween female bodily experience, perception of culti- a much higher variation in ceramic style and types vation as female work (Guyer 1991;Hall1998)andthe than the post- ad 1700 equivalents. Importantly, the surrounding landscape developed as part of living in spatial changes coincided with a shift in pyrotech- stonewalled towns. nology: from multipurpose firebowls at the centre of pre-walling, multipurpose hut floors to more special- A regional comparison: the Magaliesberg and the ized hearths in designated spaces in the post-ad 1700 Mpumalanga Escarpment stonewalled settlements. The openly visible flame for Interestingly, a comparison of the Magaliesberg roasting meat seems not to have accompanied the with the Bokoni settlements (see Fig. 1)alongthe hearths that moved to the designated cooking huts. Mpumalanga Escarpment (e.g. Collett 1979; 1982; Comparison of two sites at Olifantspoort (Fig. 1) Delius & Schoeman 2008; 2010; Delius et al. 2012; Evers excavated by Revil Mason (1986) clearly indicates that 1975; Maggs 2008; Marker & Evers 1976) reveals that ideas about fire changed as interaction around hearths differing engagements with the surrounding material changed (Fredriksen 2012, 72–83). The intimate as- world generated two political situations where social sociation between open flames, maleness and ances- wealth was articulated differently through landscapes tors broke away and went in another direction: out of stone constructions. into the front courtyard and the chiefly court. Signif- The origin and complex dynamics of the sites is icantly, an understanding of this kind of microscale a matter of current debate (see Delius & Schoeman interaction should not be limited to living members 2008; Huffman 2004b; 2007, 41, 48, 433–6, 448; Maggs of the household. When people met around the fire- 2008; for research status and periodization, see Delius place, their respective ancestors and descent lines met et al. 2012). Huffman (2007, 41) classifies the sites’ type as well. The increased compartmentalization in the of stone walling as Badfontein (Fig. 2). This is a ver- stonewalled settlements contributed to a relatively sion of the CCP which emphasizes the centre/side lower degree of interaction between female and male axis, expressed through concentric circles: the inner sexual identities—and consequently less interaction for cattle, the next marked the men’s court, and the between their ancestors. There are relatively more el- outer was the area for houses. However, Tim Maggs ements in the pre-walling households that are com- (2008, 175–8) points out that, although the homestead parable to social dynamics in exogamous households layout broadly falls into the CCP and the material than is the case for post- ad 1700 Moloko dwelling culture and basic economy are typical of the LIA, the spaces. built settlement pattern is striking in comparison to all This evidence supports Simon Hall’s (1998) other known contemporary farming societies (Fig. 3). proposition that the specific emphasis on endogamy It actually turns out to be a contrast to what we would among the Tswana developed as the process towards expect from settlements associated with Moloko pot- aggregated settlements proceeded, indicating that mi- tery and Sotho/Tswana speakers. croscale changes were inextricably linked to more re- Fredriksen (2012, 105–9) has related Maggs’ gional economies of town living (see also Hall 2012). observations to an important difference between One of the arguments made by Lane (1998) was pre- the Bokoni settlements and the Magaliesberg stone cisely that the changing layout of houses from EIA to towns. In articulating with the landscape through LIA contexts was about a shift from a preference of stonewalling, the latter was largely about demarcat- exogamous to endogamous marriage patterns associ- ing wealth through cattle enclosures and homesteads, ated with increasing male control over female labour. leaving cultivated fields as something less clearly ar- Following Lane’s (2004b) argument that settlement ticulated that lay beyond settlement space. Here, cat- movement transformed spiritual and conceptual un- tle belonged inside the settlement while cultivation derstanding of landscape, Fredriksen (2012, 99–100) was on the outside. In Mpumalanga, on the other has argued that the agglomeration process led to an hand, stone articulations included the domain of agri- increased articulation with the landscape expressing culture instead of excluding it, thereby extending the 8 Beyond Static Models

Figure 2. Huffman’s version: Badfontein walling. (Redrawn after Loubser 1994; Huffman 2007.)

articulations beyond the dwelled-in spaces. The po- terracing underscored, through everyday experience tentially deep relations between the female body and and the visual presence of the cultivated landscape, specific fields may have been important for women’s a contextually specific value attributed to land and position within the general agricultural economy and agricultural produce as sources of wealth. Thus, by ecology (Guyer 1991), and the terraced fields thereby considering architecture as something that unfolds left female experience and labour clearly visible in through practices of making, we see that different the landscape. In other words, the emphasis on what engagements with the surrounding material world was viewed socially significant for the accumulation generate culturally specific meanings. Rather than of wealth, especially from a male perspective, must treating the process of building as one of impos- have differed considerably in the two cases. ing ready-made templates or static mental repre- Consequently, intimate, bodily experienced en- sentations onto the world, people, by thinking the gagements changed the landscape, and the associa- thoughts they do in a materially and socially specific tions to it became unique in the context of contem- dwelled-in world, create and recreate a distinctive porary LIA stone-building farming communities. The architectural environment. Such a perspective con- emphasis on intimacy between bodily experience and tributes to an understanding of why the Bokoni set- materiality is an important factor for understanding tlements with Badfontein walling were different from the exceptional Bokoni development. The extensive all other versions of the CCP: through the intersecting

9 Per Ditlef Fredriksen and Shadreck Chirikure

secluded activity that necessitated the location of smelting furnaces away from the public gaze (Childs 1991; Childs & Killick 1993; Schmidt & Mapunda 1997). This association is also clear from anthropo- morphic decorations such as female breasts, genitalia, navels and other sexual iconography that appeared on furnaces in Bantu Africa (Childs 1991; Chirikure & Rehren 2004; Collett 1993). This procreational paradigm has structured relations between smelters and non- smelters across much of southern Africa in the last 200 years. However, the question still remains as to the de- gree to which one may subsume ethnographic infor- mation about spatial organization of metal produc- tion into models such as the CCP and, following the direct historical approach, extend such models back- wards to the LIA and the EIA. In our opinion, evi- dence of variation in iron smelting spatiality in more recent times and in the Iron Age opens discussion of alternatives. Clearly, a survey of late nineteenth-century ethnography and oral histories indicates that several communities located their smelting away from vil- lages. Examples include the residents of Nyanga in northeastern Zimbabwe (Soper 2006) and both the BaHurutse of Kaditshwene (Campbell 1967)andthe Figure 3. Maggs’ version: sketch plan of a typical larger Venda (Stayt 1931) in northern South Africa. How- Bokoni homestead. (Redrawn after Maggs 2008.) ever, an archaeological study by Shadreck Chirikure and Thilo Rehren (2004) demonstrates variation be- tween Lowland and Upland Nyanga (see Fig. 1). In trajectories of people, animals, material culture, archi- Lowland Nyanga, most iron-smelting furnaces were tecture and landscape, a unique process of dwelling built inside stone enclosures and situated in what unfolded. appear to be hidden areas, implicating a concept of seclusion. One furnace was explicitly decorated with Theme II: spatial variability in metalworking and female genitalia, but was situated in a secluded space. the direct historical approach Interestingly, residents of Upland Nyanga also made decorated furnaces, except that these were found Metalworking offers examples of the effect of applied in low, walled smelting enclosures that were built analytical scale and the limiting nature of projections within the homestead (Chirikure & Rehren 2004). from ethnographic observation to various parts of the While it is possible that the low walls may have Iron Age following the direct historical approach. As represented some symbolic seclusion, the furnace it was still practised in the late nineteenth and early still figured prominently. Next to the furnace was twentieth centuries, metal technologies such as iron evidence from the production process: a large slag smelting lend themselves to combined ethnographic, pile, collapsed furnace remains and other remnant historical and archaeological scrutiny. materials. The technologies have become key to under- Significantly, while a version of procreational standing aspects of Iron Age societies on a regional symbolism was integral to the process in both areas, and inter-regional level, such as political interaction, the smelting inside the homestead in Upland Nyanga trade and warfare, but they also inform multi-scalar differed from the practice in lowland Nyanga. A dif- gender relations. For example, several late nineteenth- ferent production context may thus have prevailed century cases demonstrate that iron smelting often in Upland Nyanga: the variation in spatial organiza- was located outside villages, as the CCP model pre- tion of iron smelting within one area suggests that, dicts. The principal reason was the metaphoric link while observances of certain cultural principles were between iron reduction and human reproduction, a of importance, there was comparatively less regard

10 Beyond Static Models

to the kind of spatial seclusion of production pre- side/outside settlements was of comparatively less dicted by the CCP. This illustrates a crucial point. importance. On the one hand it may be concluded, in accordance Importantly, this interpretative framework de- with the CCP and on the basis of single examples, parts from the view established by the CCP and ZP, that iron smelting in Lowland Nyanga did indeed which argue that smelting invariably took place out- take place in seclusion, away from residential areas. side settlements. In our opinion, the latter approach On the other hand, however, when discussing mul- sufficiently accommodates neither variation between tiple cases on a comparative level, the conclusion is production contexts nor the fact that cultural princi- that there are indeed variations in spatiality to be ples may have been articulated or respected regardless found. of outside/inside location. This ethnographic vari- Another important nineteenth-century example ation also poses significant questions regarding the is the spatiality of iron smelting among the Njanja spatial organization of smelting in the deeper past (Mackenzie 1975), depicted in historical and ethno- (Chirikure 2007). A consideration of examples relat- graphic sources as specialists who made iron for a ing to well-known sites such as Great Zimbabwe considerable market in modern-day Zimbabwe and (ad 1290–1550), Mapungubwe (ad 1220–1290), K2 (ad Mozambique. Chirikure spent six months in the 1000–1220) and Shroda (ad 900–1000) demonstrates Njanja area collecting oral information and ethno- how richly varied the spatial organization of produc- graphic details from various Njanja smelting houses, tion was in Iron Age southern Africa. For example, focusing in particular on how innovations in the orga- Eugenia Herbert (1996) argued that control of metal nization of production affected the spatiality of iron production was an important source of power at Great smelting (Chirikure 2005; 2006). In order to satisfy in- Zimbabwe, to the extent that primary smelting took creasing demand, the Njanja reorganized the produc- place at the site. The argument was in part based on tion by simultaneously operating multiple furnaces, the recorded presence of furnaces and tuyeres by early using a shift system of labour (Mackenzie 1975). Be- excavators. Herbert’s thesis has been criticized for not cause of increased labour needed, the Njanja called being grounded in sufficient tangible evidence, com- upon women to pump the bellows and to sing, to bined with the argument that smelting would not provide entertainment in addition to helping bellows have taken place at Great Zimbabwe because of the operators to maintain their rhythm (see also Dewey need for spatial seclusion, as predicted by the ZP 1991). For these reasons, Njanja furnaces were situ- model. Interestingly, a recent study of material recov- ated where the source of labour was: within the vil- ered during salvage excavations and wall restoration lage or homestead. Njanja furnaces were decorated at the site encountered tuyeres with run-back slag and with anthropomorphic features such as breasts, while various other slag types. The new evidence from the medicines were used to neutralize the power of malev- Northern Rock Shelter on the Hill Complex at Great olent forces (Dewey 1991). Thus, as in the Upland Zimbabwe indicates that some primary smelting was Nyanga example, the location of smelting inside the taking place, thus providing support to Herbert’s pro- village/homestead did not reduce the level of obser- posal (Chirikure 2014). It is therefore possible that vance of cultural principles linked to the reduction the rulers were symbolic metalworkers, enabling con- process (see e.g. Chirikure 2005 and Hatton 1967 for trol over a powerful productive element of society, or further ethnographic examples of variation in spatial that they were elites who produced their own metal organization of metallurgy). (Chirikure 2007). This inside/outside complexity is also sup- Also, John Calabrese (2007) has concluded that ported by examples outside metallurgy. Chirikure primary metal production was a feature of the ear- (2005) collated nineteenth-century European travel- lier Mapungubwe site. His collation of evidence in- ogues that depict local houses, drums, head rests cludes analyses by metallurgists, who confirm the and other material culture used within Shona house- presence of iron working, including smelting, on holds. Interestingly, some of the houses, headrests northwestern Mapungubwe Hill. Interestingly, some and drums were decorated with female anatomical of the areas with evidence of metallurgy at Mapun- parts, such as breasts, and other iconography asso- gubwe and Great Zimbabwe are, according to the ZP ciated with fertility. Designs on drums and houses model, believed to have been royal wives’ areas. Pre- could be observed from a distance, thus demonstrat- viously, the difficulty of chemically and visually sepa- ing that procreational symbolism was indeed inte- rating smelting from smithing slags, and backed by se- gral to the worldview. Again, as shown in the Up- lected ethnographic evidence, persuaded researchers land Nyanga and Njanja cases, it seems as long as to argue that it was smithing and not smelting that cultural principles were adhered to the location in- took place inside Iron Age homesteads and villages. 11 Per Ditlef Fredriksen and Shadreck Chirikure

However, more recent studies of slags show that a ing medium and solid structures made of daga (a number of variables separate smelting from smithing thick earthen/clay paste) are essential characteristic of practices. These include the presence of tuyeres with Great Zimbabwe and related sites (Chirikure & Piki- run-back slag, partially reduced ore and slag with a rayi 2008; Chirikure et al. 2013b). It is believed that the clearly defined flow structure (Chirikure & Rehren Zimbabwe tradition crystallized an ideology of class 2006). On this basis, it may be argued that variation in distinction and sacred leadership that physically and the spatiality of iron smelting in the Zimbabwe tradi- ideologically separated rulers from the ruled (Huff- tion was mediated by a wide array of factors, ranging man 2007; Pikirayi 2001). As such, while the elites´ from resource availability to more culturally specific lived inside walls, the commoners lived outside. Gen- ones. This variation within the tradition underscores erally, most dry stone walls were sited on raised areas the need for developing alternative approaches. If we e.g hilltops, rises etc. The process of wall construc- look to East Africa, for example, Peter Schmidt (1997) tion incorporated nature in the form of boulders to has argued that iron smelting took place in areas con- create walls, which resulted in the appropriation of tiguous to settlements with interesting levels of vari- nature to produce the built environment. In the Shona ation throughout the Iron Age. world, ‘nature’ has been the domain of ancestors who Evidence for spatial variation is not limited to controlled the fertility of the land, thereby enabling the Late and Middle Iron Ages, as discussed thus society to reproduce itself. Most walls at Khami and far. Several examples are found from the EIA (e.g. Great Zimbabwe incorporate various elements of the Maggs 1982). Chirikure and Rehren (2006) discuss natural world such as boulders. According to Pikirayi the presence of iron smelting within the EIA site (2013), the construction of monumental architecture of Swart Village in northern Zimbabwe. Here, finds in the Zimbabwe tradition was therefore a process of of partially reduced ore, blocks of flow slag, tuy- constructing social and political power through the eres with run-back slag and collapsed furnace wall manipulation of ideology, including appropriation of were recovered within the centre of the village. Also, elements of ‘nature’. Maggs (1982) made the same observation in the Two architectural styles of the Zimbabwe tradi- KwaZulu-Natal region of South Africa, where smelt- tion represent distinct conceptual relationships to na- ing furnaces were located in the middle of settle- ture: on the one hand the freestanding walls typical ments which were otherwise conforming to CCP of Great Zimbabwe and related settlements, and on principles. the other the retaining walls typical of Khami. Great We emphasize that the presence of smelting in- Zimbabwe style architecture was apparently an ex- side settlements, when compared to sites where smelt- tension of the natural environment, while Khami (ad ing did take place outside, does not necessarily sug- 1400–1800) architecture exploited elements of nature, gest a different belief system. Spatial variation merely thereby transforming them into monumental built en- suggests that, throughout the Iron Age, there was vironments. At most sites of both types the walls were marked temporal, cultural and geographic difference built and rebuilt over time. At Khami, Keith Robinson in solutions to the issue of locating smelting activi- (1959) demonstrated that some of the platforms in the ties in relation to dwelled-in spaces. When moving interior of the Hill Complex had been extended over between multiple examples at different scales, greater time. Similarly, the study of the walls in the valley variability appears than what may be captured by the enclosures at Great Zimbabwe by David Collett et al. use of the ZP and CCP models. (1992) revealed that there was building and rebuilding over time. It may therefore be argued that the use of Theme III: timescales and recursiveness in the different spaces was in constant flux as the dynamics Zimbabwe tradition of daily life continued and changed. Importantly, there is a strong possibility that This sensitivity towards spatial variation may also power may have shifted within the individual set- be combined with an increased awareness of re- tlements in keeping with the well-established prin- fined timescales and recursiveness between the hu- ciple of rotational succession (Chirikure & Pikirayi man mind and the material world, as our third and 2008). It is established that from ad 1450 to 1900, the final theme discussion demonstrates. average rule of kings of the Mutapa state which di- Pikirayi (2013) has recently discussed the recur- rectly succeeded Great Zimbabwe was not more than sive relationship between the symbolism of dry stone 30 years (Chirikure et al. 2012). While power shifted to walls of the Zimbabwe tradition and the socio-cultural other places, spaces were modified to accommodate and ideological processes behind their production. changing circumstances. A key challenge is that Dry stone walls built without the aid of any bind- power and leadership changes are often difficult to 12 Beyond Static Models

date using available absolute dating methods such as Importantly, we wish to emphasize that to crit- radiocarbon dating, which produce aggregate times icize the theoretical and analytical foundation of the that subsume events of different duration (Chirikure cognitive models is not the same as saying they are in- et al. 2012; 2013b). It is therefore important to develop adequate because they are designed to do something models or approaches that make it easier to identify else (cf. Huffman 2001, 24). Rather, our main point can such subtleties through a recursive consideration of be summarized as a rhetorical question: if a cognitive both macro- and microscale events. model does not take into account the aspects of past social life one wishes to engage with, then why do Conclusions: towards a recursive future alternative approaches have to be compared to such a model at all? Are the structuralist cognitive models needed for fu- By engaging the epistemic foundation of the ture research and understanding of Iron Age social modern view of science and by identifying certain life in southern Africa? Remaining largely unchanged root causes for the epistemic standstill, we hope for three decades, we argue that the combination of to contribute to a constructive and fruitful future modifying the CCP and ZP models to work at context- debate. However, in order for advances to be made specific levels with alternative approaches may enable on all levels, including theory, the basis of the us to progress beyond present status, by acknowledg- cognitive models should be reviewed in detail. In ing intimate recursiveness and dialectics three-fold: particular, this goes for the scientific retreat from present/past, micro/macro scales and mind/matter. society aided by dichotomizing classification into In order to overcome the epistemic stasis, a theo- simplified opposites: past from present, aim from retical shift is necessary: from an emphasis on cog- practice, processualists from postprocessualists, and nition favouring abstract space on relatively fixed science from society. In our opinion, recursiveness in macroscale levels to a recognition of the importance theory as well as in daily practice is the way of the of localized, context-specific insights which takes mi- future for Iron Age archaeology in southern Africa, croscale variation into serious consideration. A major regardless of whether one prefers the label processual challenge for the CCP and ZP models is, therefore, not or postprocessual, or neither. that they are macroscale models per se, but that they This way we can work towards further recon- only operate on the macroscale and are unable to ex- struction of women’s role in society, intra-site and plain microscale variation on homestead or household inter-cultural variation, economic specialization and levels. Rather, they must be used in ways that recog- organization of production, as well as the human nizes recursiveness between various scales to account responses and contributions to climatic change. All for continuity, change and variability to be measured. these issues have implications for settlement organi- For example, the CCP and the ZP may acknowledge zation and how human beings interacted with their that fireplaces changed positions and that smelting surroundings. As shown here, responses were of- took place within settlements just as domestic use ten different, even if the worldview may have been of space was in continuous flux. A recent contribu- similar. The future should hold multi-faceted and tion to the historical archaeology of the Mapungubwe multi-scalar studies that combine continued cross- area (Huffman 2012b) provides an interesting depar- disciplinary efforts towards ever more sophisticated ture point for discussion. By combining a wide vari- archaeological methodology with critical epistemic ety of sources into an impressively detailed archae- reflection that acknowledges contextually specific ology of the recent past, the paper concludes with African experiences. This, in our view, will produce reflections on the question of scale and archaeological archaeological insights with potential to contribute to sequencing. debates in global archaeology. An added dimension to the entrenched nature of the Iron Age discourse is the somewhat disconcerting Notes misrepresentation of critics’ arguments—not only of individual researchers but also of the entire so-called 1. Our epistemic departure point therefore differs in sig- Post-Processual School—which may serve to derail nificant respects from the criticism voiced by propo- academic exchange. The uses of broad, dichotomizing nents of more ‘historicized pasts’ recently responded categorizations construct an illusory rivalry and strive to by Gavin Whitelaw (2012, with references) as deriv- ing from a historical particularism. for hegemony between ‘competing’ models. Instead 2. Now called the Association of Southern African Profes- of a hegemonic Iron Age archaeology, we hope our sional Archaeologists (ASAPA). arguments will benefit a future set of archaeologies 3. Historian David Beach (1980) pointed out at an early where plurality and symmetry are keywords. stage that, despite claims of objectivity, the number of 13 Per Ditlef Fredriksen and Shadreck Chirikure

ceramic traditions in southern Africa approximates the Childs, S.T., 1991. Style, Technology and iron smelting fur- number of practising archaeologists, thereby demon- naces in Bantu-speaking Africa. Journal of Anthropo- strating the downside of a heavy reliance on typology. logical Archaeology 10, 332–59. Childs, S.T. & D. Killick, 1993. Indigenous African metal- Per Ditlef Fredriksen lurgy: nature and culture. Annual Review of Anthropol- Department of Archaeology, Conservation and History ogy 22, 317–37. University of Oslo, P.O. Box 1019 Chirikure, S., 2005. Iron Production in Iron Age Zimbabwe: N-0315 Oslo Stagnation or Innovation. Unpublished PhD disserta- tion, University College London. Norway Chirikure, S., 2006. New light on Njanja iron working: to- & Department of Archaeology wards a systematic encounter between ethnohistory University of Cape Town and archaeometallurgy. South African Archaeological Private Bag, Rondebosch 7701 Bulletin 61(184), 142–51. South Africa Chirikure, S., 2007. Metals in society. Iron production and its Email: [email protected] position in Iron Age communities of southern Africa. Journal of Social Archaeology 7(1), 72–100. Shadreck Chirikure Chirikure, S., 2014. Metals in Society: Indigenous Mining Department of Archaeology and Metalworking in a Global Perspective.NewYork: University of Cape Town Springer. Chirikure, S., S. Hall & T. Rehren. In press. When material Private Bag, Rondebosch 7701 science and ceramic sociology meet: an archaeomet- South Africa allurgical study of crucibles from Mapungubwe Hill, Email: [email protected] South Africa. Chirikure, S, M. Manyanga & A.M. Pollard, 2012. When science alone is not enough: radiocarbon timescales, References history, ethnography and elite settlements in southern Africa. Journal of Social Archaeology 12(3), 356–79. Andah, B.W., 1995. Studying African societies in cultural Chirikure, S., M. Manyanga & F. Bandama, 2013a. context, in Making Alternative Histories. The Practice of A Bayesian chronology of Great Zimbabwe: re- Archaeology and History in Non-Western Settings,eds. threading the sequence of a vandalized monument. P.R. Schmidt & T.C. Patterson. Santa Fe (NM): School Antiquity 87, 854–72. of American Research Press, 149–81. Chirikure, S., M. Manyanga, I. Pikirayi & M. Pollard, Anderson, M.S., 2009. Marothodi. The Historical Archaeology 2013b. New pathways of sociopolitical complexity of an African Capital. Woodford: Atikkam. in Southern Africa. African Archaeological Review 30, Beach, D.N., 1980. The Shona and Zimbabwe 900–1850. An 339–66. Outline of Shona History. London: Heinemann. Chirikure, S. & I. Pikirayi, 2008. Inside and outside the dry Beach, D.N., 1998. Cognitive archaeology and Imaginary stone walls: revisiting the material culture of Great history at Great Zimbabwe. Current Anthropology Zimbabwe. Antiquity 82, 976–93. 39(1), 47–72. Chirikure, S. & G. Pwiti, 2008. Community involvement in Boeyens, J., 2003. The Late Iron Age sequence in the Marico Archaeology and cultural heritage management. Cur- and early Tswana history. South African Archaeological rent Anthropology 49(3), 467–85. Bulletin 58(178), 63–78. Chirikure, S. & T. Rehren, 2004. Ores, slags and furnaces: as- Bonner, P.L., A.B. Esterhuysen, M.H. Schoeman, pects of iron working in the Nyanga complex. African N.J. Swanepoel & J.B. Wright, 2008. Introduc- Archaeological Review 21, 135–52. tion, in Five Hundred Years Rediscovered: Southern Chirikure, S. & T. Rehren, 2006. Iron smelting in pre- African Precedents and Prospects, eds. N. Swanepoel, colonial Zimbabwe: evidence for diachronic change A.B. Esterhuysen & P. Bonner. Johannesburg: Wits from Swart Village and Baranda, northern Zimbabwe. University Press, 1–19. Journal of African Archaeology 4, 37–54. Bourdieu, P., 1977. Outline of a Theory of Practice. Cambridge: Collett, D.P.,1979. The Archaeology of the Stone-Walled Set- Cambridge University Press. tlements in the Eastern Transvaal, South Africa. Un- Calabrese, J.A., 2007. The Emergence of Social and Political published MSc dissertation, University of the Witwa- Complexity in the Shashi-Limpopo Valley of Southern tersrand. Africa, AD 900 to 1300. Ethnicity, Class and Polity.(Cam- Collett, D.P., 1982. Excavations of stone-walled ruin types bridge Monographs in African Archaeology 69.) Ox- in the Badfontein Valley, Eastern Transvaal, South ford: Archaeopress. Africa. South African Archaeological Bulletin 37(135), Campbell, J., 1967 (1822). Travels in South Africa, Undertaken 34–43. at the Request of the London Missionary Society being a Collett, D.P., 1993. Metaphors and representations associ- Narrative of a Second Journey in the Interior of that Coun- ated with Pre-Colonial iron smelting in Eastern and try. New York/London: Johnson Reprint Corporation. Southern Africa, in The Archaeology of Africa: Food,

14 Beyond Static Models

Metal and Towns,eds.T.Shaw,P.Sinclair,B.Andah Gonzalez-Ruibal, A., 2006a. The past is tomorrow. Towards & J. Okpoko. London: Routledge, 499–511. an archaeology of the vanishing present. Norwegian Collett, D.P., A.E. Vines & E.G. Hughes, 1992. The chronol- Archaeological Review 39(2), 110–25. ogy of the Valley Enclosures: Implications for the Gonzalez-Ruibal, A., 2006b. The dream of reason. An ar- interpretation of Great Zimbabwe. African Archaeolog- chaeology of the failures of modernity in Ethiopia. ical Review 10(1), 139–61. Journal of Social Archaeology 6(2), 175–201. Davison, P.,1988. The social use of space in a Mpondo home- Gonzalez-Ruibal, A., 2009. Vernacular cosmopolitanism: An stead. South African Archaeological Bulletin 43(148), archaeological critique of universalistic reason, in Cos- 100–108. mopolitan Archaeologies, ed. L. Meskell. Durham (NC): Delius, P. & M.H. Schoeman, 2008. Revisiting Bokoni: popu- Duke University Press, 113–39. lating the stone ruins of the Mpumalanga Escarpment, Green, L.J.F., 2008. ‘Indigenous knowledge’ and ‘science’: in Five Hundred Years Rediscovered: Southern African Reframing the debate on knowledge diversity. Archae- Precedents and Prospects,eds.N.Swanepoel,A.B.Es- ologies 4(1), 144–63. terhuysen & P.Bonner. Johannesburg: Wits University Gupta, A. & J. Ferguson, 1997. Beyond ‘Culture’: space, Press, 135–67. identity and the politics of difference, in Culture, Delius, P. & M.H. Schoeman, 2010. Reading the rocks and Power, Place. Explorations in Critical Anthropology,eds. reviewing red herrings. African Studies 69(2), 235–54. A. Gupta & J. Ferguson. Durham (NC): Duke Univer- Delius, P., T.M. Maggs & M.H. Schoeman, 2012. Bokoni: Old sity Press, 33–51. structures, new paradigms? Rethinking pre-colonial Guyer, J.I., 1991. Female farming in anthropology and society from the perspective of the stone-walled sites African history, in Gender and the Crossroads of Knowl- in Mpumalanga. Journal of Southern African Studies edge: Feminist Anthropology in the Postmodern Era,ed. 38(2), 399–414. M. di Leonardo. Berkeley (CA): University of Califor- Dewey, W.J., 1991. Pleasing the Ancestors: The Traditional nia Press, 257–77. Art of the Shona People of Zimbabwe. Unpublished Hall, M., 1984a. The burden of tribalism: The social context PhD thesis, University of Indiana. of Southern African Iron Age ceramic studies. Ameri- Evers, T.M., 1975. Recent Iron Age research in the Eastern can Antiquity 49(3), 455–67. Transvaal, South Africa. South African Archaeological Hall, M., 1984b. The myth of the Zulu homestead: Archae- Bulletin 30(119/120), 71–83. ology and ethnography. Africa 54(1), 65–79. Evers, T.M., 1983. ‘Oori’ or ‘Moloko’? The origins of the Hall, M., 1984c. Pots and politics: Ceramic interpretations in Sotho-Tswana on the evidence of the Iron Age of the Southern Africa. World Archaeology 15(3), 262–73. Transvaal: reply to R.J. Mason. South African Journal of Hall, M., 1986. The role of cattle in southern African agropas- Science 79, 261–4. toral societies: more than bones alone can tell, in Pre- Evers, T.M., 1988. The Recognition of Groups in the Iron Age historic Pastoralism in Southern Africa,eds.M.Hall& of Southern Africa. Unpublished PhD thesis, Univer- A.B. Smith. (South African Society Goodwin Series sity of the Witwatersrand. 5.) Cape Town: South African Archaeological Society, Fauvelle-Aymar, F-X & K. Sadr, 2008. Trends and traps in the 83–7. reconstruction of early herding societies in southern Hall, M., 2005. Situational ethics and engaged practice: The Africa. Southern African Humanities 20, 1–6. case of archaeology in Africa, in Embedding Ethics,ed. Fewster, K.J., 2006. The potential of analogies in post- L. Meskell. Oxford: Berg, 169–94. processual archaeologies: a case study from Basimane Hall, S.L., 1995. Archaeological indicators of stress in the Ward, Serowe, Botswana. The Journal of the Royal An- Western Transvaal Region between the seventeenth thropological Institute 12(1), 61–87. and the nineteenth centuries, in The Mfecane After- Fredriksen, P.D., 2007. Approaching intimacy: interpreta- math. Reconstructive Debates in Southern African History, tions of changes in Moloko household space. South ed. C. Hamilton. Johannesburg/Pietermaritzburg: African Archaeological Bulletin 62(186), 126–39. Wits University Press/Natal University Press, Fredriksen, P.D., 2011. When knowledges meet: engage- 307–21. ments with clay and soil in southern Africa. Journal Hall, S.L., 1998. A consideration of gender relations in the of Social Archaeology 11(3), 283–310. Late Iron Age ‘Sotho’ sequence of the Western High- Fredriksen, P.D., 2012. Material Knowledges, Thermodynamic veld, South Africa, in Gender in African Prehistory,ed. Spaces and the Moloko Sequence of the Late Iron Age S. Kent. Walnut Creek (CA): AltaMira Press, 235–58. (AD 1300–1840) in Southern Africa. (Cambridge Mono- Hall, S.L., 2000. Forager lithics and Early Moloko home- graphs in African Archaeology 80.) Oxford: Archaeo- steads at Madikwe. Natal Museum Journal of Humani- press. ties 12, 33–50. Giblin, J., R. King & B. Smith, 2014. Introduction: de-centring Hall, S.L., 2007. Tswana history in the Bankenveld, in A ethical assumptions by re-centring ethical debate in Search for Origins. Science, History and South Africa’s African archaeology. Azania 49(2), 131–5. ‘Cradle of Humankind’,eds.P.Bonner,A.M.Esterhuy- Giddens, A., 1984. The Constitution of Society. Berkeley (CA): sen & T. Jenkins. Johannesburg: Wits University Press, University of California Press. 162–79.

15 Per Ditlef Fredriksen and Shadreck Chirikure

Hall, S.L., 2012. Identity and political centralisation in the Huffman, T.N., 2012b. Historical archaeology of the Mapun- Western Regions of Highveld, c. 1770–c. 1830: an ar- gubwe area: Boer, Birwa, Sotho-Tswana and Machete. chaeological perspective. Journal of Southern African Southern African Humanities 24, 33–59. Studies 38(2), 301–18. Huffman, T.N., 2014. Ritual space in the Zimbabwe Culture. Hall, S.L., M.S. Anderson, J. Boeyens & F. Coetzee, 2008. Ethnoarchaeology 6(1), 4–39. Towards an outline of the oral geography, historical Ingold, T., 2000. The Perception of the Environment. Essays in identity and political economy of the late precolo- Livelihood, Dwelling and Skill. London: Routledge. nial Tswana in the Rustenburg region, in Five Hun- Karega-Munene, K. & P.R. Schmidt, 2010. Postcolonial ar- dred Years Rediscovered: Southern African Precedents and chaeologies in Africa: Breaking the silence. African Prospects,eds.N.Swanepoel,A.M.Esterhuysen& Archaeological Review 27(4), 323–37. P. Bonner. Johannesburg: Wits University Press, 55– Kuper, A., 1980. Symbolic dimensions of the Southern Bantu 85. homestead. Africa 50(1), 8–23. Hatton, J.S., 1967. Notes on Makalanga iron smelting. NADA Kuper, A., 1982. Wives for Cattle: Bridewealth and Marriage in 9(4), 39–42. Southern Africa. London: Routledge & Kegan Paul. Herbert, E.W., 1996. Metals and power at Great Zimbabwe, Kuper, A., 2003. The return of the native. Current Anthropol- in Aspects of African Archaeology: Papers from the 10th ogy 44(3), 389–402. Congress of the PanAfrican Association for Prehistory and Kusimba, C.M. & S.B. Kusimba, 2005. Mosaics and interac- Related Studies, eds. G. Pwiti & R.C. Soper. Harare: tions: East Africa, 2000 b.p. to the present, in African University of Zimbabwe, 641–7. Archaeology. A Critical Introduction,ed.A.B.Stahl.Ox- Huffman, T.N., 1982. Archaeology and ethnohistory of the ford: Blackwell, 392–419. African Iron Age. Annual Review of Anthropology 11, Lane, P.J., 1994/95. The use and abuse of ethnography in the 133–50. study of the Southern African Iron Age. Azania 29–30, Huffman, T.N., 1986. Cognitive studies of the Iron Age in 51–64. Southern Africa. World Archaeology 18(1), 84–95. Lane, P.J., 1998. Engendered spaces and bodily practices in Huffman, T.N., 1996. Archaeological evidence for climatic the Iron Age of Southern Africa, in Gender in African change during the last 2000 years in southern Africa. Prehistory, ed. S. Kent. Walnut Creek (CA): AltaMira Quaternary International 33, 55–60. Press, 173–203. Huffman, T.N., 2001. The Central Cattle Pattern and in- Lane, P.J., 2004a. The ‘moving frontier’ and the transition to terpreting the past. Southern African Humanities 13, food production in Kenya. Azania 39(1), 243–64. 19–35. Lane, P.J., 2004b. Re-constructing Tswana townscapes: to- Huffman, T.N., 2002. Regionality in the Iron Age: the case ward a critical historical archaeology, in African His- of the Sotho-Tswana. Southern African Humanities 14, torical Archaeologies, eds. A.M. Reid & P.J. Lane. 1–22. New York (NY): Kluwer Academic & Plenum Press, Huffman, T.N., 2004a. Beyond data: The aim and practice 269–99. of archaeology. South African Archaeological Bulletin Lane, P.J., 2005. Barbarous tribes and unrewarding gyra- 59(180), 66–9. tions? The changing role of ethnographic imagination Huffman, T.N., 2004b. The archaeology of the Nguni past. in African archaeology, in African Archaeology. A Criti- Southern African Humanities 16, 79–111. cal Introduction, ed. A.B. Stahl. Oxford: Blackwell, 24– Huffman, T.N., 2007. Handbook to the Iron Age. The Ar- 54. chaeology of Pre-Colonial Farming Societies in South- Lane, P.J., 2010. Developing landscape historical ecologies ern Africa. Scottsville: University of KwaZulu-Natal in Eastern Africa: An outline of current research and Press. potential future directions. African Studies 69(2), 299– Huffman, T.N., 2008. Climatic change during the Iron Age 322. in the Shashe-Limpopo Basin, southern Africa. Journal Lane, P.J., 2011. Possibilities for a postcolonial archaeology of Archaeological Science 35, 2032–47. in sub-Saharan Africa: indigenous and usable pasts. Huffman, T.N., 2009. Mapungubwe and Great Zimbabwe: World Archaeology 43(1), 7–25. the origin and spread of social complexity in southern LaViolette, A., 2008. Swahili cosmopolitanism in Africa and Africa. Journal of Anthropological Archaeology 28(1), 37– the Indian Ocean World, A.D. 600–1500. Archaeologies 54. 4(1), 24–49. Huffman, T.N., 2010. Debating the Central Cattle Pattern: a Loubser, J.N., 1994. Ndebele archaeology of the Pieters- reply to Badenhorst. South African Archaeological Bul- burg area. Navorsinge van die Nasionale Museum, Bloem- letin 65(192), 164–74. fontein 10(2), 61–147. Huffman, T.N., 2011. Debating Great Zimbabwe. Lyons, D., 2006. A critical appraisal of gender research in South African Archaeological Bulletin 66(193), African Archaeology, in Handbook of Gender in Archae- 27–40. ology, ed. S.M. Nelson. Oxford: AltaMira Press, 595– Huffman, T.N., 2012a. Ritual space in pre-colonial farm- 631. ing communities in Southern Africa. Ethnoarchaeology Mackenzie, J.M., 1975. A pre-colonial industry: the Njanja 4(2), 119–46. and the iron trade. NADA 11(2), 200–220.

16 Beyond Static Models

Maggs, T.M., 1976. Iron Age Communities of the Southern High- Schmidt, P. R., 1997. Iron Technology in East Africa: Symbol- veld. (Occasional Publications of the Natal Museum 2.) ism, Science and Archaeology. Bloomington (IN): Indi- Pietermaritzburg: Natal Museum. ana University Press. Maggs, T.M., 1982. Mgoduyanuka: Terminal Iron Age set- Schmidt, P.R. (ed.), 2009. Postcolonial Archaeologies in Africa. tlement in the Natal grasslands. Annals of the Natal Santa Fe (NM): School for Advanced Research Press. Museum 25(1), 83–113. Schmidt, P.R. & Mapunda, B, 1997. Ideology and the ar- Maggs, T.M., 1993. Three decades of Iron Age research chaeological record in Africa: Interpreting symbolism in southern Africa: some personal reflections. South in iron smelting technology. Journal of Anthropological African Archaeological Bulletin 48(158), 70–6. Archaeology 16, 73–102. Maggs, T.M., 2008. The Mpumalanga Escarpment settle- Schnapp, J., M. Shanks & M. Tiews, 2004. Archaeology,mod- ments: some answers, many questions, in Five Hun- ernism, modernity. Editors’ introduction to ‘Archae- dred Years Rediscovered: Southern African Precedents and ologies of the modern’. Modernism/Modernity 11(1), 1– Prospects, eds. N. Swanepoel, A.M. Esterhuysen & 16. P. Bonner, P. Johannesburg: Wits University Press, Schoeman, M.H., 1998a. Excavating Ndzundza Ndebele 169–81. identity at KwaMaza. Southern African Field Archae- Marker, M.E. & T.M. Evers, 1976. Iron Age Settlement and ology 7(1), 42–52. Soil Erosion in the Eastern Transvaal, South Africa. Schoeman, M.H., 1998b. Material culture ‘under the animal South African Archaeological Bulletin 31 (123/124), 153– skin’: excavations at Esikhunjini, an Mfecane period 65. Ndzundza Ndebele site. Southern African Field Archae- Marks, S., 2011. New Paradigms in History and Archaeology ology 7(1), 72–81. in South Africa. African Studies 70(1), 123–43. Schoeman, M.H., 2006. Imagining rain-places: rain-control Mason, R., 1986. Origins of Black People of Johannesburg and the and changing ritual landscapes in the Shashe- Southern Western Central Transvaal AD 350–1880. (Oc- Limpopo Confluence Area, South Africa. South casional Publication 16.) Johannesburg: Archaeologi- African Archaeological Bulletin 61(184), 152–65. cal Research Unit, University of the Witwatersrand. Segobye, A.K., 1998. Daughters of cattle: the significance of Meskell, L., 2005. Archaeological ethnography: Conversa- herding in the growth of complex societies in Southern tions around Kruger Par. Archaeologies 1(1), 81–100. Africa between the 10th and the 15th centuries AD, in Meskell, L., 2009. Introduction: Cosmopolitan heritage Gender in African Prehistory, ed. S. Kent. Walnut Creek ethics, in Cosmopolitan Archaeologies, ed. L. Meskell. (CA): AltaMira Press, 227–33. Durham (NC): Duke University Press, 1–27. Segobye, A.K., 2005. Weaving fragments of the past for Pikirayi, I., 2001. The Zimbabwe Culture: Origins and Decline a united Africa: reflections on the place of African of Southern Zambezian States. Walnut Creek (CA): Al- archaeology in the development of the continent in taMira Press. the 21st century. South African Archaeological Bulletin Pikirayi, I., 2007. Ceramics and group identity. Towards a 60(182), 79–83. social archaeology in southern African Iron Age ceramic Shepherd, N., 2002. The politics of archaeology in Africa. studies. Journal of Social Archaeology 7(3), 286–301. Annual Review of Anthropology 31 189–209. Pikirayi, I., 2013. Stone architecture and the development Shepherd, N., 2003. State of the discipline: Science, cul- of power in the Zimbabwe tradition AD 1270–1830. ture and identity in South African archaeology, 1870– Azania 48(2), 282–300. 2003. Journal of Southern African Studies 29(4), 823– Pikirayi, I. & S. Chirikure, 2011. Debating Great Zimbabwe. 44. Azania 46(2), 221–31. Smith, B., 2014. Ethics in African archaeology. Azania 49(2), Reid, A.M., 2004. Access to cattle in a Tswana capital, 136–47. in African Historical Archaeologies,eds.A.M.Reid& Soper, R., 2006. The Terrace Builders of Nyanga. Harare: P.J. Lane. New York (NY): Kluwer Academic/Plenum Weaver Press. Press, 301–24. Stahl, A.B., 2004. Political economic mosaics: Archaeology of Renfrew, C., 1982. Towards an Archaeology of Mind.Cam- the last two millennia in tropical Sub-Saharan Africa. bridge: Cambridge University Press. Annual Review of Anthropology 33, 145–72. Robinson, K.S.R., 1959. Khami Ruins. Cambridge: Cambridge Stahl, A.B., 2010. Book review: Schmidt P.R. (ed.), ‘Postcolo- University Press. nial Archaeologies in Africa’. African Archaeological Re- Sadr, K., 2012. The origins and spread of dry laid, stone- view 27(2), 165–8. walled architecture in Pre-Colonial Southern Africa. Stayt, H., 1931. The Bavenda. London: Oxford University Journal of Southern African Studies 38(2), 257–63. Press. Sadr, K. & X. Rodier, 2012. Google Earth, GIS and stone Webmoor, T. & C.L. Witmore, 2008. Things are us! A com- walled structures in Southern Gauteng, South Africa. mentary on human/things relations under the banner Journal of Archaeological Science 39(4), 1034–42. of a ‘social’ archaeology. Norwegian Archaeological Re- Schlanger, N., 2004. The past is in the present: on the histo- view 41(1), 53–70. ries and archives of archaeology. Modernism/Modernity Weiss, L., 2007. Heritage-making and political identity. Jour- 11(1), 165–7. nalofSocialArchaeology7(3), 413–31.

17 Per Ditlef Fredriksen and Shadreck Chirikure

Whitelaw, G., 2012. Anthropology and history in the South- of Oslo, Norway, and Research Associate at Department ern African Iron Age. African Studies 71(1), 127–44. of Archaeology, University of Cape Town, South Africa. Whitelaw, G., 2013. Pollution concepts and marriage for the His interests include the use of theory and development Southern African Iron Age. Cambridge Archaeological of methodology in archaeology. A main focus is on ce- Journal 23(2), 203–25. ramic technology and prehistoric and present uses of clay Widgren, M., 2000. Islands of intensive agriculture in in southern Africa, where he also conducts ethnographic African drylands: towards an explanatory frame- fieldwork. work, in The Archaeology of Drylands: Living at the Mar- gin, ed. G. Barker & D. Gilbertson. London: Routledge, 252–67. Shadreck Chirikure is Associate Professor at Department Witmore, C.L., 2006. Archaeology and modernity – or ar- of Archaeology, University of Cape Town, South Africa. chaeology and a modernist amnesia? Comments on His research deals with the interface between the hard Julian Thomas (2004) ‘Archaeology and Modernity’. and human sciences. He mainly studies material culture, Norwegian Archaeological Review 39(1), 49–52. focusing on the technology and anthropology of pre- industrial technologies such as metal and pottery produc- Author biographies tion. Recently, he has broadened his interests to include the application of Bayesian models in developing chrono- Per Ditlef Fredriksen is Associate Professor at Depart- logical models. He also publishes on heritage and risk ment of Archaeology, Conservation and History, University management.

18