A pilot investigation into forager craft activities in the middle Limpopo Valley, southern Africa

Forager lifeways in the middle Limpopo Valley, southern Africa, were considerably altered from 350 CE onwards when incoming farmer communities settled the region. This is seen archaeologically in a shift in the preference for specific Later Stone Age
of 14
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Contents lists available at ScienceDirect Journal of Archaeological Science: Reports  journal homepage: A pilot investigation into forager craft activities in the middle LimpopoValley, southern Africa Tim Forssman a, ⁎ , Trent Seiler b , David Witelson a a  Rock Art Research Institute, School of Geography, Archaeology and Environmental Studies, University of the Witwatersrand, PO Wits 2050, Johannesburg, South Africa b  Anthropology and Archaeology Department, University of Pretoria, Hat   󿬁 eld 0023, South Africa A R T I C L E I N F O  Keywords: Later stone ageForager-farmer interactionStone scrapersUse-wearCraft activitySouthern Africa A B S T R A C T Forager lifeways in the middle Limpopo Valley, southern Africa, were considerably altered from 350 CE onwardswhen incoming farmer communities settled the region. This is seen archaeologically in a shift in the preferencefor speci 󿬁 c Later Stone Age tool types and the introduction of farmer-associated items. Changes in foragerbehaviour have also been recorded at a number of sites from pre- to post-contact assemblages. Here we in-vestigate Little Muck Shelter where an overwhelming emphasis on scrapers was interpreted by Hall and Smith(2000) to indicate the production of surplus goods for trade with nearby farmers. We examine the use-wearindicators of the scraper assemblage to a) establish whether it is possible to identify activity indicators insouthern African Later Stone Age assemblages and b) determine whether di ff  erent activities were being per-formed between forager camps. Hall and Smith (2000) suggest the scrapers may indicate intensive hide pro-duction and we show here that they were additionally used in other craft activities also being performed at thesite. Along with Hall and Smith's (2000) work, our  󿬁 ndings allow for two important conclusions to be made.First, it is possible to di ff  erentiate activity behaviour at Little Muck over the past 2000years. Second, forageractivity patterns as a consequence of forager-farmer interactions varied between sites and across the landscape. 1. Introduction Over the past two millennia, southern African foragers and in-coming farmer groups interacted with one another (Mitchell, 2002:Chapter 8). These interactions led to a variety of social outcomes, in-cluding labour arrangements (Guenther, 1986; Wadley, 1996), assim- ilation (Hall, 2000; also see Solway and Lee, 1990), and shifts in forager lifeways (Mazel, 1989; Hall, 1994). Many of these changes can be seen archaeologically, such as the appearance of farmer-associated items inforager contexts (Deacon, 1984; Hall and Smith, 2000; van Doornum, 2005), shifts in forager settlement patterns (Moore, 1985; Mazel, 1989; Sadr, 2002), and the inclusion of foragers in farmer settlement struc-tures (Hitchcock, 1978; Wadley, 1996; Hall, 2000) and ritual practices (Dowson, 1994; Hammond-Tooke, 1998; Schoeman, 2006). While at the onset of contact forager groups occupied many parts of southernAfrica, by the late second millennium AD in most areas they no longerexisted or ceased practicing a traditional hunting and gathering lifestyle(e.g. Mitchell, 2002: 405 – 407). These interactions, therefore, had aprofound e ff  ect on foragers.Forager-farmer relations in the middle Limpopo Valley were quitedi ff  erent to those occurring elsewhere in southern Africa (Fig. 1).During the course of these interactions, farmer communities underwentstate formation (see Hu ff  man, 2009). This involved the gradual accu-mulation of wealth, control of the ritual landscape, and social strati 󿬁 -cation (Hu ff  man, 1982, 2009; Evers and Hammond-Tooke, 1986), which in turn a ff  ected forager society. Notably, farmer-associated goodsincluding trade wealth appeared in forager assemblages (e.g. Hall andSmith, 2000; Forssman, 2014a), indicating that they were included in distribution networks (Forssman, 2017). Their settlement structuresalso changed, seen in a gradual decline of forager material remains inrock shelter sites (van Doornum, 2007, 2008), such as stone and bonetools and jewellery, and its appearance in farmer contexts from  c . 1000CE (Forssman, 2014a; Seiler, 2017). That foragers witnessed the transformation of local society, and seem to have partaken in the farmereconomy at this time, is unlike any other contact scenario acrosssouthern Africa. Understanding the intricate network of forager-farmerrelations thus provides us with insights into the manner in which for-agers accessed parts of the farmer economy and acquired wealth.An important part of forager wealth acquisition was tangible andintangible forms of trade. Here we are interested in tangible trade,which we refer to as  ‘ crafts ’ . Craft and craft specialization is a widelydiscussed topic, beginning with V. Gordon Childe in the 1930s. It is 8 August 2016; Received in revised form 1 March 2018; Accepted 6 March 2018 ⁎ Corresponding author.  E-mail address: (T. Forssman). Journal of Archaeological Science: Reports 19 (2018) 287–3002352-409X/ © 2018 Published by Elsevier Ltd.    understandably a challenge to reconcile (Clark 1995). Thus, we follow afairly open-ended de 󿬁 nition by Costin (2015: 1) who summarises ‘ crafts ’  as  “ a wide range of non-food, tangible, utilitarian, and prestigegoods ” . She provides a more restrictive de 󿬁 nition of   ‘ craft specializa-tion ’  as: “… a di ff  erentiated, regularized, permanent, and perhaps in-stitutionalized production system in which producers depend onextra-household exchange relationships at least in part for their li-velihood, and consumers depend on them for acquisition of goodsthey do not produce themselves ” .(Costin, 1991: 4).Clark (1995) problematizes this de 󿬁 nition believing it is too re-strictive. One issue he raises, is the importance of scale. If viewing asingle site, the archaeological record may not show a cultural anomaly,thus presenting an apparently homogenous record. Therefore, oneneeds to view single site  󿬁 ndings within a broader, regional perspectivein order to determine which sites contain di ff  erent assemblages thatmay relate to craft specialization. Craft production is further de 󿬁 ned asbeing part of a co-ordinated system, which may be autonomous (Stein,1998), that accompanies a rapidly growing population (e.g. Charltonet al., 1991: 98; Beyyette, 2017: 12). These de 󿬁 nitions have speci 󿬁 cimplications in the middle Limpopo Valley which saw a populationgrowth among both foragers between the last centuries BC and 1000 CE(van Doornum, 2005) and farmers between at least 900 CE and 1300(Hu ff  man, 2005). We believe that since foragers partook in the localtrade network during farmer state formation processes (Forssman,2017), an emphasis on crafts developed during this time to accom-modate an increased demand for speci 󿬁 c items (see Forssman, 2015).In this paper we examine the forager response to farmer contact atLittle Muck Shelter by analysing the scraper assemblage to inferchanges in craft production. Our aims are to a) contribute to the ar-gument posited by Hall and Smith (2000) and provide data that de-monstrates shifts in forager activities at the site and b) compare these 󿬁 ndings to the regional sequence. Using traditional stone toolcategories to infer activity is problematic because it is based on mor-phological traits and assumed function (e.g. Dibble, 1987). While sometools' morphology may preclude them from certain tasks, most tooltypes can and were used in a variety of activities (Odell, 1975). As such,determining activity di ff  erences between sites cannot be done using theappearance of tool types alone. In order to do so a use-wear analysis isrequired (Rots and Williamson, 2006). Here we present the results fromsuch a study on stone scrapers recovered from Square L42 at LittleMuck (Fig. 2). The backed tool assemblage was not included in thisanalysis being so few in number ( n =27). The results demonstrate thevaried responses to farmer contact and aid in our understanding of therole forager communities played in local trade networks. 2. Little Muck Shelter Little Muck is a multi-component site with an internal sheltered areaand an outside courtyard. While both areas were occupied, the majorityof the Later Stone Age material came from the shelter. The site wasexcavated in 1998 by a team of archaeologists from the University of the Witwatersrand headed by Simon Hall. Six 1x1m squares were ex-cavated to varying depths, but only Square L42 has been published. Allartefacts were sorted in the  󿬁 eld and formal tools were stored sepa-rately in rigid containers. Since the tools were not collected with re-sidue analysis in mind, it is unlikely that a focused micro-analysis willbe successful, but it is not anticipated that macro-traces have beensigni 󿬁 cantly altered.The excavated deposit is composed of a series of culturally homo-genous stratigraphic units identi 󿬁 ed based on colour, compaction andcomposition (see Fig. 2) (Hall and Smith, 2000). A lack of ceramics in the lower units, ARB 2 and GS 2 (collectively referred to as ARB 2), maysuggest that it pre-dates the local appearance of farmers. While this isproblematic, since it assumes foragers would have acquired ceramics asfarmers arrived in the area, the levels above contain Happy Rest cera-mics (500 – 750 CE; Hu ff  man, 2007: 219), the  󿬁 rst undisputed farmer-produced facies that appeared in the valley. The Happy Rest levels Fig. 1.  The middle Limpopo valley with sites mentioned in the text and other prominent sites in the area. T. Forssman et al.  Journal of Archaeological Science: Reports 19 (2018) 287–300 288  contain two stratigraphic units, ARB overlying GS (collectively referredto as ARB), but both are thought to date between  c . 350 CE and 900.However, the division between the Happy Rest and underlying pre-ceramic units is unclear. Above ARB is PGA 3 with a suggested date of between 900 CE and 1000 based on Zhizo ceramics identi 󿬁 ed in thisunit. The division between PGA 3 and the lower (ARB) and upper (PGA2) units are clearly marked suggesting PGA 3 has experienced minimaldisturbance. The Units PGA 2, PGA, EA and PAH all date to the Leo-pard's Kopje period (1000 CE to 1300), or possibly thereafter, and onthe surface there appears to be a nineteenth century Venda use or oc-cupation (see Hall and Smith, 2000). This stratigraphic outline is basedon the excavation in Square L42, from which the assemblage analysedhere was recovered.From ARB 2 there is a massive increase in artefacts (Hall and Smith,2000: 35). Speci 󿬁 cally, this is in the density of stone and bone tools,ostrich and landsnail shell, ochre pieces and food waste. Hall and Smith(2000) suggest this shift may be because foragers moved into the  ‘ freespace ’  of the middle Limpopo Valley to avoid farmers settling aroundthe Soutpansberg, 80km south, or due to climatic improvements. Therealso appears to be a shift in tool production at the site. In the pre-contact phase there are more scrapers per 10l bucket (10/bucket) thanbacked tools (1.4/bucket), whereas after the appearance of farmers,  c .350 CE, the backed tool density remains low (1.6/bucket) but scrapersincrease nearly threefold (28/bucket) (see Fig. 3 for typological ex-amples of scrapers and backed tools). Hall and Smith (2000: 36) arguethat  “ [t]he high numbers of scrapers suggest production over and aboveimmediate forager needs, and consequently, an obvious intensi 󿬁 cationof hide production for local trade and barter ” . Changes at this stage of the site's occupation are thought to be linked to social interactionsbetween its forager occupants and incoming or nearby farmers.Although there is an increase in artefacts from the pre-contact intothe Happy Rest layer (ARB), the density of most categories only in factpeaks in the overlying layer (PGA 3). The amount of scrapers once againincreases massively (37.8/bucket) but so do all other archaeological 󿬁 nds, including faunal remains, landsnail and ostrich eggshell frag-ments, ochre and bone tools (see Hall and Smith, 2000; van Doornum, 2000). It is also in PGA 3 that glass beads appear for the  󿬁 rst time (Halland Smith, 2000: 35). Hall and Smith (2000) argue that this increase in Fig. 2.  Plan view of Little Muck (above) showing the location of Square L 42 from which the assemblage under investigation was recovered (adapted from Hall and Smith, 2000: 34 & 35)and the south wall pro 󿬁 le showing the chronological sequence (below). T. Forssman et al.  Journal of Archaeological Science: Reports 19 (2018) 287–300 289  Later Stone Age artefacts represents a shift in the site's use; it now be-came a workshop for the explicit purpose of producing crafts and goodsfor nearby farming communities. Hall and Smith (2000: 36) suggestthat the number of bone points and linkshafts and  “ the possible addi-tion of iron projectiles ”  (although none were reported) may indicatehunting was taking place in order to procure hides for trade. Metal doesnot seem to have replaced stone weapon armaments (backed tools; seeLombard and Pargeter, 2008; Shea, 2009) given that backed tools are consistently low in frequency throughout the assemblage, including inpre-contact levels, and that metal is rarely found at most farmerhomesteads during this early occupation period (Calabrese, 2000). Itcould be that hunting was occurring elsewhere with production-relatedactivities happening on site; if so, Hall and Smith (2000: 36) suggestthat foragers may have been contracted by farmers to prepare hides. Inaddition, little other evidence exists that suggests foragers were capi-talising on other tradeable resources, such as bead production, unlessthis manufacturing was taking place elsewhere (Hall and Smith, 2000:36). It was nonetheless  “ a period of intense interaction during whichmuch of the debris accumulated marks forager involvement withneighbouring Zhizo economies ”  (Hall and Smith, 2000: 36).The overlying layers PGA 2, PGA, EA and PAH are thought to date tothe Leopard's Kopje phase (1000 – 1300 CE) based on the appearance of diagnostic ceramics. During this period, the accumulation of wealth,control of the ritual landscape, political growth, and the appearance of social strati 󿬁 cation resulted in the establishment of the Mapungubwestate (1220 – 1300 CE) (Hu ff  man, 2009). When compared to  󿬁 rst mil-lennium contact or the pre-contact levels, there is a considerable de-cline in the frequency and density of most archaeological material inthe upper units, including a total absence of ochre. And yet, glass beadfrequencies increase and all diagnostic metal artefacts were recoveredfrom these levels (undiagnostic metal  󿬂 akes were found in PGA 3) (Halland Smith, 2000: 35). Found outside the shelter are also gaming( mankala ) boards pecked into the exposed bedrock and grinding hol-lows (see Fig. 2). For these reasons, Hall and Smith (2000: 37) believe the site was appropriated by farmer groups and that  “ foragers [duringthe Leopard's Kopje phase] became excluded from barter and craft ex-changes and were forced to select other responses ” . It is possible,however, that foragers were permitted access to the site because there isa low density of Later Stone Age remains in these levels. Yet, this maybe due to mixing between the terminal Later Stone Age and the farmer-occupation layers. Another possibility is that foragers assimilated intothe farmer economy only maintaining certain aspects of their own.Nevertheless, the desire from a farmer perspective to appropriate thesite may be due to the rock shelter's rock art sequence, which mighthave been seen as  “ critically ambiguous and, consequently, appropriatefor rites of passage ”  (Hall and Smith, 2000: 37).Little Muck's sequence can therefore be expressed in a series of stages: a pre-contact camp (<350 CE), an early contact camp with anemphasis on craft production (350 – 900 CE), an intensi 󿬁 ed workshop(900 – 1000 CE), and, possibly, a culturally appropriated place(1000 – 1300 CE) (we do not consider the later Venda occupation). Forthe sake of comparability with Hall and Smith's (2000) work, we followthese divisions in the presentation and discussion of our data. 3. Materials and methods Tools identi 󿬁 ed as  ‘ scrapers ’ –  having one or more retouched edgesat an acute angle (35° to 75°) (Walker, 1994: 3)  –  were separated intoraw materials and divided into three size classes following Deacon(1984): small (<20mm), medium (20 – 30mm) and large (>30mm). Fig. 3.  Examples of backed tools and scrapers from Dzombo Shelter: A – D, I & L, backed tools; E, J, small side scraper; F, small end-side scraper; G, H, K & N, small end scraper; and M,medium end scraper. T. Forssman et al.  Journal of Archaeological Science: Reports 19 (2018) 287–300 290  They were then further divided into morphological categories based onthe location of the retouched working edge in relation to the maximumlength of the artefact. Primary scraper forms are 1) end scrapers, wherethe working edge is at a distal or proximal end; 2) side scrapers, where alateral edge is worked; and 3) circular scrapers, where three or moreedges are worked. Combinations of end and side scrapers are 1) end-side scrapers, where one proximal or distal and a lateral edge is worked;and 2) side-side scrapers, where both lateral edges are worked. A  󿬁 nalcategory, scraper-adze, contains a typical scraper edge and a secondconcave, steeply retouched edge. While these morphological traits areused to determine whether a tool is a  ‘ scraper ’ , it is not assumed thatthey were all used in scraping activities. In order to establish this a use-wear analysis is required.The use-wear analysis was conducted in two phases (artefacts werenot cleaned in any solvent to avoid potentially removing organic re-sidues; e.g. Rots and Williamson, 2006: 1288). In Phase 1, all scraperswere analysed using handheld low-power magni 󿬁 cation (10×) in orderto separate scrapers with and without trace evidence, considered hereto be any artefact with evidence of use-wear. Both faces were examinedand extra attention was given to the working edges (see Langejans,2011). Scrapers that appeared to have use-wear or questionable oruncertain trace indicators were re-analysed in Phase 2 using a NikonSMZ 745T stereomicroscope with a magni 󿬁 cation of between 10× and300×. The face on which the use-wear was identi 󿬁 ed (dorsal or ven-tral) was recorded along with the location following D'errico (2012):edge area (the zone along the working edge), inner edge (immediatelybehind the working edge) and inner surface area (innermost area). Tothis we added the categories opposite working edge (directly oppositethe retouched end) and restricted portion (occurring in a limited area).Four di ff  erent use-wear (trace) indicators were considered: rounding,polish, edge damage and macro-fractures. The combination of thesefeatures may suggest use, including speci 󿬁 c activities and materialworked (e.g. Keeley, 1980; Moss, 1983; Vaughn, 1985; Shea, 1992; Jahren et al., 1997; Hardy and Garu 󿬁 , 1998; Lombard and Wadley,2007; Rots, 2005, 2014; Miller, 2014). All identi 󿬁 ed use-wear in-dicators were photographed and measured. Provided below is a sum-mary of each use-wear feature used in this study and its identi 󿬁 cation. 3.1. Rounding  The extent of rounding is dependent on the time spent using the tooland the material that was worked (Donahue and Burroni, 2004). Threecategories were used to record this: slight, moderate and considerable.Slight rounding refers to tools possessing evidence of rounding alongprotruding edges, such as  󿬂 ake scar ridges. Tools with moderaterounding exhibit a greater degree of rounded edges that are clearlyvisible and have  ‘ smoothed ’  jagged or sharp portions along the workingedge. Lastly, considerable rounding refers to tools with thoroughlyrounded edges where protrusions as well as the concavities betweenthem have been clearly rounded. A single tool could possess multiplerounding forms in di ff  erent portions of the tool's edge. These toolsmight still have sharp portions of the working edge, jagged regions or ‘ fresh ’  working. However, the  󿬂 aking scars could also be as a result of incomplete sharpening of the edge or from interrupted resharpeningduring the activity (Rots, 2005: 62), which may impair our ability todetermine the extent of rounding. Therefore, in cases where multiplerounding categories were noted only the most extreme degree of rounding was recorded. The percentage of the a ff  ected edge that ex-hibited rounding was also recorded. 3.2. Polish Polish most commonly occurs at the working edge on both dorsaland ventral surfaces. It appears in broad bands from the edge area to-wards the inner surface area (Binneman and Deacon, 1986) and isshinier or rougher than the surrounding surface (Rots and Williamson,2006). The extent of polish depends on the time spent using the tool,the hardiness of the material worked and the tool's raw material(D'Errico, 2012). Striations can form on the polished section of thestone tool when hard particles such as sand are trapped between thetool and object being grooved or scraped (Donahue and Burroni, 2004;D'Errico, 2012). Polish forms from working various hard and soft ma-terials such as antler, bone, domestic produce, hide, reeds, siliceousgrasses, wild cereals and wood (see González-Urquijo and Ibáñez-Estévez, 2003; D'Errico, 2012). Di ff  erent materials also produce dif-ferent polish types which are usually but not always identi 󿬁 able(Binneman and Deacon, 1986; D'Errico, 2012). Hafting might also produce a bright polished spot on the haft limit of a tool (Rots, 2005).Binneman and Deacon's (1986) experimental study on stone adzeswas used in this analysis. They found that greasy and dull polish de- 󿬁 ned as smooth well-polished areas that vary in intensity across thedomed areas of the tool formed from working  ‘ clean ’  wet and dry wood,respectively. Where pitting was found within this matrix, an abrasivehad been present but this might also result in striations. Charcoal polishis far brighter and is formed in the direction of use. Also used here wasShea and Klenck's (1993) matted polish which develops from hideworking (see Rots, 2005). Noticed in this study is the impact raw ma-terial has on identifying polish; cryptocrystalline silicates (CCS), fromwhich most of Little Muck's scrapers were made, is a  󿬁 ne-grainedsmooth material which makes it di ffi cult to distinguish polish from thenatural, unpolished surface and from a gloss or increased lustre causedby heating (see Domanski and Webb, 1992; Wadley et al., 2017). Ad- ditionally, polishes are identi 󿬁 ed with a certain level of subjectivity,questioning their reliability (cf. González-Urquijo and Ibáñez-Estévez,2003). 3.3. Edge damage Along the edge of an artefact one might also  󿬁 nd crushing and step 󿬂 aking on both the dorsal and ventral surfaces. Binneman and Deacon(1986) identi 󿬁 ed these features on adzes used in wood-working andShea and Klenck (1993) noted stepped  󿬂 aking as a result of workingrigid surfaces, which includes wood and bone (Rots, 2005: 67). Fatiguewear might also form as a result of sheer stress causing micro-cracks inthe tool surface (Diamond, 1979; Arnell et al., 1991). Polish and rounding can assist with establishing whether some forms of edge da-mage are a result of use, manufacture or post-depositional processes;rounding of scar edges in a polished area indicates the damage occurredduring use (Donahue and Burroni, 2004). However, if polish orrounding does not occur it could also indicate that the artefact was notused as repeatedly, intensively or for the same purpose after the da-mage occurred; it thus does not reliably exclude the possibility of post-depositional interference. 3.4. Macro-fractures Macro-fractures diagnostic of impact include bifacial and unifacialspin-o ff  fractures (forming o ff  of a bending fracture at the impact-end of the artefact), step terminating fractures (longitudinal fractures endingin a 90° step) and impact burination bending fractures (a bendingfracture along the lateral edge terminating in a 90° step; see Lombard,2005). Other fracture types are notches (arguably related to impact; seeYaroshevich et al., 2010; but see Lombard and Pargeter, 2008), hinge and feather terminations and snap fractures. It is not necessary toprovide a thorough review here since only a few macro-fractures wereidenti 󿬁 ed in Little Muck's assemblage (see Fischer et al., 1984 andLombard, 2005 for a review of fracture types).While use-wear analysis has the potential to indicate tool use, theabsence of use-wear does not necessarily mean an artefact was not used;the occurrence of polish, rounding, damage and residues (the latter notconsidered here) is dependent on repeated use, material hardness,viscosity, elasticity and surface texture (Yamada, 1993; Donahue and T. Forssman et al.  Journal of Archaeological Science: Reports 19 (2018) 287–300 291


Mar 27, 2018
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks